How is Anthropic, OpenAI and xAi going to compete against the likes of Google that can spend $200 billion a year? It’s an impossible war and all these investors are throwing their money into a bottomless insatiable pit of money.
Until the funding stops for one reason or another and then everyone loses all their money at once like a star that collapses into a black hole singularity in a femtosecond.
As someone who thought Google+ doomed facebook, because of Gmail accounts and everyone with Google as their homepage already, I learned not to overestimate Google’s abilities.
1. Google had recently exploited their home page to push chrome browser successfully altering the browser market. They pushed anyone visiting Google to chrome with a popup on the home page. The same opportunity was there for G+, but with updates from friends.
2. Everyone already had a Google account and many millennials were using Google Talk at the time. It appeared Google could undermine the network effects.
3. The UI of G+ appeared better
4. Facebook had released the newsfeed otherwise known as ‘stalker mode’ at the time and people recoiled at the idea of broadcasting their every action to every acquaintance. The circles idea was a way of providing both privacy and the ability to broadcast widely when needed.
5. Google had tons of money and devoted their world class highly paid genius employees to building a social network.
You can see parallels to each of these in AI now. Their pre existing index of all the world’s information, their existing search engine that you can easily pop an LLM in, the huge lead in cash, etc. They are in a great position but don’t underestimate their ability to waste it.
So is Gemini tbh. It's the only agent I've used that gets itself stuck in ridiculous loops repeating "ok. I'm done. I'm ready to commit the changes. There are no bugs. I'm done."
Google somehow manages to fumble the easiest layups. I think Anthropic et al have a real chance here.
Google's product management and discipline are absolute horsesh*t. But they have a moat and its extreme technical competence. They own their infra from the hardware (custom ASICs, their own data centers, global intranet, etc.) all the way up to the models and product platforms to deploy it in. To the extent that making LLMs work to solve real world problems is a technical problem, landing Gemini is absolutely in Google's wheelhouse.
Hard to bet against Hassabis + Google's resources. This is in their wheelhouse, and it's eating their search business and refactoring their cloud business. G+ seemed like a way to get more people to Google for login and tracking.
A couple months ago things were different. Try their stronger models. Gemini recently saved me from a needle in a haystack problem with buildpacks and Linux dependencies for a 14-year-old B2B SaaS app that I was solving a major problem for, and Gemini figured out the solution quickly after I worked on it for hours with Claude Code. I know it's just one story where Gemini won, and I have really enjoyed using Claude Code, but Google is having some success with the serious effort they're putting into this fight.
I think they had no choice but to release that AI before it was ready for prime time. Their search traffic started dropping after ChatGPT came out, and they risked not looking like a serious player in AI.
They recently replaced “define: word” (or “word meaning”) results with an “ai summary” and it’s decidedly worse. It used to just give you the definition(s) and synonyms for each one. Now it gives some rambling paragraphs.
I use Claude every day. I cannot get Gemini to do anything useful, at all. Every time I've tried to use it, it has just failed to do what was required.
I thought it was a far superior UI to facebook when it launched. I tried to use it but the gravity of the network effect was too strong on facebook's side.
In the end I'd rather if both had failed. Although one can argue that they actually did. But that's another story.
I very much wanted Google Plus to succeed. Circles was a great idea in my opinion. Google Plus profiles could be the personal home page for the rest of us but of course, Google being Google...
That being said, tying bonuses for the whole company on the success of Google+ was too much even for me.
Everything was obviously DOA after it dies. I also thought it wouldn't last but it wouldn't be the first or last tech company initiative that lived on long after people thought it would die. Weird things happen. "Obviously" isn't a good filter.
Theoretically Apple can spend just as much. What are the outcomes though? All those giants have their own business that are established and profitable.
It’s the new kids in the block that will make the difference.
You know those lists on twitter about how many companies US has in top 10 and are presented as a win? Those are actually lists of capital concentrations blocking innovation. It looks like US is winning but for some reason life is better in EU and innovation is faster in China.
It’s companies like OpenAI Anthropic that will move US ahead. Even if some core innovation or and capital comes from the establishment.
> Theoretically Apple can spend just as much. What are the outcomes though?
The GP was talking about Google specifically, and their outcomes on AI are nothing to scoff at. They had a rocky late start, but they seem to have gotten over that. Their models are now very much competitive with the startups. And it's not just that have more money to spend. They probably have more training data than anyone in the world, and they also have more infrastructure, more manpower, more of a global footprint than the startups.
The Innovator's Dilemma is an anecdotal, maybe a statistical relationship at best, but not a fundamental law of nature. When an established company has everything it should take to become a leader in a new industry in theory, and in practice their products are already on par with the industry leaders, you know at some point it becomes rational to think that maybe they might become a leader.
Google didn’t have a late start, they invented the tech, had bespoke hardware in place that supported it and have money to spend.
I don’t have any idea what comes next but Google and Microsoft look bad right now because they can’t execute a product strategy.
My personal bias is that either ms or Google or both will land just fine after it all shakes out but they started with a lead and are now playing catch up.
they did have a late start in terms of productionizing the models. It's definitely improved but there was a time where Gemini and the associated tools werent as good as claude/oai
Sometimes I worry about the incentives for innovation in the US.
Step 1, find something to innovate on, sell the promise of it to investors.
Step 2, build a prototype or worst case, build it for real and start generating income from your truly innovate and unique product.
Step 3, get acquired by a large company and then shut down because your product competed with theirs.
End result, general public possibly benefited from your innovation, but in the long run, it was temporary.
Maybe the incentives would be better if it were harder for large companies to acquire small ones? If the path to riches where driven primarily by delivering value to customers. Would love to hear other's opinions on this.
> Theoretically Apple can spend just as much. What are the outcomes though? All those giants have their own business that are established and profitable.
Ah! Well, if we put aside "The Innovator's Dilemma" and pick up Reis and Trout's "marketing Warfare," we get the answer. Apple does have an existing business, but investing in AI does not cannibalize it. They can throw money at it, try to find a way to make it work really, really well for consumers on very specific custom hardware in their devices...
Likewise, someone like Google has all the money in the world to throw at it, but they aren't investing in a new market, they're defending their search business against everyone just asking a generative AI Chatbot questions. I\But it's possible for them to screw this up internally over turf wards, just ask the engineers who tried to make search better but were kneecapped by Prabhakar Raghavan who demanded that search be poor enough to drive people to click sponsored results.
In the "Marketing Warfare" model, Apple is attempting a flanking attack: An outsider trying to disrupt the AI giants with an approach that they can't imitate without undermining their value proposition. On-device AI flanks the big giants that areservcie-centric.
And in that model, Google is playing defence, which is what every leader is supposed to do. Their job is to "cover every move," which they are doing in textbook fashion. If AI goes away, Google dry their tears and continue to mine ad revenue.
I'm very curious if they're going to ditch Google by providing on-device search. A monthly Open Crawl is under 100 terabytes, and if you clean that down to raw text and deduplicate and maybe pick out what you don't care about, the dataset might already fit onto my iPhone. They could do a lot without making a network call and reach out to a server for anything the device doesn't have, but a lot of user queries might never need to leave the phone. In another couple years, storage will be even higher.
I was hopeful for on-device AI too but any AI processing so far sucks up the battery, heats up the phone and most importantly isn't even nearly good enough. Without a breakthrough in battery, chips or the models and algorithms the way forward is thin clients that connect to some servers close to a solar farm or nuclear energy plant.
> On-device AI flanks the big giants that areservcie-centric.
Wouldn't on-device AI also support Google's position? If search is to be protected, on-device AI (small models) would be capable of basic usage, but inept at answering knowledge questions specifically, necessitating a search service be preserved. They have already launched local models in Chrome and Android. Meanwhile none of the big AI competition can profit off of local models, so this is a unique opportunity for big-G.
That said, I disagree with the premise you propose. It's 2026, and about 40% of their revenue over the last few years comes from non-search products (depending on quarter). Oh and Apple doesn't seem to be investing enough in AI products, because it's just making them look bad, not providing a "flanking attack".
Google is pulling in tons of AI revenue - from subscriptions, personal and enterprise, and Google Cloud (APIs etc). Cloud is seeing a ton of growth lately, and I'm sure that's largely from AI services that are uniquely available there. As long as they can serve models with a better cost structure (thanks TPUs) they can squeeze out better margins than their competitions.
well, it's basically existential, so the incentive is there to not only get it very right but also to limit the delta with how right anyone else gets it. The same can't really be said of the long tail of products Google have done.
Look to GCP as an example. It had to be done, with similar competitive dynamics, it was done very well.
I hadn't heard that, that's interesting. Any sources you'd recommend to hear more about it?
I think it's a slightly different point though. What I'm saying isn't about where the idea came from or whether it was part of some precient top down bet / strategy from the very beginning.
It's more where did the strategy evolve to (and why) and did they mess it up. GCP and Android are good examples of where it at a minimum became obvious over time that these were massively important if not existential projects and Google executed incredibly well.
My point is just that there's therefore good reason to expect the same of LLMs. After all the origin story of the strategy there has a similar twist. Famously Google had been significantly involved in early LLM/transformer research, not done much with the tech, faltered as they started to integrate it, course corrected, and as of now have ended up in a very strong position.
> well, it's basically existential, so the incentive is there to not only get it very right but also to limit the delta with how right anyone else gets it. The same can't really be said of the long tail of products Google have done.
I've yet to see anything that threatens Google's ad monopoly.
It's not that a dominant position goes away overnight. In fact that would be precisely the impetus to spur the incumbent to pivot immediately and have a much better chance of winning in the new paradigm.
It's that it, with some probability, gets eaten away slowly and the incumbent therefore cannot let go of the old paradigm, eventually losing their dominance over some period of years.
So nobody really knows how LLMs will change the search paradigm and the ads business models downstream of that, we're seeing that worked out in real time right now, but it's definitely high enough probability that Google see it and (crucially) have the shareholder mandate to act on it.
That's the existential threat and they're navigating it pretty well so far. The strategy seems balanced, measured, and correct. As the situation evolves I think they have every chance of actually not being disrupted should it come to that.
Because Google has the money to build 10 different versions/iterations of Gemini and can essentially force one to work. They have most people's data and most people use them for mail/search/browser/maps as well.
In my opinion though this is a race to the bottom rather than a winner takes all situation so I don't think anyone is coming out ahead once the dust settles.
Does it matter? Microsoft won by default with Teams because it actually turns out no one cares about chat or even has a choice in it: employees use whatever the company picks.
If you're going to say "other than the US" then you've got to say at a minimum "other than the US and China", but really "other than the US and China and Japan and Korea and Taiwan and Thailand and Russia and most of Central Asia".
Only mentioning the US is wildly americentric even by HN standards.
The tech behind wave eventually made its way into Google docs though and pioneered collaborative document editing, so wasn't a complete failure even though the product itself was killed.
No comment on Google+, Google has a storied history of failure on any kind of social media/chat type products.
Where Google wins is just simply having enough money to outlive anyone else. As the saying goes "the market can remain irrational longer than you can remain solvent" In this case, Google is the market and they can just keep throwing money at the wall until OpenAI, Anthropic, etc. go under.
They target those ads by ingesting as many signals as possible from as many input devices & sensors as they can possibly convince people to use. They make a lot of money from advertising b/c they have managed to convince the most number of people to give them as many behavioral signals as possible & they will continue to do so. They kill products only when the signal is not valuable enough to improve their advertising business but that's clearly not the case w/ AI.
Depend on the definition of the "product". For example some banal cloud storage in which everyone competes. And it's an "old" product, despite being invisibly improved behind the scenes, just like at any other provider. Google has pretty competitive storage AND they are fully abusing Android integration for AND they have pretty good bundling of that storage with other products, including, you've guessed it - LLM Gemini. So say a person is not a professional user of LLMs like a developer burning tokens in a dozen accounts simultaneously. A person has a phone and eventually memory runs out, so he buys a one click Google storage for 4 bucks. And suddenly he has Gemini Pro included too. So why pay 20 bucks to Anthropic, when Google costs 1/5 of that AND has other stuff bundled too?
So maybe Google is lagging on truly new products (btw, does Gemini itself with its TPUs count as a new product? I'd say yes), but "old" products are entrenched enough to carry them and compete.
Google Drive is easily the worst of the desktop cloud storage options. It’s okay for Google Docs but not other files if that’s what you’re talking about..
Im not sure what you consider successful. They've been struggling to get market share vs azure, and the product isnt that good. lots of rough edges, and piss poor support
Their API business model seems to be hope enough people accidentally go over free tier: $0 for the first 5000 monthly places lookups, $40 per 1000 after that
Few years ago, we had Google Bard, the ancestor of Gemini, which was supposed to be an AI LLM, and when you right-clicked the page, it was a fake page with hardcoded sentences in a .js file...
Does Alphabet/Google have any other significant alternative revenue streams though besides their ad revenue? And won't that decrease significantly the more people use AI tools for research than firing up a google web search? I find myself using Claude more and more doing web research and comparing products/reviews...without getting a single ad served up from Google.
The conclusion Google is engaged in consumer capitalism is wild.
They're engaged in computing research and merely engage in consumer capitalism as a consequence of political and social constraints.
Products are a means to an end not the goal.
OpenAI and Anthropic are product companies and are more likely to fail like most product companies do as they will lack broad and wide depth.
Google has experience in design, implementation, and 24/7 ops with every type of SaaS there is. They can bin LLMs tomorrow and still make bank. Same cannot be said for OAI or Anthropic.
I'm trying paid tier Gemini and it doesn't allow to keep have personal chat history when you disable training on your data, on reload of the page your chat is gone. Even free tier of ChatGPT allows disabling training on your data while allowing to keep such basic functionality.
Some technical advancements are not worth it if you do not respect your users.
Google is evil in passive way, like sprawling bureaucracy making you life slowly worse and worse but also doing some stuff to at least some fraction of population. OpenAI and Sam are determined and energetic evil, laser focused on making whole human population jobless and homeless in shortest way possible and not producing anything else of value, no other products. I'd rather prefer the former evil out of the two.
Another basic feature that’s missing is sharing a Gemini chat as a link anyone can view.
OpenAI figured this out: it’s awesome marketing when people send each other links to the app with a convenient text box to continue the conversation. It’s viral.
Google meanwhile set this up so that “anyone with the link can view” is actually “anyone with the link and a Google account”.
That’s grade A failure of marketing.
The PM in charge of that decision ought to be walked off a plank.
Because it's Google they can't build products and they only care about benchmarking.
The product they released so far are all half assed experiments.
Gemini 3 Pro is now being beaten by open source models because they can't fix or don't want to fix the problems with the Gemini models being completely useless.
The same for Microsoft.
Microsoft had GitHub Copilot, and Microsoft Copilot and both of them are useless to Claude Code and Claude Cowork.
You can have all the money in the world, but nothing is stopping you from building useless garbage.
I guess it depends on your spending. GPT-5.2 and -5.3-Codex are certainly much cheaper: you get much more from the same $20 sub. When I was using Claude as primary I would daily hit limits and have to wait vs on GPT with more usage it only happened to me once in a few months when I was vibecoding non-stop for a week or two to port my personal Windows tools to Linux with multiple other projects being worked on in parallel.
Anecdotally GPT was also smarter than Claude which prompted my move from Claude in the first place: Gemini and Claude back in October failed to get their own harness PID.
Work smarter, not harder: find an innovation that makes LLM computation an order of magnitude more efficient. Easier said than done, of course, but vastly more efficient biological systems provide an existence proof that this is possible in principle. DeepSeek already showed how it's possible to get good results with limited resources.
Well there's a good reason that OpenAI partnered up with Microsoft. The calculation is that the established big techs - Amazon, Apple, Microsoft, Google, Meta are all going to be significantly impacted by AI so it's not unreasonable to look at Anthropic at 10% of their market cap as a reasonable value. Would it be worth Apple to bring Anthropic in house? They failed to deliver AI themselves, they know the risks of being dependent on Google. If AI goes far enough it may totally remove Apple's differentiation.
Some of the Big Techs are building their own in house stuff (Meta, Google), but it wouldn't be crazy to see acquisitions by the others, especially if the market cools slightly. And then there's the possibility that these companies mature their revenue streams enough to start actually really throwing off money and paying off the investment.
> they know the risks of being dependent on Google
I wouldn't argue it's that risky. Look at their past entanglements:
1. Google Default Search Bribe - brings in $20B a year for literally doing nothing
2. Google Maps: Google let them build their own custom app using Google's backend, and it worked fine all the way up until Apple chose to exit that arrangement
actually I can't think of any others, but is there an example of Apple getting burned by Google?
Well they have an illegal monopoly over display ads that they defended with a moat of money when FB tried to butt in, so that's an example of them being not great to rely on.
Slight counter point - claude code is basically the only developer tool that ever been happy to pay money for. Getting the entire software industry to give you $200/mo/person is quite the market.
These companies are operating outside of "normal" economics. They seem to be fed money against all rational valuations because it is a Manhattan project - the US needs to 'beat' the Chinese with this and as such these companies will probably not be allowed to fold.
SpaceX makes 16B in revenue per year, with 7B in ebitda (which doesn't account for the cost of rockets)... so assume what, 3B in free cash flow per year? And that's being generous.
That's about what Google creates in free cash every 2 weeks.
They have the best coding agent around. I would say roughly half the industry no longer writes code by hand and claude is/was the only agent that let me drop my editor all together. That enabler alone is enough signal beyond hype.
Frankly google models and the UIs google designs around their models just aren't as powerful and more training, more data and more compute is no longer tipping the scale. Anthropic did something to make their model better at coding than almost anything else.
And all of this is just what's happening right now. The money being invested is a gamble not on right now... but on the trendline to the future. I agree that the LLMs are overloaded with hype, but the people who compare it to crypto aren't thinking straight. Whether there's over investment or not a paradigm has shifted. Maybe there will be a collapse, but it won't collapse into a singularity. If it collapses at all, a new world will emerge, and that new world will generate more value than all the money currently invested in AI.
Have you ever used anything that is on google cloud console? Or tried not to get randomly ratelimited with a single request to a vertex llm model? They are shooting themselves in the foot for solid 20 years, any of these players can compete with google in this frontier
I think GP is probably implying that this particular vertical requires obscene amounts of capital to keep up, which makes it really hard for a startup if you’re going up against businesses with giant free cash flow machines.
It’s the same reason Reid Hoffman sold his AI startup early… he realized he just couldn’t beat Google/FB/MSFT long term if it devolved into a money race.
Just because something is in do-or-die situation doesn't mean that they have some kind of magical advantage over fat cat. Being in that situation means there is very real possibility of doing "die" part and we have lots of examples of them doing so.
Basically "we have youtube subscribers" is the only thing that isn't all about AI, but even that i'm sure they're trying to figure out how to shoehorn AI into that product
The most efficient way Google could spend that money is probably to buy a company and not poke at it too much. I have no confidence that large rich companies can actually innovate beyond buying small innovators or spawning business units and not poking at them much.
Sundar "the manager" has presided over an enormous growth of the businesses he was handed. He also presided over the complete collapse of the internal culture. OTOH he may have fired Dianne Green, so that's something. Overall, at best Meh.
Demis ran a startup that burnt cash on vanity projects and continues to burn cash on vanity projects. Gemini is barely open source quality AI, but Google makes it nearly free and has the best distribution on the planet.
Gemini has been a joke since 1.0. No release has hurt Google's brand more. 3.0 was STOA for about 2 days, easily Gemini's best release.
Anthropic and OAI are moving at amazing pace, Google can not keep up at all.
> "It has been less than three years since Anthropic earned its first dollar in revenue. Today, our run-rate revenue is $14 billion, with this figure growing over 10x annually in each of those past three years."
Wild although not entirely surprising. Congrats, Anthropic.
Pay attention to the outflow of tech investment in the stock market. That money is going to move into OpenAI and Anthropic IPOs. The valuations will be as big you are thinking because the market believes these companies will represent an entire basket of startups.
That's a really interesting tidbit. Thanks for sharing.
And thinking about it it makes sense since the decision to pay the outrageous rates for an ad during the Superbowl must be driven by strong emotions (confidence or desperation). In this case, considering there's no clear moat for any of the big players, I believe it's the latter.
I really wonder is there even enough dump money from them to sell the stock they hold. Not to mention even raising any new capital... Is there really enough bag holders that will run after these stock with large enough piles of money?
Could you be more specific? Because NVDA has a consistent 20 year growth of something like 400x and +30%/yr, so I don't think the bitter lessons are there.
I would hold off congratulating them until they’re actually in the black. They are still burning billions a year lol the revenue is impressive but their expenses are still solidly north of it.
How is this not a red flag? It’s a series G and they are still begging for money to burn. When do they convert their balance sheet to profit? Is it after a series AAF when they are worth more than Apple or nvidia?
Those growth stats for Claude Code are pretty wild:
> Claude Code was made available to the general public in May 2025. Today, Claude Code’s run-rate revenue has grown to over $2.5 billion; this figure has more than doubled since the beginning of 2026. The number of weekly active Claude Code users has also doubled since January 1 [six weeks ago].
Doubling both annual run-rate revenue and weekly active users in the first six weeks of this year!
No doubt it's astounding, and I have no experience underwriting these kinds of scale of investments...
...but if I were a recent employee diluted as part of this raise (even with the massive uplift in revenue) I'd be very skeptical about an sort of financial outcome for myself.
OpenAI has "open" in their name but is closed off to public access and input
Google used to have a motto "don't be evil"
Who enforces the definition of language? Who demands compliance?
Soon as we go down the path of policing and insistence on one true dogma, we veer into religious holy war type behavior.
Obsession with semantics of syntax is a sort of theism even if the syntax and semantics do not refer to the commonly accepted tropes of a specific religion.
In this case, it has to do with how they're classified under Delaware law as a corporate entity.
I'm not a lawyer (I don't even play one on TV, damn you Odenkirk) so I can't tell you what that means as far as case law for companies getting punished for behaving badly, but in this case, there is supposedly some sort of legal backing for the classification.
I know PBC; legal semantics that dictate various financial terms. I also know the specifics because for some reason those specifics matter to some people, and it's been personally beneficial to have more than an awareness of such legalese. Makes true believers feel I am living up to their "one of us!" means testing.
I accept they matter to others but reject such exists as anything but contemporary ethno-objects. Similar to how I acknowledge Christianity exists but am not a disciple of the dogma; I have to accept others believe are but do not have to pledge allegiance myself.
Aside from engineering and healthcare, machine operation, with real safety implications, everything else is all just parroting and social role-play.
Being a VHS copy-paste of generational semantics isn't the flex you want it to be. Patronize harder though.
They've been very clear about their mission, they're doing more than anyone else when it comes to it, and if you've ever interviewed with them you'd know how critical it is to them.
But I guess it's easier to make a glib comment than look these things up.
Anthropic has one of the best moats of any business that's been created in the last 50 years.
Numerous companies have tried and failed competing with SoTA foundational models. If Anthropic had no moat, Apple and Meta wouldn't be paying them billions for coding asistance.
Meta, Amazon, Apple, and Nvidia would all have SoTA competitors to Claude. They all tried and have not produced a competitor.
Instead you have three companies that stand alone making billions from foundational models.
The open models are not far behind. Is it really a "moat" when it's so short lived and you need a brand new moat after six months? That's just ordinary competition.
Two years ago, I considered investing in Anthropic when they had a valuation of around $18B and messed up by chickening out (it was available on some of the private investor platforms). Up 20x since then ...
It was always obvious that Anthropic's focus on business/API usage had potential to scale faster than OpenAI's focus on ChatGPT, but the real kicker has been Claude Code (released a year ago).
It'd be interesting to know how Anthropic's revenue splits between Claude Code, or coding in general, other API usage, and chat (which I assume is small).
Eh, I think you made the best decision you could given the info you had.
I’ve poked around on EquityZen and was shocked at how little information is available to investors. In some cases I did not even see pitch decks, let alone one of the first companies I looked at had its top Google result: CEO recently arrested for fraud and business is almost worthless now.
Unless you are willing to take a blind punt or have insider information, those platforms are opaque minefields and I don’t fault you for not investing.
Matt Levine has a fun investment test: when presented with an opportunity, you should always ask, “and why are you offering it to me?”
Meaning, by the time it gets offered to retail investors (even accredited ones are retail) we’re getting the scraps that no one else wants.
Hiive and Forge Global are the ones I know of. You must be an "accredited investor" which means nothing at all except that you have a million dollars or make $200k/yr.
Like you can buy shares of Anthropic as long as you prove you make over 200K? That easy? Shouldn't they approve of the purchase? Sorry, noob in this space!
They have to approve and it's not as simple - it's just that if you make $200k a year or have $1m in the bank, the government assumes you're a knowledgeable investor and allows you to bypass certain protections.
If you are NOT knowledgeable and simply have money ... well it'll soon be parted.
The secondary platform verifies you and then you indicate interest. If there’s a seller you may get to buy. Company may ROFR. Priority goes to bigger buyers.
A few years back, well ok maybe almost ten now, but regardless- a recruiter reached out to me about a role at a "series G" company like it was a selling point, and I was just kinda like ok maybe thats signaling its relatively stable and can raise money, but at the same time, that's a lot of rounds to have preferences ensure unprivileged shareholders get nothing, and also to have most of the hockey stick growth already tapped out.
This was in the middle of the boom when companies were fighting over talent, so I found it odd.
I wonder how good it is for companies to be allowed to grow so big and still be private? Would it makes sense to require any company with more than a billion dollar valuation to be subject to all the same SEC requirements that public companies are? Could companies be blocked from raising money once the reach a crazy valuation like $1 billion?
That doesn’t make sense at all. The raison d’être of SEC is to protect regular mom-and-pop investors. A private company just doesn’t allow anyone to invest in them. Why should SEC rules apply? On what legal basis can you force a private company to divulge its financial details? Would you be happy if you, as an individual, have to divulge your account statements if your own net wealth reaches one million?
> The raison d’être of SEC is to protect regular mom-and-pop investors
That's not the sole reason. They (should) also enforce a fair even playing field by preventing market manipulation (e.g. how Elon was tweeting about stock prices) and a few other things to "facilitate efficient markets and the formation of capital".
> Why should SEC rules apply?
Because private companies still fall under the jurisdiction of the SEC? See e.g. Theranos.
> On what legal basis can you force a private company to divulge its financial details?
On the to-be-created legal basis that aims to prevent bubble formation and the resulting fallout to the wider society?
> Would you be happy if you, as an individual, have to divulge your account statements if your own net wealth reaches one million?
Sure, why not? It's not a totally unheard of idea. In Norway everyones salary can be looked up.
yeah it's a slippery slope forcing companies to go public at X valuation. who decides that? what number makes sense? etc. but i do think we need to somehow fix massively overpriced companies going public and dumping on retail
it's still the case, but there are never 1,000 investors, there's a couple dozen VC firms, SPVs, and individuals
if you're smart.
I don't think this is an SEC problem, they are fully aware that people subject to their jurisdiction can jump through many hoops to circumvent them. This shows consent on the investor's part well enough, and capital formation regulations do not burden the investor at all, they are only constitutional because they burden the organization raising capital, who simply needs to do a cursory check - not an in-depth one. (level of depth is based on which regulatory exemption is chosen)
So as long as you separate concerns the SEC is satisfied.
If employees get stock options and decide to exercise on exit, they count against the 500 unaccredited investor limit that would trigger reporting requirements. So companies that issue stock options do have an outside risk that enough employees will exit, exercise their stock options, and trigger a reporting requirement.
Goldman Sachs recently stoked fear about software stocks due to claimed AI competition.
What if their strategy is this: slowly drive down software stocks, keep talking about AI, buy the downward market. Then cash in on the IPOs of OpenAI and Anthropic.
Then let OpenAI and Anthropic implode. Goldman Sachs had no problems underwriting webvan at the end of 1999, which then imploded in 2000.
Anyway, I just valued my dog at $1 billion post-money. You can buy it at pets.com.
Matt Levine has put this forward in his newsletter - if you're moderately influential you can go on TV and tell people that "X industry will be dead in 10 years" because of AI and then profit from the inevitable stock dip.
Because we live in the worst possible timeline the end result for AI companies does seem to be "too big to fail", where these massive investments will get foisted on working class people via a bailout or an IPO and index inclusion.
I don't see why any AI company would ever be "too big to fail". I can't see why any government would be motivated to take the political hit of bailing them out.
Claude Code is trivially an attempt to hobble the rest of the software business: the PID controller, the control vectors, the ever change loss surface, the bash and JSON jank at the foundations, the no one is this stupid context management, the some-data-critical-to proceed | tail -n 5, the sed editing, the speculative execution of partial frames.
OpenRouter and Opencode show you how behind it is, that bootstraps you off of them. They have issues too and Zen is starting to feel icky, but they let you speed run to the next thing.
You're attributing way too much intent to what is the viewpoint of some random analyst at Goldman Sachs (who doesn't even control any purse strings). A year ago there was another big hullabaloo when a GS team wrote a long post about how AI companies would never make enough revenue.
I haven't used it to replace workers though, only to replace Google search. My company is pushing copilot but it's only $16/user/mo. Hardly lucrative and no moat.
Claude Code and Cowork are incredible products, and can do much more than just search. Lots of people are paying hundreds of dollars a month for them.
If you’re just using AI for search then I can see why you’d not see the value. But many people really are getting a huge amount of value out of agents, and are already paying for it.
That said, agentic search connected to your companies information sources is very valuable on its own. We have just connected up our internal zendesk, Jira, confluence, and github in Claude Code and it’s incredible how useful it is to find information spread across different services in 1 minute instead of it personally taking me 15 minutes of manual search.
Start Claude Code in a big repo, give it a bug report and ask it to come up with a fix, and watch it do hours of work in minutes. It doesn't have 100% success rate, but its ability to navigate code bases and understand how different parts play together has become seriously impressive
Only for well defined tasks. There's not really a practical upper bound. We will keep throwing more complex tasks at it to the extent that it can handle them. Like if you just need fancy OCR, then a specific model will probably suffice, but there will be an appetite for human- or superhuman-level intelligence that never gets tired and has no rights.
These companies are spending billions on custom datasets for a gazillion of valuable tasks and are clamping down on exfil for distillation. It's not guaranteed open source models will continue to keep pace.
China might purchase the data and train their models just to make the AI bubble pop. A few billions to throw a wrench in your competing superpower economy might be totally worth it
If they did that it would be pursuing a commodify-the-complement strategy of some kind, not just "popping" a bubble. Same as nVidia publishing their own open models. If anything the value of everything AI would rise even further due to Jevons' Paradox.
Funny I consider this valuation modest considering what the max extent of the investment thesis is here.
SaaS and legal market caps have already contracted a multiple of the combined OpenAI + Anthropic valuations just based on the _threat_ of what they may be able to accomplish.
They'll have the data + knowledge edge over open alternatives and be able to implement + deploy (see the story about Anthropic employees being at GS for 6 months already[0])
Then they will fall back on the selling their other real competitive products - hardware accelerators, phones and PCs, cloud storage and cloud compute, enterprise software, databases, operating systems, office and media suits... Oh wait...
I am struggling with this because I have an Anthropic offer vs another equivalent offer that is all cash.
But project out forwards.
- What happens when Google builds a similar model? Or even Meta, as far behind as they are? They have more than Anthropic in cash flow to pour into these models.
- What happens when OSS is "enough" for most cases? Why would anyone pay 60% margins on inference?
What is Anthropic's moat? The UX is nice, but it can be copied. And other companies will have similarly intelligent models eventually. Margins will then be a race to the bottom, and the real winners will be GPU infra.
If they outlast the competition it might be a really hard market to enter. Models are expensive to train, and they'll get outdated. You're on a time limit to make a profit off of it
Google and Meta might be the only real threats against this given how much cash they have and so far Meta is just flopping
I've been in this situation before. Anthropic has a stupid business model but the market can stay irrational longer than you can stay solvent. If you get in there you will be aligned with people who structurally do not lose.
Big picture, sure. We can talk about the millions that corporations will make and who's going to do what. But you're a person. $1 million in options is probably meaningful for you. Companies aren't IPOing, but the secret is that they're still paying employees cash for their options. SpaceX employees have had what's called a tender, which means they get to sell some of their hypothetical SpaceX options for cold hard cash in the bank that you can use to pay your mortgage. There's zero guarantee that Anthropic will do such a thing before the bubble bursts, but if they do, and you're there, who cares about a software company moat when you have enough money to buy a castle in Napa and pay to have a real actual moat with water in it and crocodiles, if that's what you want.
Others are made of different stuff, and are going to go right back to work, even though they could go off to a beach for the rest of forever somewhere.
> who cares about a software company moat when you have enough money to buy a castle in Napa and pay to have a real actual moat with water in it and crocodiles, if that's what you want.
Doesn't this require their private market valuations to go well into the trillions?
No, it’s not. This is a dangerous perspective, usually held by engineers who think that accounting doesn’t matter and don’t understand it.
You MUST accrue the lifetime value of the assets against the capital expense (R&D in this case) to determine the answer to this question.
The company (until this announcement) had raised $17B and has a $14B revenue rate with 60% operating margin.
It is only negative on margin if you assume the prior 14B (e.g. Claude 4.6 plus whatever’s unreleased) will have no value in 24 months. In that case, well, they probably wasted money training.
If you think their growth rate will continue, then you must only believe the models have a useful 9 months or so life before they are break even.
Anthropic is, according to Dario, profitable on every model <<—- they have trained if you consider them individually. You would do best to think “will this pattern continue?”
What is the lifetime value of an individual pretraining run, and what is the cost to do it? Whether it is a net positive seems to still be an open question.
Actually there is a chart of answers to this question, because the frontier providers have been delivering new models for some time. The answer is that so far they have been net positive.
Sorry - if a model costs (say) 20B to train, lasts 12 months before it becomes obsolete, generates 2B/month revenue, but with 1B/month inference costs, then it has lost 8B.
Or are you suggesting that in fact each model comes out ahead over its lifespan, and all this extra cash is needed because the next model is so much more costly to train that it is sucking up all the profits from the current, but this is ok because revenue is expected to also scale?
I'm not suggesting, I'm saying that this is what has happened so far, full stop, based on multiple public statements from people like Dario.
Basically every model trained so far has made money for Anthropic and OpenAI. Well maybe not GPT4.5 - we liked you but we barely knew thee..
The cash spend is based on two beliefs a) this profitability will continue or improve, and b) scaling is real.
Therefore, rational actors are choosing to 2-10x their bets in sequence, seeing that the market keeps paying them more money for each step increase in quality, and believing that either lift off is possible or that the benefits from the next run will translate to increased real cash returns.
What's obscure to many is that these capital investments are happening time shifted from model income. Imagine a sequence of model training / deployments that started and finished sequenced: Pay $10m, make $40m. Pay $100m, make $400m. Pay $1bn, make $4bn. Pay $10bn, (we are here; expectation is: make $40bn).
If you did one of those per year, the company charts would look like: $30m in profits, $300m in profits, $3bn in profits. And in fact, if you do some sort of product-based accrual accounting, that's what you would see.
Pop quiz, if you spend in the first month your whole training budget, and the cycles all start in November, what would the cash basis statement look like for the same business model I just mentioned?
-$10m, -$60m, -$600m, $-6bn.. This is the same company with different accounting periods.
Back in reality, shortly into year 1, it was clear (or a hopeful dream) that the next step (-100 / +400) was likely, and so the company embarked on spending that money well ahead of the end of the revenue cycle for the first model. They then did it again and again. As a result naive journalists can convince engineers "they've never made money". Actually they've made money over and over and are making more and more money, and they are choosing to throw it all at the next rev.
Is it a good idea or not to do that is a question worth debating. But it's good to have a clear picture of the finances of these companies; it helps explain why they're getting the investment.
Anthropic's marketing somehow punches hard. Not sure why, but the stuff they do sticks. Not because the products are great, but because the way they communicate about it gives people the right feeling. They do have legitimately the best coding model now for most tasks, and for narrative prose, but the marketing stuck and people stan'd them even when they were trailing.
It's Web 2.0 all over again. No moat, winner-take-all (economies-of-scale/network-effect). Just have to out-spend everyone else, and then figure out whether it was worth it all after you win.
Having a cutting edge model that requires tens of billions of dollars to train + a massive concentration of talent and experience + brand + one of, if not the best, coding experiences in Claude Code
At least from the software engineer pleb perspective, their moat is that their tools seem to work well more often than not. I wasn't comfortable with the idea of using GitHub CoPilot as our GenAI solution at work, and apparently that was a widespread feeling, because we switched to Claude Code, and it's been a relatively smooth transition from manual coding to GenAI agentic loops.
These scammers from FTX did put $500 million in Anthropic early on, for about 14% of the company. Later on this was diluted to 8%.
8% of a $380 billion valuation would be a cool 30 billion which I think would have covered the entirety of the fraud and left money for SBF and its friends.
But thankfully around June 2024, the clawback of stolen funds by FTX had its Anthropic shares sold for about $450 million.
I'm glad to know SBF and its scammers friends are going to see exactly jack fucking shit of that money.
They're being told to by management. That says approximately nothing about the relative merits of the two products, but it certainly says something. What, exactly? Who knows...
Other companies have a similar top-down "use Claude" mandate as well.
Is it nothing though? How many Apple employees do you think use Windows? And how many Anthropic employees do you think use GitHub Copilot? I would assume the answer to both is approximately 0.
How is Anthropic, OpenAI and xAi going to compete against the likes of Google that can spend $200 billion a year? It’s an impossible war and all these investors are throwing their money into a bottomless insatiable pit of money.
Until the funding stops for one reason or another and then everyone loses all their money at once like a star that collapses into a black hole singularity in a femtosecond.
As someone who thought Google+ doomed facebook, because of Gmail accounts and everyone with Google as their homepage already, I learned not to overestimate Google’s abilities.
How in the world would you have thought that? Genuinely curious.
It was obviously DOA and waaaayyy outside G'scompetence.
1. Google had recently exploited their home page to push chrome browser successfully altering the browser market. They pushed anyone visiting Google to chrome with a popup on the home page. The same opportunity was there for G+, but with updates from friends.
2. Everyone already had a Google account and many millennials were using Google Talk at the time. It appeared Google could undermine the network effects.
3. The UI of G+ appeared better
4. Facebook had released the newsfeed otherwise known as ‘stalker mode’ at the time and people recoiled at the idea of broadcasting their every action to every acquaintance. The circles idea was a way of providing both privacy and the ability to broadcast widely when needed.
5. Google had tons of money and devoted their world class highly paid genius employees to building a social network.
You can see parallels to each of these in AI now. Their pre existing index of all the world’s information, their existing search engine that you can easily pop an LLM in, the huge lead in cash, etc. They are in a great position but don’t underestimate their ability to waste it.
So is Gemini tbh. It's the only agent I've used that gets itself stuck in ridiculous loops repeating "ok. I'm done. I'm ready to commit the changes. There are no bugs. I'm done."
Google somehow manages to fumble the easiest layups. I think Anthropic et al have a real chance here.
Google's product management and discipline are absolute horsesh*t. But they have a moat and its extreme technical competence. They own their infra from the hardware (custom ASICs, their own data centers, global intranet, etc.) all the way up to the models and product platforms to deploy it in. To the extent that making LLMs work to solve real world problems is a technical problem, landing Gemini is absolutely in Google's wheelhouse.
without modern product moat, legacy products and infrastructure is useless. It's like Microsoft saying I have excel, Azure and CoPilot.
Maybe it's incentive is to 'close the ticket' as fast as possible.
Hard to bet against Hassabis + Google's resources. This is in their wheelhouse, and it's eating their search business and refactoring their cloud business. G+ seemed like a way to get more people to Google for login and tracking.
> it's eating their search business
Fact not in evidence. Google's search and advertising revenue continues to grow.
Indeed. The stupid AI on Google’s search page is so bad, I really wonder why the released it publicly.
Makes CoPilot look like something from a Sci-Fi movie.
A couple months ago things were different. Try their stronger models. Gemini recently saved me from a needle in a haystack problem with buildpacks and Linux dependencies for a 14-year-old B2B SaaS app that I was solving a major problem for, and Gemini figured out the solution quickly after I worked on it for hours with Claude Code. I know it's just one story where Gemini won, and I have really enjoyed using Claude Code, but Google is having some success with the serious effort they're putting into this fight.
I think they had no choice but to release that AI before it was ready for prime time. Their search traffic started dropping after ChatGPT came out, and they risked not looking like a serious player in AI.
They recently replaced “define: word” (or “word meaning”) results with an “ai summary” and it’s decidedly worse. It used to just give you the definition(s) and synonyms for each one. Now it gives some rambling paragraphs.
Interesting that you consider the most cutting edge technology in the category to be "the easiest layups".
I think they've been gaming benchmarks.
I use Claude every day. I cannot get Gemini to do anything useful, at all. Every time I've tried to use it, it has just failed to do what was required.
I thought it was a far superior UI to facebook when it launched. I tried to use it but the gravity of the network effect was too strong on facebook's side.
In the end I'd rather if both had failed. Although one can argue that they actually did. But that's another story.
I very much wanted Google Plus to succeed. Circles was a great idea in my opinion. Google Plus profiles could be the personal home page for the rest of us but of course, Google being Google...
That being said, tying bonuses for the whole company on the success of Google+ was too much even for me.
I very much wanted Google Wave to succeed. It seemed like a really cool way to communicate.
I guess we sort of got it with Slack though
Everything was obviously DOA after it dies. I also thought it wouldn't last but it wouldn't be the first or last tech company initiative that lived on long after people thought it would die. Weird things happen. "Obviously" isn't a good filter.
Google+ had an exclusive club appeal at launch because it wasn't instantly globally accessible, but slowly opened up instead.
It became clear they where desperate about user numbers when thay forced the merge of Youtube accounts. Or something like that.
Facebook was the same way when it started.
Theoretically Apple can spend just as much. What are the outcomes though? All those giants have their own business that are established and profitable.
It’s the new kids in the block that will make the difference.
You know those lists on twitter about how many companies US has in top 10 and are presented as a win? Those are actually lists of capital concentrations blocking innovation. It looks like US is winning but for some reason life is better in EU and innovation is faster in China.
It’s companies like OpenAI Anthropic that will move US ahead. Even if some core innovation or and capital comes from the establishment.
> Theoretically Apple can spend just as much. What are the outcomes though?
The GP was talking about Google specifically, and their outcomes on AI are nothing to scoff at. They had a rocky late start, but they seem to have gotten over that. Their models are now very much competitive with the startups. And it's not just that have more money to spend. They probably have more training data than anyone in the world, and they also have more infrastructure, more manpower, more of a global footprint than the startups.
The Innovator's Dilemma is an anecdotal, maybe a statistical relationship at best, but not a fundamental law of nature. When an established company has everything it should take to become a leader in a new industry in theory, and in practice their products are already on par with the industry leaders, you know at some point it becomes rational to think that maybe they might become a leader.
Google didn’t have a late start, they invented the tech, had bespoke hardware in place that supported it and have money to spend.
I don’t have any idea what comes next but Google and Microsoft look bad right now because they can’t execute a product strategy.
My personal bias is that either ms or Google or both will land just fine after it all shakes out but they started with a lead and are now playing catch up.
they did have a late start in terms of productionizing the models. It's definitely improved but there was a time where Gemini and the associated tools werent as good as claude/oai
Sometimes I worry about the incentives for innovation in the US.
Step 1, find something to innovate on, sell the promise of it to investors. Step 2, build a prototype or worst case, build it for real and start generating income from your truly innovate and unique product. Step 3, get acquired by a large company and then shut down because your product competed with theirs.
End result, general public possibly benefited from your innovation, but in the long run, it was temporary.
Maybe the incentives would be better if it were harder for large companies to acquire small ones? If the path to riches where driven primarily by delivering value to customers. Would love to hear other's opinions on this.
> Theoretically Apple can spend just as much. What are the outcomes though? All those giants have their own business that are established and profitable.
Ah! Well, if we put aside "The Innovator's Dilemma" and pick up Reis and Trout's "marketing Warfare," we get the answer. Apple does have an existing business, but investing in AI does not cannibalize it. They can throw money at it, try to find a way to make it work really, really well for consumers on very specific custom hardware in their devices...
Likewise, someone like Google has all the money in the world to throw at it, but they aren't investing in a new market, they're defending their search business against everyone just asking a generative AI Chatbot questions. I\But it's possible for them to screw this up internally over turf wards, just ask the engineers who tried to make search better but were kneecapped by Prabhakar Raghavan who demanded that search be poor enough to drive people to click sponsored results.
In the "Marketing Warfare" model, Apple is attempting a flanking attack: An outsider trying to disrupt the AI giants with an approach that they can't imitate without undermining their value proposition. On-device AI flanks the big giants that areservcie-centric.
And in that model, Google is playing defence, which is what every leader is supposed to do. Their job is to "cover every move," which they are doing in textbook fashion. If AI goes away, Google dry their tears and continue to mine ad revenue.
I'm very curious if they're going to ditch Google by providing on-device search. A monthly Open Crawl is under 100 terabytes, and if you clean that down to raw text and deduplicate and maybe pick out what you don't care about, the dataset might already fit onto my iPhone. They could do a lot without making a network call and reach out to a server for anything the device doesn't have, but a lot of user queries might never need to leave the phone. In another couple years, storage will be even higher.
I was hopeful for on-device AI too but any AI processing so far sucks up the battery, heats up the phone and most importantly isn't even nearly good enough. Without a breakthrough in battery, chips or the models and algorithms the way forward is thin clients that connect to some servers close to a solar farm or nuclear energy plant.
> On-device AI flanks the big giants that areservcie-centric.
Wouldn't on-device AI also support Google's position? If search is to be protected, on-device AI (small models) would be capable of basic usage, but inept at answering knowledge questions specifically, necessitating a search service be preserved. They have already launched local models in Chrome and Android. Meanwhile none of the big AI competition can profit off of local models, so this is a unique opportunity for big-G.
That said, I disagree with the premise you propose. It's 2026, and about 40% of their revenue over the last few years comes from non-search products (depending on quarter). Oh and Apple doesn't seem to be investing enough in AI products, because it's just making them look bad, not providing a "flanking attack".
Google is pulling in tons of AI revenue - from subscriptions, personal and enterprise, and Google Cloud (APIs etc). Cloud is seeing a ton of growth lately, and I'm sure that's largely from AI services that are uniquely available there. As long as they can serve models with a better cost structure (thanks TPUs) they can squeeze out better margins than their competitions.
+1 for mentioning Christensen & Trout and Reis.
> What are the outcomes though?
NVIDIA, and contractors who build data centers, and manufacturers who supply them, will all get rich.
The new kids have an easier time focusing. the big kids can integrate AI with their existing products and user data
In the long term, big kids win no? The big kids are also going to have an easier time with hardware at scale too
"but for some reason life is better in EU" citation needed
Google fucks up 90% of their products, why do you think Gemini is in the 10%?
well, it's basically existential, so the incentive is there to not only get it very right but also to limit the delta with how right anyone else gets it. The same can't really be said of the long tail of products Google have done.
Look to GCP as an example. It had to be done, with similar competitive dynamics, it was done very well.
Look to Android as another.
You must not know the stories of why GCP came to be.
It was an idea from the creators of Kubernetes and the execs at Google fought it the whole way
I hadn't heard that, that's interesting. Any sources you'd recommend to hear more about it?
I think it's a slightly different point though. What I'm saying isn't about where the idea came from or whether it was part of some precient top down bet / strategy from the very beginning.
It's more where did the strategy evolve to (and why) and did they mess it up. GCP and Android are good examples of where it at a minimum became obvious over time that these were massively important if not existential projects and Google executed incredibly well.
My point is just that there's therefore good reason to expect the same of LLMs. After all the origin story of the strategy there has a similar twist. Famously Google had been significantly involved in early LLM/transformer research, not done much with the tech, faltered as they started to integrate it, course corrected, and as of now have ended up in a very strong position.
> well, it's basically existential, so the incentive is there to not only get it very right but also to limit the delta with how right anyone else gets it. The same can't really be said of the long tail of products Google have done.
I've yet to see anything that threatens Google's ad monopoly.
The threat to Google is that browsers themselves get displaced.
I mean I guess this is classic disruption theory.
It's not that a dominant position goes away overnight. In fact that would be precisely the impetus to spur the incumbent to pivot immediately and have a much better chance of winning in the new paradigm.
It's that it, with some probability, gets eaten away slowly and the incumbent therefore cannot let go of the old paradigm, eventually losing their dominance over some period of years.
So nobody really knows how LLMs will change the search paradigm and the ads business models downstream of that, we're seeing that worked out in real time right now, but it's definitely high enough probability that Google see it and (crucially) have the shareholder mandate to act on it.
That's the existential threat and they're navigating it pretty well so far. The strategy seems balanced, measured, and correct. As the situation evolves I think they have every chance of actually not being disrupted should it come to that.
Because Google has the money to build 10 different versions/iterations of Gemini and can essentially force one to work. They have most people's data and most people use them for mail/search/browser/maps as well.
In my opinion though this is a race to the bottom rather than a winner takes all situation so I don't think anyone is coming out ahead once the dust settles.
Google built ten different chat products, how did that go?
Does it matter? Microsoft won by default with Teams because it actually turns out no one cares about chat or even has a choice in it: employees use whatever the company picks.
No one uses Teams for personal use. LLMs are used daily for personal use by hundreds of millions of people at this point.
It's bundled with office and no serious business can live without excel.
The world, other than the US, runs on WhatsApp. Business, support and payments are done there. So people do care.
If you're going to say "other than the US" then you've got to say at a minimum "other than the US and China", but really "other than the US and China and Japan and Korea and Taiwan and Thailand and Russia and most of Central Asia".
Only mentioning the US is wildly americentric even by HN standards.
Gosh doesn't that sound familiar.
This was the same argument made for Google Wave and Google+ and both completely tanked
The tech behind wave eventually made its way into Google docs though and pioneered collaborative document editing, so wasn't a complete failure even though the product itself was killed.
No comment on Google+, Google has a storied history of failure on any kind of social media/chat type products.
Where Google wins is just simply having enough money to outlive anyone else. As the saying goes "the market can remain irrational longer than you can remain solvent" In this case, Google is the market and they can just keep throwing money at the wall until OpenAI, Anthropic, etc. go under.
Google Docs has no features remotely like what Google Wave was.
And there was collaborative editing long before Google Wave.
Social media has strong network effects that keeps competitors at bay. What network effects are OpenAI/Anthropic/etc accumulating?
Yes, but Gemini is actually good and so are their APIs.
Agree. Look at how miserably MSFT has failed at integrating AI tastefully in their business.
Google makes money selling ads. Nothing else matters.
They target those ads by ingesting as many signals as possible from as many input devices & sensors as they can possibly convince people to use. They make a lot of money from advertising b/c they have managed to convince the most number of people to give them as many behavioral signals as possible & they will continue to do so. They kill products only when the signal is not valuable enough to improve their advertising business but that's clearly not the case w/ AI.
Google has barely released a successful product in 20 years.
Depend on the definition of the "product". For example some banal cloud storage in which everyone competes. And it's an "old" product, despite being invisibly improved behind the scenes, just like at any other provider. Google has pretty competitive storage AND they are fully abusing Android integration for AND they have pretty good bundling of that storage with other products, including, you've guessed it - LLM Gemini. So say a person is not a professional user of LLMs like a developer burning tokens in a dozen accounts simultaneously. A person has a phone and eventually memory runs out, so he buys a one click Google storage for 4 bucks. And suddenly he has Gemini Pro included too. So why pay 20 bucks to Anthropic, when Google costs 1/5 of that AND has other stuff bundled too?
So maybe Google is lagging on truly new products (btw, does Gemini itself with its TPUs count as a new product? I'd say yes), but "old" products are entrenched enough to carry them and compete.
Google Drive is easily the worst of the desktop cloud storage options. It’s okay for Google Docs but not other files if that’s what you’re talking about..
In a world where OneDrive exists?
I get 2TB (which I use) and AI Studio for $20, that's the best deal out there for me.
Which one would you say is the best?
Google is good at buying existing products and scaling them, which is exactly what they did with DeepMind.
Deepmind was their worst acquisition ever. It is a vanity project that burns cash.
Google Cloud is good and successful. Except they can't implement billing hard caps, or pretend they can't.
Im not sure what you consider successful. They've been struggling to get market share vs azure, and the product isnt that good. lots of rough edges, and piss poor support
Neither does AWS and you can argue they aren't good but they're objectively successful, so it doesn't seem like a good metric.
Their API business model seems to be hope enough people accidentally go over free tier: $0 for the first 5000 monthly places lookups, $40 per 1000 after that
I thought that the likes of Android, Google Docs, Google Translate, etc. were fairly successful. Chrome and ChromeOS also seem fairly popular too.
A lot of those are getting pretty close to 20 years ago.
This year:
chromeos is 17
android is 18
chrome is 18
google docs is 20
google translate is 20
In retrospect, it is wild how good/successful google was 17-20 years ago!
Few years ago, we had Google Bard, the ancestor of Gemini, which was supposed to be an AI LLM, and when you right-clicked the page, it was a fake page with hardcoded sentences in a .js file...
None of them looks like their original form, none would survive without google’s enormous investment.
Because the product quality doesn't matter if the competition isn't making any money.
google the only ai which invests mixing llm ai with real ai, and it seems work well.
race to the bottom. google in house cheaper inference hardware. anthropic buys it.
Persistence. Google has a lot more endurance then OpenAI does in this game.
The current AI market is going to destroy anyone who's specialized into it compared to having alternative revenue streams to subsidize it.
Does Alphabet/Google have any other significant alternative revenue streams though besides their ad revenue? And won't that decrease significantly the more people use AI tools for research than firing up a google web search? I find myself using Claude more and more doing web research and comparing products/reviews...without getting a single ad served up from Google.
Do they though?
Google does things I hate with their products. But the money printing machine keeps going whrrr faster and faster.
The conclusion Google is engaged in consumer capitalism is wild.
They're engaged in computing research and merely engage in consumer capitalism as a consequence of political and social constraints.
Products are a means to an end not the goal.
OpenAI and Anthropic are product companies and are more likely to fail like most product companies do as they will lack broad and wide depth.
Google has experience in design, implementation, and 24/7 ops with every type of SaaS there is. They can bin LLMs tomorrow and still make bank. Same cannot be said for OAI or Anthropic.
I'm trying paid tier Gemini and it doesn't allow to keep have personal chat history when you disable training on your data, on reload of the page your chat is gone. Even free tier of ChatGPT allows disabling training on your data while allowing to keep such basic functionality.
Some technical advancements are not worth it if you do not respect your users.
Yeah I’m never using a Google product. The sole purpose of their company is to be evil. At least other companies are indifferent.
Google is evil in passive way, like sprawling bureaucracy making you life slowly worse and worse but also doing some stuff to at least some fraction of population. OpenAI and Sam are determined and energetic evil, laser focused on making whole human population jobless and homeless in shortest way possible and not producing anything else of value, no other products. I'd rather prefer the former evil out of the two.
How does their top tier subscription compare in usage limits to the $200/mo Claude usage limits?
Claude has easily the worst and most opaque usage limits in the industry.
Another basic feature that’s missing is sharing a Gemini chat as a link anyone can view.
OpenAI figured this out: it’s awesome marketing when people send each other links to the app with a convenient text box to continue the conversation. It’s viral.
Google meanwhile set this up so that “anyone with the link can view” is actually “anyone with the link and a Google account”.
That’s grade A failure of marketing.
The PM in charge of that decision ought to be walked off a plank.
This doesn't sound right. I just opened a shared link in a fresh incognito window and it works fine.
Try sharing from the Google AI Studio Playground.
E.g.: https://aistudio.google.com/app/prompts?state=%7B%22ids%22:%...
OK. I think you moved the goalposts a fair distance there.
Because it's Google they can't build products and they only care about benchmarking.
The product they released so far are all half assed experiments.
Gemini 3 Pro is now being beaten by open source models because they can't fix or don't want to fix the problems with the Gemini models being completely useless.
The same for Microsoft.
Microsoft had GitHub Copilot, and Microsoft Copilot and both of them are useless to Claude Code and Claude Cowork.
You can have all the money in the world, but nothing is stopping you from building useless garbage.
Claude is clearly the most superior product right now.
Gemini is absurdly expensive for low quality (3000 USD of tokens are not even worth what you get @ Anthropic for 200 USD).
The same can be said about Claude (no or tiny Opus on Pro) vs GPT-5.2 high (5.3-codex if you like terminal bench hacking).
You mean GPT-5.3-Codex is a much better value than Claude Opus for programming ? If yes then I'm very interested as I am using Claude there
I guess it depends on your spending. GPT-5.2 and -5.3-Codex are certainly much cheaper: you get much more from the same $20 sub. When I was using Claude as primary I would daily hit limits and have to wait vs on GPT with more usage it only happened to me once in a few months when I was vibecoding non-stop for a week or two to port my personal Windows tools to Linux with multiple other projects being worked on in parallel.
Anecdotally GPT was also smarter than Claude which prompted my move from Claude in the first place: Gemini and Claude back in October failed to get their own harness PID.
Outside of anecdata I rely on https://artificialanalysis.ai/models/capabilities/coding for now.
I really like github copilot.
I also tried open code cli and desktop, but how well copilot is integrated into the ide is a plus for me.
What makes them "useless garbage"?
Work smarter, not harder: find an innovation that makes LLM computation an order of magnitude more efficient. Easier said than done, of course, but vastly more efficient biological systems provide an existence proof that this is possible in principle. DeepSeek already showed how it's possible to get good results with limited resources.
Well there's a good reason that OpenAI partnered up with Microsoft. The calculation is that the established big techs - Amazon, Apple, Microsoft, Google, Meta are all going to be significantly impacted by AI so it's not unreasonable to look at Anthropic at 10% of their market cap as a reasonable value. Would it be worth Apple to bring Anthropic in house? They failed to deliver AI themselves, they know the risks of being dependent on Google. If AI goes far enough it may totally remove Apple's differentiation.
Some of the Big Techs are building their own in house stuff (Meta, Google), but it wouldn't be crazy to see acquisitions by the others, especially if the market cools slightly. And then there's the possibility that these companies mature their revenue streams enough to start actually really throwing off money and paying off the investment.
> they know the risks of being dependent on Google
I wouldn't argue it's that risky. Look at their past entanglements:
1. Google Default Search Bribe - brings in $20B a year for literally doing nothing
2. Google Maps: Google let them build their own custom app using Google's backend, and it worked fine all the way up until Apple chose to exit that arrangement
actually I can't think of any others, but is there an example of Apple getting burned by Google?
Well they have an illegal monopoly over display ads that they defended with a moat of money when FB tried to butt in, so that's an example of them being not great to rely on.
Android.
Slight counter point - claude code is basically the only developer tool that ever been happy to pay money for. Getting the entire software industry to give you $200/mo/person is quite the market.
That just means they're basically Adobe. What's their valuation?
> Getting the entire software industry to give you $200/mo/person is quite the market.
Quite the fantasy, you mean.
These companies are operating outside of "normal" economics. They seem to be fed money against all rational valuations because it is a Manhattan project - the US needs to 'beat' the Chinese with this and as such these companies will probably not be allowed to fold.
it's a race of parallel discoveries sprinting towards commoditization and indistinguishably
any real breakthrough will be instantaneously reverse engineered and replicated
none of the not-googles can win, because there is no win state
We're starting to see the nuance come out in what these models are working towards.
* OpenAI - chat that has some character to it.
* Claude - working through thoughts and coding
* Gemini - general reasoning (still blown away by gemini's reasoning, but cannot understand it's inability to tool call - maybe that's been fixed)
> It’s an impossible war and all these investors are throwing their money into a bottomless insatiable pit of money.
Anthropic went from zero to $14 billion in revenue in less than 3 years, growing at 10x per year.
That's what they're investing in.
Also Anthropic seems laser-focused, unlike some of their competitors who are throwing stuff against the wall to see what sticks.
Revenue, but what about profit? Google can be cash positive but I’m not sure Anthropic can be the same.
At least xAI now has a revenue generation backer. SpaceX.
Others must pull up their revenue number.
SpaceX makes 16B in revenue per year, with 7B in ebitda (which doesn't account for the cost of rockets)... so assume what, 3B in free cash flow per year? And that's being generous.
That's about what Google creates in free cash every 2 weeks.
SpaceX can also raise their prices for government launches to pretty much anything and still get business, because they are essentially a monopoly.
They have the best coding agent around. I would say roughly half the industry no longer writes code by hand and claude is/was the only agent that let me drop my editor all together. That enabler alone is enough signal beyond hype.
Frankly google models and the UIs google designs around their models just aren't as powerful and more training, more data and more compute is no longer tipping the scale. Anthropic did something to make their model better at coding than almost anything else.
And all of this is just what's happening right now. The money being invested is a gamble not on right now... but on the trendline to the future. I agree that the LLMs are overloaded with hype, but the people who compare it to crypto aren't thinking straight. Whether there's over investment or not a paradigm has shifted. Maybe there will be a collapse, but it won't collapse into a singularity. If it collapses at all, a new world will emerge, and that new world will generate more value than all the money currently invested in AI.
Have you ever used anything that is on google cloud console? Or tried not to get randomly ratelimited with a single request to a vertex llm model? They are shooting themselves in the foot for solid 20 years, any of these players can compete with google in this frontier
I had to install the separate gcloud and gsutil utilities and use one to synthesize a login session for the other this week.
Took fully 10 minutes to install from homebrew.
I do not believe in this company.
Google is playing the datacenter game differently because they have their own hardware.
how does any startup beat an incumbent?
I think GP is probably implying that this particular vertical requires obscene amounts of capital to keep up, which makes it really hard for a startup if you’re going up against businesses with giant free cash flow machines.
It’s the same reason Reid Hoffman sold his AI startup early… he realized he just couldn’t beat Google/FB/MSFT long term if it devolved into a money race.
Until a version of Gemini is released that can fix one (1) bug in my codebase, I'm not worried for any of those companies.
Google's only focus isn't on Gemini. Anthropic is do-or-die
Also Theranos was do-or-die and we know how it ended.
I fail to see what that has to do with this?
Just because something is in do-or-die situation doesn't mean that they have some kind of magical advantage over fat cat. Being in that situation means there is very real possibility of doing "die" part and we have lots of examples of them doing so.
Theranos didn't get out-competed by the fat cat. They defrauded their investors, got caught, the founder went to jail and the company died.
very different; not a relevant comparison
Look at Sundar's most recent remarks and tell me Google's isn't only focusing on AI: https://blog.google/company-news/inside-google/message-ceo/a...
Basically "we have youtube subscribers" is the only thing that isn't all about AI, but even that i'm sure they're trying to figure out how to shoehorn AI into that product
They're frenemies. The likes of Google also host Anthropic and OpenAI.
The most efficient way Google could spend that money is probably to buy a company and not poke at it too much. I have no confidence that large rich companies can actually innovate beyond buying small innovators or spawning business units and not poking at them much.
Google has invested in Anthropic. I don't trust that Google will compete on fair grounds with Anthropic on coding. Their common enemy is OpenAI.
The same way that Google+ never overtook Facebook
Given the amount money that they are spending for vastly subpar products maybe they need to quadruple their capex
How long is Google going to be able to keep selling search engine ads?
I’d guess they want to outlast OpenAI and then get bought by Apple or Amazon.
Google will buy Anthropic if it comes to it. Google already owns ~30% of anthropic and Anthropic is running on Google hardware.
Google also has the most to lose.
just want google to have good web apps again, it's so bad on desktop
Google is invested in Anthropic
Culture.
I mean, you gotta spend the 200B on the right things.
Let's be real.
Google leadership is pathetic.
Sundar "the manager" has presided over an enormous growth of the businesses he was handed. He also presided over the complete collapse of the internal culture. OTOH he may have fired Dianne Green, so that's something. Overall, at best Meh.
Demis ran a startup that burnt cash on vanity projects and continues to burn cash on vanity projects. Gemini is barely open source quality AI, but Google makes it nearly free and has the best distribution on the planet.
Gemini has been a joke since 1.0. No release has hurt Google's brand more. 3.0 was STOA for about 2 days, easily Gemini's best release.
Anthropic and OAI are moving at amazing pace, Google can not keep up at all.
They made a C compiler. Google has no C compiler. Win. /s
> "It has been less than three years since Anthropic earned its first dollar in revenue. Today, our run-rate revenue is $14 billion, with this figure growing over 10x annually in each of those past three years."
Wild although not entirely surprising. Congrats, Anthropic.
Next year 140 billion the following year 1.4 trillion, 14 trillion the year after that?
Better be otherwise it’s a 27+ year wait for breakeven!
27 year PE ratio is a value stock at this point
Don't confuse revenue with earnings.
Shit, explain Reddit with a P/E much higher than nVidia.
Makes pretty much no money, has no real opportunity to make money, only a small segment of fanatics actually like it… and yet… infinite stock price.
I’ll give you the point they aren’t actively and intentionally losing money. My point is that everything is fake and… lame.
At $14b revenue Anthropic is likely running a HUGE loss on their P&L.
Pay attention to the outflow of tech investment in the stock market. That money is going to move into OpenAI and Anthropic IPOs. The valuations will be as big you are thinking because the market believes these companies will represent an entire basket of startups.
It Is more likely that people are cashing out very liquid assets (tech stocks) to pay back their loans in Yen as interest rates are rising over there.
Tech stocks with all the hype are second only to crypto in terms of how easy and fast are to sell (hence BTC dropped and now tech stocks IMHO).
Btw, I was too young to fully remember, but wasn't the year before the dot com crash also full of IPOs?
Apparently the last two times the Super Bowl Ads were dominated by Tech companies was 2000 for dotCom and 2022 for Crypto.
That's a really interesting tidbit. Thanks for sharing.
And thinking about it it makes sense since the decision to pay the outrageous rates for an ad during the Superbowl must be driven by strong emotions (confidence or desperation). In this case, considering there's no clear moat for any of the big players, I believe it's the latter.
If your thesis was correct, why wouldn't some of those "outflows" go to GOOG or NVDA?
They would. You can see how resilient GOOG has been during this recent draw-down, and how much growth it has had even as AI sells off.
AI sells off… if this is a selloff than I see what everyone is talking about when they are saying we in a bubble :)
And why would anyone participate in their IPOs? They would be crazily overvalued, like Figma or worse.
I really wonder is there even enough dump money from them to sell the stock they hold. Not to mention even raising any new capital... Is there really enough bag holders that will run after these stock with large enough piles of money?
To be fair, Facebook was at the time viewed by many as crazily overvalued.
Might as well long NVDA?
There are many bitter lessons ...
Could you be more specific? Because NVDA has a consistent 20 year growth of something like 400x and +30%/yr, so I don't think the bitter lessons are there.
I would hold off congratulating them until they’re actually in the black. They are still burning billions a year lol the revenue is impressive but their expenses are still solidly north of it.
Don't worry about it: they'll make it up in volume
https://www.youtube.com/watch?v=CXDxNCzUspM
What are the benefits ?
If you give me $1T to spend, I, too, can probably make $14B (this is a metaphor)
How is this not a red flag? It’s a series G and they are still begging for money to burn. When do they convert their balance sheet to profit? Is it after a series AAF when they are worth more than Apple or nvidia?
Have you seen their revenue growth?
Those growth stats for Claude Code are pretty wild:
> Claude Code was made available to the general public in May 2025. Today, Claude Code’s run-rate revenue has grown to over $2.5 billion; this figure has more than doubled since the beginning of 2026. The number of weekly active Claude Code users has also doubled since January 1 [six weeks ago].
Doubling both annual run-rate revenue and weekly active users in the first six weeks of this year!
No wonder. At an early stage startup and every single person here/we talk to has the $200 CC.
How do you even rationally value that growth??
No doubt it's astounding, and I have no experience underwriting these kinds of scale of investments...
...but if I were a recent employee diluted as part of this raise (even with the massive uplift in revenue) I'd be very skeptical about an sort of financial outcome for myself.
Kind of amusing that there is basically no mention of their original mission at all here.
What was their original mission?
My sense is that startup mission statements are ~meaningless. Builders try to build great things that lots of other people will find valuable.
>What was their original mission?
Beat OpenAI. The Founders came from OpenAI so there was obviously some disagreement about the direction there or they simply wanted more control.
Friendly AI
To maximise shareholder value.
It's a benefit corp
Technically speaking, it's to maximize shareholder value while serving the public interest. They're a public benefit corporation.
OpenAI has "open" in their name but is closed off to public access and input
Google used to have a motto "don't be evil"
Who enforces the definition of language? Who demands compliance?
Soon as we go down the path of policing and insistence on one true dogma, we veer into religious holy war type behavior.
Obsession with semantics of syntax is a sort of theism even if the syntax and semantics do not refer to the commonly accepted tropes of a specific religion.
In this case, it has to do with how they're classified under Delaware law as a corporate entity.
I'm not a lawyer (I don't even play one on TV, damn you Odenkirk) so I can't tell you what that means as far as case law for companies getting punished for behaving badly, but in this case, there is supposedly some sort of legal backing for the classification.
Laws are words with zero meaning if they go unenforced.
Politicians are not interested in assuring such.
Public is busy arguing semantics online; they are not interested in assuring such.
In other words, you didn’t know they’re a PBC, and you don’t know what that means?
I know PBC; legal semantics that dictate various financial terms. I also know the specifics because for some reason those specifics matter to some people, and it's been personally beneficial to have more than an awareness of such legalese. Makes true believers feel I am living up to their "one of us!" means testing.
I accept they matter to others but reject such exists as anything but contemporary ethno-objects. Similar to how I acknowledge Christianity exists but am not a disciple of the dogma; I have to accept others believe are but do not have to pledge allegiance myself.
Aside from engineering and healthcare, machine operation, with real safety implications, everything else is all just parroting and social role-play.
Being a VHS copy-paste of generational semantics isn't the flex you want it to be. Patronize harder though.
They've been very clear about their mission, they're doing more than anyone else when it comes to it, and if you've ever interviewed with them you'd know how critical it is to them.
But I guess it's easier to make a glib comment than look these things up.
Anthropic has one of the best moats of any business that's been created in the last 50 years.
Numerous companies have tried and failed competing with SoTA foundational models. If Anthropic had no moat, Apple and Meta wouldn't be paying them billions for coding asistance.
Meta, Amazon, Apple, and Nvidia would all have SoTA competitors to Claude. They all tried and have not produced a competitor.
Instead you have three companies that stand alone making billions from foundational models.
The open models are not far behind. Is it really a "moat" when it's so short lived and you need a brand new moat after six months? That's just ordinary competition.
They're the least incompetent in the space.
Big companies are handcuffed by Innovators Dilemna etc.
$14B revenue run rate is the interesting number here.
Yeah, up from $1B a year ago.
Two years ago, I considered investing in Anthropic when they had a valuation of around $18B and messed up by chickening out (it was available on some of the private investor platforms). Up 20x since then ...
It was always obvious that Anthropic's focus on business/API usage had potential to scale faster than OpenAI's focus on ChatGPT, but the real kicker has been Claude Code (released a year ago).
It'd be interesting to know how Anthropic's revenue splits between Claude Code, or coding in general, other API usage, and chat (which I assume is small).
Eh, I think you made the best decision you could given the info you had.
I’ve poked around on EquityZen and was shocked at how little information is available to investors. In some cases I did not even see pitch decks, let alone one of the first companies I looked at had its top Google result: CEO recently arrested for fraud and business is almost worthless now.
Unless you are willing to take a blind punt or have insider information, those platforms are opaque minefields and I don’t fault you for not investing.
Matt Levine has a fun investment test: when presented with an opportunity, you should always ask, “and why are you offering it to me?”
Meaning, by the time it gets offered to retail investors (even accredited ones are retail) we’re getting the scraps that no one else wants.
what are the private investor platform you mentioned ? and what are the requirements to join in?
Hiive and Forge Global are the ones I know of. You must be an "accredited investor" which means nothing at all except that you have a million dollars or make $200k/yr.
Like you can buy shares of Anthropic as long as you prove you make over 200K? That easy? Shouldn't they approve of the purchase? Sorry, noob in this space!
They have to approve and it's not as simple - it's just that if you make $200k a year or have $1m in the bank, the government assumes you're a knowledgeable investor and allows you to bypass certain protections.
If you are NOT knowledgeable and simply have money ... well it'll soon be parted.
The secondary platform verifies you and then you indicate interest. If there’s a seller you may get to buy. Company may ROFR. Priority goes to bigger buyers.
EquityBee got me investors to exercise my Brex options, in exchange for giving up some small beta
I don't remember which one I was looking at - they all have different requirements.
Soon we will lack letters for funding rounds!
G is tame. Wait until you hear of Databricks’ Series K…
https://www.thesaasnews.com/news/databricks-raises-1b-series...
A few years back, well ok maybe almost ten now, but regardless- a recruiter reached out to me about a role at a "series G" company like it was a selling point, and I was just kinda like ok maybe thats signaling its relatively stable and can raise money, but at the same time, that's a lot of rounds to have preferences ensure unprivileged shareholders get nothing, and also to have most of the hockey stick growth already tapped out.
This was in the middle of the boom when companies were fighting over talent, so I found it odd.
Emojis would be far more appropriate for AI startups
They could stop at F and treat it as hexadecimal by adding digits: Series 4F, etc
considering their series F was only ~5 months ago this doesn't seem too far-fetched!
I wonder how good it is for companies to be allowed to grow so big and still be private? Would it makes sense to require any company with more than a billion dollar valuation to be subject to all the same SEC requirements that public companies are? Could companies be blocked from raising money once the reach a crazy valuation like $1 billion?
That doesn’t make sense at all. The raison d’être of SEC is to protect regular mom-and-pop investors. A private company just doesn’t allow anyone to invest in them. Why should SEC rules apply? On what legal basis can you force a private company to divulge its financial details? Would you be happy if you, as an individual, have to divulge your account statements if your own net wealth reaches one million?
It's a corporation, it exists at society's leisure.
It might be necessary to create a legal basis, but it's just a matter of doing it. If the owners don't like it they can dissolve it.
> The raison d’être of SEC is to protect regular mom-and-pop investors
That's not the sole reason. They (should) also enforce a fair even playing field by preventing market manipulation (e.g. how Elon was tweeting about stock prices) and a few other things to "facilitate efficient markets and the formation of capital".
> Why should SEC rules apply?
Because private companies still fall under the jurisdiction of the SEC? See e.g. Theranos.
> On what legal basis can you force a private company to divulge its financial details?
On the to-be-created legal basis that aims to prevent bubble formation and the resulting fallout to the wider society?
> Would you be happy if you, as an individual, have to divulge your account statements if your own net wealth reaches one million?
Sure, why not? It's not a totally unheard of idea. In Norway everyones salary can be looked up.
yeah it's a slippery slope forcing companies to go public at X valuation. who decides that? what number makes sense? etc. but i do think we need to somehow fix massively overpriced companies going public and dumping on retail
Nobody said anything about forcing them to go public, just to force them to adhere to reporting regulations.
It’s a major issue in VC. Main Street doesn’t get access until it’s time to offload. Prevents capital recycling for early stage as well.
there used to be rules that companies must be public if they have other 1000 investors. is this no longer the case?
it's still the case, but there are never 1,000 investors, there's a couple dozen VC firms, SPVs, and individuals
if you're smart.
I don't think this is an SEC problem, they are fully aware that people subject to their jurisdiction can jump through many hoops to circumvent them. This shows consent on the investor's part well enough, and capital formation regulations do not burden the investor at all, they are only constitutional because they burden the organization raising capital, who simply needs to do a cursory check - not an in-depth one. (level of depth is based on which regulatory exemption is chosen)
So as long as you separate concerns the SEC is satisfied.
If employees get stock options and decide to exercise on exit, they count against the 500 unaccredited investor limit that would trigger reporting requirements. So companies that issue stock options do have an outside risk that enough employees will exit, exercise their stock options, and trigger a reporting requirement.
Oh dang, no wonder they’re auto coding so much garbage in public (crap c compiler, crap browser, crap salesforce).
The timing of the Claude Code guerilla marketing campaign that seems to have started around new years is now making much more sense.
It's wild watching people fall for it
I did a search for a nations GDP to compare that to. That’s Chile, I think.
Goldman Sachs recently stoked fear about software stocks due to claimed AI competition.
What if their strategy is this: slowly drive down software stocks, keep talking about AI, buy the downward market. Then cash in on the IPOs of OpenAI and Anthropic.
Then let OpenAI and Anthropic implode. Goldman Sachs had no problems underwriting webvan at the end of 1999, which then imploded in 2000.
Anyway, I just valued my dog at $1 billion post-money. You can buy it at pets.com.
Matt Levine has put this forward in his newsletter - if you're moderately influential you can go on TV and tell people that "X industry will be dead in 10 years" because of AI and then profit from the inevitable stock dip.
Because we live in the worst possible timeline the end result for AI companies does seem to be "too big to fail", where these massive investments will get foisted on working class people via a bailout or an IPO and index inclusion.
I don't see why any AI company would ever be "too big to fail". I can't see why any government would be motivated to take the political hit of bailing them out.
What if the president of the nation happens to hold stock in these companies?
His newsletter (and podcast) are fantastic.
I doubt talking heads on the TV can move markets.
what goes up, must come down
spinning wheel, round and round
https://youtu.be/zhnEjxsjjuA
Claude Code is trivially an attempt to hobble the rest of the software business: the PID controller, the control vectors, the ever change loss surface, the bash and JSON jank at the foundations, the no one is this stupid context management, the some-data-critical-to proceed | tail -n 5, the sed editing, the speculative execution of partial frames.
OpenRouter and Opencode show you how behind it is, that bootstraps you off of them. They have issues too and Zen is starting to feel icky, but they let you speed run to the next thing.
You're attributing way too much intent to what is the viewpoint of some random analyst at Goldman Sachs (who doesn't even control any purse strings). A year ago there was another big hullabaloo when a GS team wrote a long post about how AI companies would never make enough revenue.
Is G for Grand Theft?
Is the G for Grand Theft?
Annoyed parent voice: What happened to the $13 billion I gave you 4 months ago?
check your credit card statement Dad
> The number of customers spending over $100,000 annually on Claude (as represented by run-rate revenue) has grown 7x in the past year.
Looks like major uptake from businesses. But all these articles keep saying there isn’t any actual value creation?
When will we see the first $1T valuation for a private company? What do you call a herd of 1000 unicorns together?
> What do you call a herd of 1000 unicorns together?
As millipede, clearly therefore millicorn.
The collective noun for a group of unicorns in AI is known as a hallucination, as in: a hallucination of unicorns.
We already have: https://www.theguardian.com/science/2026/feb/02/elon-musk-sp...
Great, they can pay me the $60k they owe me for pirating my books...
No, no didn't you hear? If they were to actually play by the rules, then there wouldn't be an industry! /$
Is everyone competing to steal Google's ad cash-cow? This is the only way these investments make sense.
I think the idea is to reduce labor costs by replacing the human workers.
I haven't used it to replace workers though, only to replace Google search. My company is pushing copilot but it's only $16/user/mo. Hardly lucrative and no moat.
Claude Code and Cowork are incredible products, and can do much more than just search. Lots of people are paying hundreds of dollars a month for them.
If you’re just using AI for search then I can see why you’d not see the value. But many people really are getting a huge amount of value out of agents, and are already paying for it.
That said, agentic search connected to your companies information sources is very valuable on its own. We have just connected up our internal zendesk, Jira, confluence, and github in Claude Code and it’s incredible how useful it is to find information spread across different services in 1 minute instead of it personally taking me 15 minutes of manual search.
Start Claude Code in a big repo, give it a bug report and ask it to come up with a fix, and watch it do hours of work in minutes. It doesn't have 100% success rate, but its ability to navigate code bases and understand how different parts play together has become seriously impressive
OTOH my company spends well into the $hundreds/user/day on Claude.
They are insane
up to 5 paid max CC accounts per dev currently, can request (and easily get approved) for more if needed
How are they not overvalued? At some point OSS will be sufficient for most businesses, what then?
>OSS will be sufficient for most businesses
Only for well defined tasks. There's not really a practical upper bound. We will keep throwing more complex tasks at it to the extent that it can handle them. Like if you just need fancy OCR, then a specific model will probably suffice, but there will be an appetite for human- or superhuman-level intelligence that never gets tired and has no rights.
K2.5 run on your own gear is an Orca tube away from 4.5, R4 will be stronger.
The gap is growing and the second derivative is positive. They don't catch up until it saturates.
They’ll become commodity AI compute providers while training and selling premium foundation models.
These companies are spending billions on custom datasets for a gazillion of valuable tasks and are clamping down on exfil for distillation. It's not guaranteed open source models will continue to keep pace.
China might purchase the data and train their models just to make the AI bubble pop. A few billions to throw a wrench in your competing superpower economy might be totally worth it
If they did that it would be pursuing a commodify-the-complement strategy of some kind, not just "popping" a bubble. Same as nVidia publishing their own open models. If anything the value of everything AI would rise even further due to Jevons' Paradox.
And yet open models have been tailing closer lately?
Funny I consider this valuation modest considering what the max extent of the investment thesis is here.
SaaS and legal market caps have already contracted a multiple of the combined OpenAI + Anthropic valuations just based on the _threat_ of what they may be able to accomplish.
They'll have the data + knowledge edge over open alternatives and be able to implement + deploy (see the story about Anthropic employees being at GS for 6 months already[0])
[0] https://www.cnbc.com/2026/02/06/anthropic-goldman-sachs-ai-m...
What’s your max extent? I was just doing some napkin math to think about where they’ll cap out.
Then they will fall back on the selling their other real competitive products - hardware accelerators, phones and PCs, cloud storage and cloud compute, enterprise software, databases, operating systems, office and media suits... Oh wait...
What do you value a company at that has gotten to $14b in revenue in 3 years and has 60%+ margin on inference? Just out of curiosity.
60%+ margin on inference: source ?
+ r&d costs
Of course, if one does not "pay" for investment, benefits are easily made ..
I am struggling with this because I have an Anthropic offer vs another equivalent offer that is all cash.
But project out forwards.
- What happens when Google builds a similar model? Or even Meta, as far behind as they are? They have more than Anthropic in cash flow to pour into these models.
- What happens when OSS is "enough" for most cases? Why would anyone pay 60% margins on inference?
What is Anthropic's moat? The UX is nice, but it can be copied. And other companies will have similarly intelligent models eventually. Margins will then be a race to the bottom, and the real winners will be GPU infra.
If they outlast the competition it might be a really hard market to enter. Models are expensive to train, and they'll get outdated. You're on a time limit to make a profit off of it
Google and Meta might be the only real threats against this given how much cash they have and so far Meta is just flopping
If you have an offer, you can and should ask this question of whomever you're coordinating with. They will give you an honest answer.
I've been in this situation before. Anthropic has a stupid business model but the market can stay irrational longer than you can stay solvent. If you get in there you will be aligned with people who structurally do not lose.
Big picture, sure. We can talk about the millions that corporations will make and who's going to do what. But you're a person. $1 million in options is probably meaningful for you. Companies aren't IPOing, but the secret is that they're still paying employees cash for their options. SpaceX employees have had what's called a tender, which means they get to sell some of their hypothetical SpaceX options for cold hard cash in the bank that you can use to pay your mortgage. There's zero guarantee that Anthropic will do such a thing before the bubble bursts, but if they do, and you're there, who cares about a software company moat when you have enough money to buy a castle in Napa and pay to have a real actual moat with water in it and crocodiles, if that's what you want.
Others are made of different stuff, and are going to go right back to work, even though they could go off to a beach for the rest of forever somewhere.
> who cares about a software company moat when you have enough money to buy a castle in Napa and pay to have a real actual moat with water in it and crocodiles, if that's what you want.
Doesn't this require their private market valuations to go well into the trillions?
It would have to be a small castle.
Is their overall margin also about 60% too? Or something saner like 30%?
Their overall margin is negative.
No, it’s not. This is a dangerous perspective, usually held by engineers who think that accounting doesn’t matter and don’t understand it.
You MUST accrue the lifetime value of the assets against the capital expense (R&D in this case) to determine the answer to this question.
The company (until this announcement) had raised $17B and has a $14B revenue rate with 60% operating margin.
It is only negative on margin if you assume the prior 14B (e.g. Claude 4.6 plus whatever’s unreleased) will have no value in 24 months. In that case, well, they probably wasted money training.
If you think their growth rate will continue, then you must only believe the models have a useful 9 months or so life before they are break even.
Anthropic is, according to Dario, profitable on every model <<—- they have trained if you consider them individually. You would do best to think “will this pattern continue?”
What is the lifetime value of an individual pretraining run, and what is the cost to do it? Whether it is a net positive seems to still be an open question.
Actually there is a chart of answers to this question, because the frontier providers have been delivering new models for some time. The answer is that so far they have been net positive.
Sorry - if a model costs (say) 20B to train, lasts 12 months before it becomes obsolete, generates 2B/month revenue, but with 1B/month inference costs, then it has lost 8B.
Or are you suggesting that in fact each model comes out ahead over its lifespan, and all this extra cash is needed because the next model is so much more costly to train that it is sucking up all the profits from the current, but this is ok because revenue is expected to also scale?
I'm not suggesting, I'm saying that this is what has happened so far, full stop, based on multiple public statements from people like Dario.
Basically every model trained so far has made money for Anthropic and OpenAI. Well maybe not GPT4.5 - we liked you but we barely knew thee..
The cash spend is based on two beliefs a) this profitability will continue or improve, and b) scaling is real.
Therefore, rational actors are choosing to 2-10x their bets in sequence, seeing that the market keeps paying them more money for each step increase in quality, and believing that either lift off is possible or that the benefits from the next run will translate to increased real cash returns.
What's obscure to many is that these capital investments are happening time shifted from model income. Imagine a sequence of model training / deployments that started and finished sequenced: Pay $10m, make $40m. Pay $100m, make $400m. Pay $1bn, make $4bn. Pay $10bn, (we are here; expectation is: make $40bn).
If you did one of those per year, the company charts would look like: $30m in profits, $300m in profits, $3bn in profits. And in fact, if you do some sort of product-based accrual accounting, that's what you would see.
Pop quiz, if you spend in the first month your whole training budget, and the cycles all start in November, what would the cash basis statement look like for the same business model I just mentioned?
-$10m, -$60m, -$600m, $-6bn.. This is the same company with different accounting periods.
Back in reality, shortly into year 1, it was clear (or a hopeful dream) that the next step (-100 / +400) was likely, and so the company embarked on spending that money well ahead of the end of the revenue cycle for the first model. They then did it again and again. As a result naive journalists can convince engineers "they've never made money". Actually they've made money over and over and are making more and more money, and they are choosing to throw it all at the next rev.
Is it a good idea or not to do that is a question worth debating. But it's good to have a clear picture of the finances of these companies; it helps explain why they're getting the investment.
Absolutely wild valuation given their lack of a moat isn't it?
Microsoft is deeply entwined in OpenAI and has obvious reasons to dogfood, yet their people are using Anthropic solutions.
Valuation behemoth OpenAI has been forced by the market to use Anthropic standards a couple times, having no comparable solutions of their own.
… I can see it.
Anthropic's marketing somehow punches hard. Not sure why, but the stuff they do sticks. Not because the products are great, but because the way they communicate about it gives people the right feeling. They do have legitimately the best coding model now for most tasks, and for narrative prose, but the marketing stuck and people stan'd them even when they were trailing.
> Anthropic's marketing somehow punches hard. Not sure why
The fish rots from the head and marketing depends on being relatable.
https://www.youtube.com/watch?v=qMAg8_yf9zA
Take a scroll through the comments.
Anthropic develops tools for developers and power users which are the actual people doing the evangelizing and marketing for them.
It's Web 2.0 all over again. No moat, winner-take-all (economies-of-scale/network-effect). Just have to out-spend everyone else, and then figure out whether it was worth it all after you win.
Having a cutting edge model that requires tens of billions of dollars to train + a massive concentration of talent and experience + brand + one of, if not the best, coding experiences in Claude Code
These are all moats.
Ah yes the 6m months of product development that’s mostly vibe coded. However will anyone build something better?
It wouldn’t be surprising at all if in 2 months everyone has moved on to another harness. In fact I think it’s more likely than not
The moat seems rather small right now. There are 7 different companies represented in the top 10 models on openrouter.
Have you tried qwen3-coder-next? Model moats are going bye bye
Couldn’t their excellent model and coding experience generate another excellent coding CLI tool?
> tens of billions of dollars to train
Source??
> cutting edge model that requires tens of billions of dollars to train
seems like there are a lot of those out there these days, and the costs are falling
> a massive concentration of talent and experience
Apparently 3000 employees? There's plenty of talent to be found elsewhere. Plus employees can be hired away.
> brand
meh.
> one of, if not the best, coding experiences
Seems easy enough to replicate, given how quickly they built it.
Ain't that for the entire ai field.
FOMO, pretty much
They have a moat on hype.
At least from the software engineer pleb perspective, their moat is that their tools seem to work well more often than not. I wasn't comfortable with the idea of using GitHub CoPilot as our GenAI solution at work, and apparently that was a widespread feeling, because we switched to Claude Code, and it's been a relatively smooth transition from manual coding to GenAI agentic loops.
They did say they were going to cover the electricity bills...
Is this a fair valuation?
I don’t think we will know for a while.
it's crazy that Google is spending something like 4x this in a year just for capex
wonder how much of that $30B will make it their way and pay that down
has me wondering if Anthropic is one of those confidential TPU buyers
likely because Google has stake in Anthropic
These scammers from FTX did put $500 million in Anthropic early on, for about 14% of the company. Later on this was diluted to 8%.
8% of a $380 billion valuation would be a cool 30 billion which I think would have covered the entirety of the fraud and left money for SBF and its friends.
But thankfully around June 2024, the clawback of stolen funds by FTX had its Anthropic shares sold for about $450 million.
I'm glad to know SBF and its scammers friends are going to see exactly jack fucking shit of that money.
"Post-money" is the euphemism for the glorious end of capitalism, when we will be paying in corporate scrip, Arasaka-style? :)
"Post-money valuation is a way of expressing the value of a company after an investment has been made" [1]
[1] https://en.wikipedia.org/wiki/Post-money_valuation
380? Weren’t they talking numbers like 500B? Isn’t this a bad sign?
I can’t find any news stories mentioning a $500B target for Anthropic. You might be thinking of OpenAI’s October raise.
https://www.cnbc.com/2025/10/02/openai-share-sale-500-billio...
Anthropic is the clear category leader in enterprise AI
Citation needed.
As the clear category leader in HN posting, I agree
Google has an edge, always a "personal experience" comment about leaving OpenAI/Anthropic in the dust every time new model gets posted.
It's marketing copy--I wouldn't expect them to say otherwise.
It is quite obviously Microsoft. They use the same (in my dictionary illegal) tactic they did with Teams.
Microsoft engineers use their offerings over OpenAI - their partner.
That isn’t nothing.
They're being told to by management. That says approximately nothing about the relative merits of the two products, but it certainly says something. What, exactly? Who knows...
Other companies have a similar top-down "use Claude" mandate as well.
It is approximately nothing, since lots of MS engineers use Apple products too.
Is it nothing though? How many Apple employees do you think use Windows? And how many Anthropic employees do you think use GitHub Copilot? I would assume the answer to both is approximately 0.
Apple does pretty well for themselves - are you sure that’s not a positive signal for Anthropic?
I mean, I do think it is true, I’m not sure if this is like fastest toddler in the preschool or whatever.