ASML is the flashy, public-facing partner. Mistral is also working with the French government and defense industry for applications that are unlikely to be publicly announced, but are bound to bring in much more money.
The European-sovereignty angle is what makes Mistral's strategy harder to read from the outside. If a meaningful share of revenue comes from EU government and defense contracts, the public benchmarks against OpenAI/Anthropic stop being the right scoreboard — they're optimizing for different procurement rules and a different definition of "trustworthy." Curious if anyone has visibility into how those contracts are structured: per-seat, on-prem, or something closer to hosted with sovereign keys?
"...a long-term collaboration agreement to explore the use of AI models across ASML’s product portfolio as well as research, development and operations..."
ASML is one of the clients Mistral keeps referencing, for example here: https://mistral.ai/news/forge But it isn't clear exactly what they've been doing together. The Forge page only mentions they "train models on the proprietary data that powers their most complex systems and future-defining technologies."
ASML’s EUV money printer has nothing to do with their ability to deploy that money in illiquid investments instead of to their own shareholders
ASML buying equity in one company in tangentially related industry (just because they’re in Europe and the pickings are slim in both offerings and growth capital) has nothing to do with any synergy or integration with ASML’s utility (and bottleneck) to the chip supply chain
remember when you were studying for standardized tests as a teenager? this is what the high scoring answer would be
A friend of mine works for ASML and it’s suggested they were nudged quite heavily “from above” to make this investment, rather than it being an actual strategic play by ASML. Basically “sovereign EU AI” is the play here.
I check almost every mistral model. But to be honest after agentic coding has become a thing, I only test their models in a chat app. I tried their coding TUI a few months back and found that it was way behind. It didn't even support skills back then. I wish they go strong just because they are a big AI player in Europe and I have personal connection to France.
This is why Elon’s appealing. I thought they surely had the talent to be considered a fourth at this point. (Oh someone mentioned they’re politically unpopular at work.)
No they really don't because Grok is not a competitor. The big 3 are the big 3 because they have historically traded the top place for model intelligence. Grok has never crested that high. People like to think Grok isnt as popular because of elon and politics, but if Grok was the best coding model, nobody would give a fuck. Google has also not led since gemini 2.5 pro. It's really the big 2 at this point.
mistral was never competitive and is getting less so, but that doesn't matter they cant be allowed to fail and have a long time to find their lane. They're smart and have an audience of like 600m people and the largest governments by spending who would use them if they were good enough.
I would speculate that the purpose is not just to have a domestically owned model that’s as good as the Chinese one, but for the investment to be used to build a domestic industry of people that know how to do it themselves, so as not to rely on foreign parties, so it doesn’t have to be better, just has to be good enough.
If you're a European AI lab and you see you can't compete with the Amerikans on compute or spending, and the Chinese are open sourcing their LLM work and keeping the Amerikans honest, then there's no real need for you to focus on LLMs
I was skeptical when I saw the headline. And I still am. But AI for manufacturing and industry seems like a good way to differentiate and focus on a vertical that others are ignoring.
What I am curious about is what has Emmi actually built? Who uses it? I was hoping to see something like a demo on the website but couldn’t find anything concrete.
They built a transformer-based mold flow simulator. Mold flow analysis is used in injection molding to predict whether and how the plastic will fill a cavity. It's helpful to make sure the mold will actually be able to fill, and to try to engineer lighter weight or different geometry.
Mold flow simulators already exist, but they're slow. So if you're a mold engineer, and you want to try different material properties, each run takes time.
By making this into a transformer (no idea what that model looks like) it can run pretty fast, because it's not mathematically solving it, it's running a learned function approximator. So for the mold engineer, suddenly they can just change values around and get pretty fast analysis.
They appear to not be selling it as a product per se, but partnered with a German company called Simcon to sell it, whose website lists it as a preview, so it's unclear how good it is compared to conventional simulations.
They seem to be doing let's say case studies on how AI based simulations can help industry.
They did injection molding for example, and I'm sure they're testing similar approach to everything which can be modeled by PDEs, which is well pretty much everything ever of engineering interest (I'm assuming this is somehow connected to a research project funded by Engel, one of the leading injection molding machine companies, located in the same Region): https://www.emmi.ai/models/neuralmould
"Early 2025: First enterprise contracts secured"
"Today: Powering Fortune 500 engineering teams"
- I guess that is all that is publicly available for now.
I assume it's more a talent acquisition as the company only exists for 18 months or so and you can't acquire enterprise customers that quickly, not even as an AI vendor.
The reason we have so much M&A is that large companies can’t really innovate, especially when they’re publicly traded and have shareholders who hate risk-taking
It’s interesting that a French company can compete at international level to some extent, given the regulations, labor laws and generally the business unfriendly environment. I suspect they capitalize on the preference of European governments to use EU products, but might be wrong.
This comment is ridiculous. France's economy is bigger than 90 % of "unregulated" countries.
European regulations help protect from USA's tech monopolies. French labor laws and social security and state-funded scientific schools helped build one of the most competent international AI scientist generation.
All of europe got crushed by the US on the domain of internet. "Regulated" or not.
Indeed, French labor laws and their downstream effects have pushed the most talented French researchers to US-based frontier labs, thus building one of the most competent cohorts of international AI scientists.
This has more to do with the humongous amounts of money sloshing around in VC funds and the disproportionate importance of the US in the global financial markets. They just followed the money. Those who are successful in securing funding then tend to come back eventually.
Curious to see, it sounds like a rather pretty irrational decision. I don’t see many YC companies suddenly running toward France after securing funding.
The impacts of French labor law have been studied. The European Union is a natural laboratory for the impacts of labor policy. There is a strong correlation between looser requirements for terminating employees and startup formation and risk-taking in business. The Nordics and the Baltics have produced more successful technology firms than countries where dismissing employees is onerous.
The impact of labor laws in France is profound. To avoid hiring employees, many firms bring people on short-term contracts. This disproportionately impacts young people. A tradeoff of the risk of being unable to discharge employees is reduced salaries for the gainfully employed. I have met several French people in tech who left because their post-tax compensation was so low relative to employment in the United States.
This information is unsourced and parroted from various articles in The Economist, Money & Macro podcast, and accounts of European citizens I've known.
Taxation, admin hell, government involvement led by highly uncompetent unelected people that got to make decision on the sovereignty and future of
Companies. The only chance for mistral is to escape the grasp of france and its low iq visionless political “elites” or else theyll endup like dailymotion.
Building a company in france and europe is hell even mistral ceo said this a few days ago in front of french officials
What was the jurisdiction ? France is among the most difficult countries for entrepreneurs in the EU (maybe the Germany too, with their written paperwork and rigidity).
In top of high taxes, there is a very adversarial administration and a philosophy of “I’ve told you, you should never have started this project”.
In the baltics it is easy for example, but your experience can vary a lot depending on the country as the EU business environment is not uniform.
Its not ridiculous but accurate. You right the scientist generation is huge from europe but many leave europe… Brain drain in france is huge nowadays because of what op states.
And if you look at mistral biggest customers it would be lying not to say they are done through political ties. No shame in saying that. US gov and agencies created FAANG through those same mecanism
We are using it for "old-fashioned" use cases (sentiment, classification) for some clients here in Europe. Mistral Small 3.2 8bit is good enough for most well-defined cases.
It may just be greenwashing to check “AI sovereignty” off the list.
Nice, also note that ASML is a big investor in Mistral AI, which made the industrial AI ambitions already more credible. https://www.asml.com/en/news/press-releases/2025/asml-mistra...
ASML is the flashy, public-facing partner. Mistral is also working with the French government and defense industry for applications that are unlikely to be publicly announced, but are bound to bring in much more money.
The European-sovereignty angle is what makes Mistral's strategy harder to read from the outside. If a meaningful share of revenue comes from EU government and defense contracts, the public benchmarks against OpenAI/Anthropic stop being the right scoreboard — they're optimizing for different procurement rules and a different definition of "trustworthy." Curious if anyone has visibility into how those contracts are structured: per-seat, on-prem, or something closer to hosted with sovereign keys?
can you explain how it makes it more credible? is the assertion that asml is using mistral as part of its research/manufacturing?
ASML is one of the bigger investor of Mistral
https://mistral.ai/news/mistral-ai-raises-1-7-b-to-accelerat...
From the link:
"...a long-term collaboration agreement to explore the use of AI models across ASML’s product portfolio as well as research, development and operations..."
ASML is one of the clients Mistral keeps referencing, for example here: https://mistral.ai/news/forge But it isn't clear exactly what they've been doing together. The Forge page only mentions they "train models on the proprietary data that powers their most complex systems and future-defining technologies."
ASML knows like no other the importance of doing research in secrecy.
the assertion is that the people at asml are likely in a good position to assess ai use in complex industry
It doesn’t
ASML’s EUV money printer has nothing to do with their ability to deploy that money in illiquid investments instead of to their own shareholders
ASML buying equity in one company in tangentially related industry (just because they’re in Europe and the pickings are slim in both offerings and growth capital) has nothing to do with any synergy or integration with ASML’s utility (and bottleneck) to the chip supply chain
remember when you were studying for standardized tests as a teenager? this is what the high scoring answer would be
A friend of mine works for ASML and it’s suggested they were nudged quite heavily “from above” to make this investment, rather than it being an actual strategic play by ASML. Basically “sovereign EU AI” is the play here.
Abbreviated title leaves out key detail: "Mistral AI Acquires Emmi AI to Create the Leading AI Stack for Industrial Engineering"
That’s more realistic, I was going to say…
Is Mistral still competitive? I completely forgot they existed because of how much press the Big 3 get (Google, Anthropic and OpenAI).
I check almost every mistral model. But to be honest after agentic coding has become a thing, I only test their models in a chat app. I tried their coding TUI a few months back and found that it was way behind. It didn't even support skills back then. I wish they go strong just because they are a big AI player in Europe and I have personal connection to France.
> Big 3
This is why Elon’s appealing. I thought they surely had the talent to be considered a fourth at this point. (Oh someone mentioned they’re politically unpopular at work.)
Mistral is super welcome competition, good luck!
No they really don't because Grok is not a competitor. The big 3 are the big 3 because they have historically traded the top place for model intelligence. Grok has never crested that high. People like to think Grok isnt as popular because of elon and politics, but if Grok was the best coding model, nobody would give a fuck. Google has also not led since gemini 2.5 pro. It's really the big 2 at this point.
Competitive in what sense? In training LLMs? Then they would be below Korea then UAE models.
mistral was never competitive and is getting less so, but that doesn't matter they cant be allowed to fail and have a long time to find their lane. They're smart and have an audience of like 600m people and the largest governments by spending who would use them if they were good enough.
I wonder if they could’ve caught up to Qwen & Gemma by now by now distilling them?
If their best cloud-run offering is far more intelligent than the laptop Gemma / Qwen than nevermind
I would speculate that the purpose is not just to have a domestically owned model that’s as good as the Chinese one, but for the investment to be used to build a domestic industry of people that know how to do it themselves, so as not to rely on foreign parties, so it doesn’t have to be better, just has to be good enough.
I love this point
If you're a European AI lab and you see you can't compete with the Amerikans on compute or spending, and the Chinese are open sourcing their LLM work and keeping the Amerikans honest, then there's no real need for you to focus on LLMs
Not really. They have usable models, but it probably Anthropic, OpenAI, Google -> Chinese labs -> NVIDIA, AllenAI, Mistral.
Built what you want to use yourself. AI for engineering and physics sounds like the perfect product a company like ASML (Mistral investor) could use.
I was skeptical when I saw the headline. And I still am. But AI for manufacturing and industry seems like a good way to differentiate and focus on a vertical that others are ignoring.
What I am curious about is what has Emmi actually built? Who uses it? I was hoping to see something like a demo on the website but couldn’t find anything concrete.
They built a transformer-based mold flow simulator. Mold flow analysis is used in injection molding to predict whether and how the plastic will fill a cavity. It's helpful to make sure the mold will actually be able to fill, and to try to engineer lighter weight or different geometry.
Mold flow simulators already exist, but they're slow. So if you're a mold engineer, and you want to try different material properties, each run takes time.
By making this into a transformer (no idea what that model looks like) it can run pretty fast, because it's not mathematically solving it, it's running a learned function approximator. So for the mold engineer, suddenly they can just change values around and get pretty fast analysis.
They appear to not be selling it as a product per se, but partnered with a German company called Simcon to sell it, whose website lists it as a preview, so it's unclear how good it is compared to conventional simulations.
They seem to be doing let's say case studies on how AI based simulations can help industry.
They did injection molding for example, and I'm sure they're testing similar approach to everything which can be modeled by PDEs, which is well pretty much everything ever of engineering interest (I'm assuming this is somehow connected to a research project funded by Engel, one of the leading injection molding machine companies, located in the same Region): https://www.emmi.ai/models/neuralmould
"Early 2025: First enterprise contracts secured" "Today: Powering Fortune 500 engineering teams" - I guess that is all that is publicly available for now.
I assume it's more a talent acquisition as the company only exists for 18 months or so and you can't acquire enterprise customers that quickly, not even as an AI vendor.
I'm glad Mistral is doing well, but...
I am so tired of M&A. Buy instead of competing or - heaven forfend - cooperating.
The reason we have so much M&A is that large companies can’t really innovate, especially when they’re publicly traded and have shareholders who hate risk-taking
Mistral is not large, though.
Thta's amazing, Mistral is winning silently lol
It’s interesting that a French company can compete at international level to some extent, given the regulations, labor laws and generally the business unfriendly environment. I suspect they capitalize on the preference of European governments to use EU products, but might be wrong.
This comment is ridiculous. France's economy is bigger than 90 % of "unregulated" countries.
European regulations help protect from USA's tech monopolies. French labor laws and social security and state-funded scientific schools helped build one of the most competent international AI scientist generation.
All of europe got crushed by the US on the domain of internet. "Regulated" or not.
Indeed, French labor laws and their downstream effects have pushed the most talented French researchers to US-based frontier labs, thus building one of the most competent cohorts of international AI scientists.
This has more to do with the humongous amounts of money sloshing around in VC funds and the disproportionate importance of the US in the global financial markets. They just followed the money. Those who are successful in securing funding then tend to come back eventually.
Curious to see, it sounds like a rather pretty irrational decision. I don’t see many YC companies suddenly running toward France after securing funding.
Do you really believe this? Lol
Could you please elaborate what labour law drives the labour out of france?
Haha yeah. I'm being a bit silly but I do.
The impacts of French labor law have been studied. The European Union is a natural laboratory for the impacts of labor policy. There is a strong correlation between looser requirements for terminating employees and startup formation and risk-taking in business. The Nordics and the Baltics have produced more successful technology firms than countries where dismissing employees is onerous.
The impact of labor laws in France is profound. To avoid hiring employees, many firms bring people on short-term contracts. This disproportionately impacts young people. A tradeoff of the risk of being unable to discharge employees is reduced salaries for the gainfully employed. I have met several French people in tech who left because their post-tax compensation was so low relative to employment in the United States.
This information is unsourced and parroted from various articles in The Economist, Money & Macro podcast, and accounts of European citizens I've known.
Taxation, admin hell, government involvement led by highly uncompetent unelected people that got to make decision on the sovereignty and future of Companies. The only chance for mistral is to escape the grasp of france and its low iq visionless political “elites” or else theyll endup like dailymotion.
Building a company in france and europe is hell even mistral ceo said this a few days ago in front of french officials
As somebody who's built a company in Europe, I don't know what you're talking about. I suspect you don't, either.
What was the jurisdiction ? France is among the most difficult countries for entrepreneurs in the EU (maybe the Germany too, with their written paperwork and rigidity).
In top of high taxes, there is a very adversarial administration and a philosophy of “I’ve told you, you should never have started this project”.
In the baltics it is easy for example, but your experience can vary a lot depending on the country as the EU business environment is not uniform.
Its not ridiculous but accurate. You right the scientist generation is huge from europe but many leave europe… Brain drain in france is huge nowadays because of what op states.
And if you look at mistral biggest customers it would be lying not to say they are done through political ties. No shame in saying that. US gov and agencies created FAANG through those same mecanism
I have yet to see someone recommending Mistral for anything tbh.
Their Voxtral[1] speech models are really good.
[1] https://mistral.ai/news/voxtral
+1 for this, I've found Voxtral to be the best combination of price/speed/accuracy.
We are using it for "old-fashioned" use cases (sentiment, classification) for some clients here in Europe. Mistral Small 3.2 8bit is good enough for most well-defined cases.
It may just be greenwashing to check “AI sovereignty” off the list.