Mistral is an interesting AI company because they clearly have a contrarian business strategy to the other AI labs. They're also landing big customers in Europe for the right reasons. People dump on them because they're not benchmaxxxing which is pretty shortsighted - do you really want to be in a benchmark arms race with China, or do you want to make money and deploy sovereign AI compute in Europe?
When Apple is not racing for the frontier it's a strategy, but when it's Mistral it's a mistake.
The two companies have read the market the same.
It's always very dangerous for first movers and their investors, and the commodification of intelligence seems even more likely each time a chinese open model release. It's less exciting to do business that way, but if you're building to stand the test of time, it's wiser that way.
I am not sure it's correct to lump apple and Mistral's strategies together. Apple's business is selling hardware/services and their stores, but Mistral's business is AI.
Apple's strategy seems to be "wait till real business shakes out" but Mistral's strategy seems to be "go after profitable niches and avoid unwinnable fights".
> When Apple is not racing for the frontier it's a strategy, but when it's Mistral it's a mistake.
I think what Mistral is doing is smart within their financial constraints, but this comparison is misleading. Mistral is an LLM company; Apple is a consumer hardware and services company.
It's smart for Apple not to join the LLM arms race, because they can just pick the cheapest supplier and let other companies take the financial losses. Mistral is in a very different situation; they are the supplier.
Apple is in an entirely different market than Mistral (consumer electronics vs AI lab focusing on enterprise consulting). Which obviously means their optimal strategies are different. It doesn't really matter for Apple if it's Gemini or other model running behind their AI features. If anything it saves them a lot of money and provides a lot of flexibility.
AI company not keeping up with AI companies is by no definition the same as combinedhardwaresoftwareservicesentertainmentlifestyletechcompany not keeping up with AI companies
They aren't following the same strategy. Apple does race for the frontier in their main field: beautiful, well-integrated hardware and software. Mistral does not race for the frontier in their main field: AI creation. They're grabbing bits and bobs from people that need (or feel they need) sovereign AI. It's like if you go into making phones for the military. You aren't going for the best phones; you're going for making sure you're the only company that has the right connections to keep that customer.
That's uncharitable. Defense companies do play the game (like anyone in the enterprise space), but they also do ship stuff that is uniquely set up to satisfy the use-case. I don't think an unmodified iPhone would do great in a warzone with no cell signal.
Yea people tend to give Apple the benefit of the doubt because they’re the most successful and valuable company in human history.
Mistral is not Apple, and is not emulating their strategy. Please show me Mistral’s half a $Trillion in yearly revenue coming from consumer hardware/software.
Then I’ll agree with you that they’re taking the Apple strategy.
I root for Mistral and hope they'll be successful, perhaps I'll buy a subscription too once they're good enough for coding aid (perhaps they are now, didn't do any test with their models recently).
First of all, they release the models' weights, perdonally I don't consider any other option as viable (no OpenAI and definitely no Anthropic, thank you).
I especially like their Vibe Chat web offer, the allowed monthly usage with a free account is incredibly generous (still have to hit a limit) and the deep research feature (5/month for free) is also valuable.
I don't know anything about the alleged regulation maxxing problems, I don't perceive them as a problem for my causal/personal usage anyway.
The gap has only been increasing, though. Devstral 2 was obviously not great compared to Claude/GPT but kind of acceptable if you were willing to compromise. There has been no real progress since then and frontier labs are massively ahead.
I don't care about benchmarks. Benchmarks show that Opus 5 is a stronger model than Fable 5 which is obviously not the case.
But I do care about capability and so far only Anthropic and, very recently with Astra, OpenAI can deliver on coding quality. And capability matters immensely. There is a world of difference between being able to do something and not being able.
A capability isn't binary. There is a massive difference between can produce an impressive demo and can reliably complete the task without a human babysitting it.
It's not this week's change. Fable was the step change for programming. And most of truly useful and powerful capabilities arrived in the last eight months.
AFAIK: Mistral does not even try to compete in this field. There are other use cases for LLMs beside coding. As Mistral AI wrote:
> During the first wave of generative AI, the central question was who could build the most powerful model. Organizations and governments are now asking a different one: how to harness the power of AI for their mission-critical needs without surrendering control over the infrastructure and intelligence loop. Demand for that combination of performance with control, choice and independence is growing internationally, as enterprises and governments weigh the long-term technology dependencies, data governance requirements and deployment choices that come with any AI investment.
> Mistral is the only AI company in the world building the full stack required to answer that question: open-weight models, the infrastructure and the compute capacity they run on, and the products that bring them into production; ensuring that customers are never locked into a single vendor's roadmap, pricing or availability.
> Mistral’s full-stack and open approach also allows organizations to build on it without exposing their most valuable data, workflows and institutional knowledge to anyone outside their own walls. That's what makes Mistral’s stack the sovereign AI layer, meaning retaining control across four dimensions: data that stays inside the organization's boundaries, models that are controllable and customizable, compute that is private and predictable, and systems in production that are fully controllable and auditable.
People said this for Opus 4.6 too. Every release the models get RLHF'ed into accomplishing a new task and the people who need to do this task think there was a step change.
Yeah that's kinda my point. I'm not sure if the models have gotten that much smarter, but they're certainly getting more capable. That's not the same thing though.
There are things GPT 6.0 can accomplish for me that 5.3 was not able to. But there are also things it still fails at, and it doesn't seem to be much better at the big picture. It does spam about 100x more tests though and I wonder if just RLHFing it to test everything constantly is carrying it more. 6.0 writes so many tests and spends so much time verifying it's work in python sandboxes. Slow as hell but it tends to get things right the first time more which is good, I guess. I don't love the thought of a 500loc feature adding +4000loc due to tests though.
After living in the US for several years, I was happy to return to the EU where regulations exist to protect against the worst corporate behavior.
My American bank sold my credit card transactions to advertisers. My American mobile operator had insane fees for roaming and other features that are basic in Europe. Sending a bank transfer in America was unreliable and slow and expensive because there was nothing like SEPA instant 24/7 free transfer (I guess FedWire does that nowadays, I don’t know if consumers actually have guaranteed access at all banks like they do in EU).
When American AI companies become established Fortune 100 members, they’ll start abusing their customers just like all the others in that club, those banks and phone operators and
Microsoft and the rest. Google once pretended to be different, now they have the corporate cancer. No reason to believe the same won’t happen to OpenAI and Anthropic.
There is regulation that is effective and useful (e.g. plenty of consumer protection laws) and regulation that actually makes markets more competitive and efficient. Then there is a lot of what EU is doing which leads to less efficient markets and more stagnation.
Even a lot of their attempts to increase competitiveness like forcing Apple to allow alternative app stores have been halfassed and not very effective.
Sure but it's not the government calling for regulation, it's Mistral. At times they seem like a government mouthpiece to check the waters, which I could accept if their models were at least any good
They've both been insisting that humanity's survival is contingent on regulation since the start and employ armies of lobbyists to try to shape the policies they want that hold significant sway in their own country. But of course Americans doing this is smart and entrepreneurial...
Meh, OpenAI and Anthropic are doing the same. The US shtick is "our models are too powerful, government, please hold them back (and buy 10% of the company)".
Edit: forgot, they also call for banning Chinese models and open weights models...
> Markets, left to their own devices, do not end up automatically being competitive in favor of consumers.
True. Does not mean all regulation is increases competition and marker efficiency by default. If anything EU has long abandoned the core tenets of Ordoliberalism and "Social Market Economy"
Great example. An highly regulated field (aerospace) put so many barriers of entry for startups and competition that the incumbents simple have to reason to innovate or even compete anymore.
I heard there used to be dozens of plane-building companies when the field was unregulated, with new, more efficient models and lower prices appearing regularly. I wonder what happened.
Right now the weakly regulated American market is delivering staggering competition where massive corporations are fighting tooth and nail over my 100 bucks.
I kept hearing the same about previous age tech and there also Americans rule the Internet: Google Search, Workspace or m365, AWS, Cloudflare... Even Linus moved to the US.
They've been pretty efficient in squashing any potential competition because of that weak regulation, yes. Don't confuse it with competing on the merits of their products.
There's a big difference between governments regulating to benefit their own control and power, and regulating in favor of consumers.
Most regulation in the EU have jack-all to do with consumers. There's a couple of niceties - having all prices displayed in comparable unites (EUR/kg, EUR/liter etc). But apart from that the regulation have broadly served as moats around the largest businesses.
I wouldn't even call it a communication problem - I believe it's more a media problem, than anything. Good news are minimized and bad news are amplified to the maximum - nowadays even broadsheet newspapers operate with tabloid headlines, and that is really bad.
We somehow need to "want" a slower news cycle and simultaneously push back against the algorithmic outrage. The world is more predictable and safer than ever but the perception of the average person is not this, it's manufactured outrage at every possible opportunity and topic.
Despite the average european objectively being better off and safer than a decade ago, almost none feel that way and that is almost 100% due to the news algorithms being tuned to keep you in a constant dread cycle.
> cooling off [...] seller liability [...] Right to repair
Let's not forget I am paying directly for those when I buy a product. Those benefits are not free, they cost money. The money has to come from somewhere. It can't come from profits (since those are the sacred reason the company exists in the first place) thus they will come from higher prices.
Higher product prices mean less competitive companies. It also means fewer startups which must implement the required regulation no matter the market says. Then you need protectionism against Chinese competition.
But the worst is that I must pay for those benefits whether if I want them or not. Thus my choice is reduced. I cannot vote with my wallet and send price signals in the market to tell sellers that I want those benefits or not and at what price. The market gets corrupted. The economy falters. Outside competitors gain. The Government needs to intervene again and again, growing stronger and more intrusive "for my own benefit".
It's the decay spiral we are witnessing here in the EU right now.
That's not even true in the concrete case we're talking about here.
As part of the AI Act, an important decision on a person (such as a hiring decision) cannot be made without human involvement.
You think that's not in favor of "consumers"? (also note the implicitly submissive framing of humans as receptacles for consumption instead of subjects with rights!)
This is a bottom feeder mentality. Europe has enough bright people and resources to truly compete in the AI race. There is something wrong when the only selling point is that it's local.
We're regularly getting demonstrations that being at the frontier is no moat at all. "Run open weights locally" was literally on the HN front page a few days ago as a primary concern/competitor for the big US labs. So there's a big and meaningful gap between "not frontier" and "bottom feeder". There's just tons of applications where you don't need "the best" model, especially not tomorrow's best model.
I wouldn't. I can think of at least 10 regulations (labor and environmental chef amongst them) that would render such a maker uncompetitive from the start. Chinese makers would eat its lunch.
Those things are not mutually exclusive. We can have open weight, yet close-to-frontier models. China has very strong models that can also run locally.
To compete, we’d have to throw a lot of the regulations and laws into the trash (especially anything regarding copyright) and do an order of magnitude more investment.
I’m surprised that there’s no domestic chip production either, we don’t have our own CPUs or GPUs, meanwhile China is spinning up manufacturing so they don’t have to work around the Nvidia export restrictions as much.
I like Mistral and there’s cool stuff going on like how EuroLLM models know Latvian language and all the other EU ones better than way bigger models, but we don’t have anything frontier.
At the very least, they should be distilling Kimi K3 and GLM 5.3 as much as possible and working on MoE models like ~35B and ~120B versions to match Qwen.
No, Europe doesn't have the capital markets required to build the required data centers. Underwriting gigawatt data centers and frontier models demands a fully realized European Capital Markets Union, a single energy regulator with cross-border grid integration, and shared fiscal borrowing power.
Realistically they will have to deploy the Chinese models or their finetuned versions though since their models are completely out of date and not competitive. Outside of maybe government contracts it will be hard to compete against Azure/AWS who promise to run their models in EU datacenters and not store any data since actual companies normally prefer frontier models with decent performance (cost/performance is pretty decent as well if you are fine with e.g. Luna which is massively better than anything Mistral can offer).
There are two separate issues here. One of which is very simple. That issue is where to run the models. For many companies this has to be in the EU, on EU terms. Mostly, this is not really optional from a compliance point of view. It's why all the big cloud providers have data centers in places like Frankfurt, Amsterdam, etc. and why a lot of new data centers are being built in Ireland. Of course a lot of those investments are being made by US companies. But they all have legal entities in the EU because otherwise they'd have no business here. And they can't afford to miss out on that business because it's a huge market.
The second one is about which model to run and who controls and oversees quality control. OpenAI and Anthropic seem to insist that only they can do that. But of course here in the EU we see that a bit differently. The big US based hyper-scalers are neither liked nor trusted here at this point. We don't trust the Chinese model makers much either. But with open weight models, we can at least pick different models and run them on our own terms.
Also, what most companies need is not necessarily the latest fashionable model straight from the Silicon Valley cat walk but something that will work reliably and predictably for years. Factories are not going to install the latest model in their production lines every few weeks. Same with most banks, insurers, etc. I actually know people that do business with those in relation to AI development in Germany. Companies like that are very much obsessing about self hosting their models. Sending customer data off premises is a big concern for them. They are building stuff that will be used for many years. In five years, nobody will care which model was best in autumn of 2026. But a lot of software built this year that uses AI might still be running.
You have to see Mistral's investment in that context. They could make a lot of money in the EU if they do a decent enough job. Lots of conservative companies here that are going to pick something that's good enough and then they'll be using that for many years.
With how they’re currently being used we might as well call them bendmarks.
Every newly released model is paraded as SOTA showing peak or near peak performance on cherry-picked bendmarks the model was either fine-tuned on, or tested under specific conditions optimal for that model.
I tried them via OpenRouter. I loved their OCR. I really disliked their code generation. It was about six months ago -- so it was a geological era ago in this world. However Mistral is legally favored in Europe. In fact from my point of view , using them presents no trouble with GDPR (I live and work in Europe). I'm NOT a lawyer but I'm a technician that define itself 'privacy savy'.
I fail to see how this is relevant with regards to areas.
Whether AI is hosted by the USA, Europe or China - they all are awful and eliminating real jobs while also driving up RAM prices etc... Why should I want to support any of these?
Because it's increasingly obvious that this is the next step in our capabilities as humans competing with that of the discovery of bacteria and transistors.
Because humanity stalled out- and coasted for the last 50 years and it shows. And now it must compete and git good or git gone. No more fat ponies paraded as race-horses.
The last 50 years is a wild take considering thats the time in with the computers became a bigger thing, mobile phones, smart phones. CRISPR, mRNA Vaccincs, HIV Antivirals, HPV Vaccines, We friggin confirmed the Higgs Boson and Gravity Waves. Found thousands of other worlds outside the solar system. Lithium Ion Batteries, the modern solar panel making solar power the cheapest energy source in most of the world. Blue LEDs making RGB LEDs a thing. Reusable Orbital Rockets.
Mistral is not that bad as the comments here suggest. I am not using it as a frontier model but with simple RAG tasks and its doing great. Also OCR is pretty decent. It's a positive development that Europe is at least trying. Alternative would be: do nothing.
The problem is that they're in a weird position between US models and Chinese models. Not as performant as US models, not as cheap as Chinese models.
Especially as Chinese models are getting better Mistral is getting less and less relevant.
It pains me because I want them to succeed, but despite them denying it I believe they'll end up restrict their activity to (1) selling hosting for Chinese models (they're already hosting GLM) and (2) selling AI-related consultant service (they're also doing that already).
Which is interesting since who are the investors and what exactly are their roles? Samsung - European? BlackRock - European? Salesforce Ventures - European? Etc.
So yes it might well be
> [...] the largest equity fundraising round ever completed by a European technology company, three years after the company's launch.
When it comes to what really matters, they're far behind and probably will not catch up. It is my conviction that in this space, if you're not the best in the world, you're losing. Everyone is fundamentally selling the same thing, so if you don't have the most intelligent or cheapest model in the world, you're losing. Sure, Mistral has a "Made in Europe" edge, but any open-weight self-hosted chinese model might as well have been made by Von der Leyen herself.
Also, €3B is nothing in this market, especially when you're competing with more efficient competitors. €3B in Europe is probably the same as €10B in the USA and €20B or €30B in China.
> It is my conviction that in this space, if you're not the best in the world, you're losing
I disagree on that point. Models are getting good enough that you can switch them and barely notice. I'm switching between Opus, GPT Codex and GLM 5.3 for coding and I can barely tell the difference.
I think they'll become more like telcos than anything, selling a commodity. It's even truer when any provider can host open weight models like GLM-5.3.
Basically a world with dozens of Baseten, with AI labs having a hard time monetizing, just like editors of open source software.
I can tell the difference between the SOTA and GLM 5.3, and paying a few hundred for the better models is definitely worth it.
Mistral does not offer a model that makes sense to use. I understand they now host the already outdated GLM 5.2, courtesy of China providing the weights. And Mistral offers this for 2x-3x the price of other providers.
Europe is a different beast. They build and distribute rails, they don’t compete on frontier capabilities. When the civic benefits of AI become clear, EU is in a position to mandate their distribution. US develops capabilities that remain stuck in heterogeneous corporate silos without interop. Payments is a good point of comparisons between the two approaches.
I don't get it when people all claim that AGI is a winner takes all game. It is not (unless it is used as a weapon). When one company reaches AGI, there will be a dozen very close to AGI, given time. Also a winner is not going to drive everyone else out of business, it is the opposite, one winner will have people betting on the second and the third winners, the technology will also help other develops. Once you have a good enough model, everything will be incremental. I think the hardware capability will be the real burden, not the model itself. If AGI is as powerful as it sounds, maybe hardware won't be a problem any more.
Well, then Samsung Electronics appears to be dumb for leading this investment round? They probably only do it because they're European...oh, wait...
> so if you don't have the most intelligent or cheapest model in the world, you're losing
So how does this match up to the fact that there is currently OpenAI and Anthropic, both raking in money? They can't both have the smartest model at the same time, can they? And all those inference companies selling API access to open weight models on OpenRouter, which are apparently also earning billions already? While the former are probably bound to have much higher cost for research and training than they are currently earning, which may be called "losing", the latter don't have that problem, they can simply price their API access such that the money earned covers their costs, no training and practically no research necessary. In your theory these companies shouldn't have a cent of earnings.
> So how does this match up to the fact that there is currently OpenAI and Anthropic, both raking in money? They can't both have the smartest model at the same time, can they?
They are simultaneously first: the two leapfrog each other with regularity, and are meaningfully ahead of the competition.
OpenAI and Anthropic are so close on benchmarks and release date that they are essentially ex-aequo at this point. The market is just hedging their bets.
There is no close second, because the AA Index points are expentially harder to get as you get closer to the first.
The value of these companies is not defined entirely by the state of their current models. A much more important signal is their chance of having the best model in the future. Just like with anything having to do with investment, this is the castle in the sky. Also, shame on me for simplifying things so much - but I still believe in my original sentiment.
That's kinda damning with faint praise, but I agree, I am glad to see something. I've been disappointed in their coding ability -- it's where I'm most focused, I've built and we are selling a (specialised) coding agent -- and hopefully investment will give them the ability to achieve more in their research and model development.
They have been focusing largely on government and business not consumer, which is fine. Perhaps coding is not something they want to achieve, but they do provide Codestral. It's a signal it's a market of interest to them.
Ehm, sorry, but no, this is not how lithography works. You cannot "hide" functionality in circuitry your machines are producing if your machines' job is to shine light through a mask.
You'd have to be the producer of the machine producing the mask. Or, even better, the software that produces the plan according to which a machine produces a mask.
I didn't say to hide it. My suggestion is to create international AI safety rules which companies must be in compliance with if they want to buy and service ASML machines.
Europe absolutely needs a home-grown AI lab, especially with Pax Americana looking increasingly shaky.
LLMs embody value systems, and American and European values are not the same (yes, there are overlaps, but also key differences).
More nefariously, I can also imagine LLMs that silently degrade their reasoning when used in a national security context of a non-US country.
So, Mistral may not be competitive with OpenAI and Anthropic, but in many contexts that doesn’t matter. And, perhaps this gap could be closed with more funding (the three billion funding figure is a rounding error next to US labs). I’m sort of surprised that the EU isn’t stepping in to support them.
>I’m sort of surprised that the EU isn’t stepping in to support them.
On a similar note: Why does it have to be the EU to step up?
Why doesn't EU rather speed up making VC investments more attractive, so that EU and banks don't do the majority of investing?
*I don't have answers to these questions. It just frustrates me how many investments here come from politicians and banks, rather than from investors, people, and companies.
> LLMs embody value systems, and American and European values are not the same (yes, there are overlaps, but also key differences).
There are also some areas where values across Europe are quite divergent. Think LGBT rights and social acceptance, religion & secularism, immigration & multiculturalism.
And countries don’t fall into neat “liberal west vs conservative east”. Spain is exceptionally liberal on LGBT issues while remaining more religious than some northern countries; Denmark is socially liberal but has adopted relatively restrictive immigration policies.
This "Non-Signatory Participants : Taiwan *" is so weird precisely because Taiwan is the one who started the Silicon shield few years ago by making TSMC central to the whole ecosystem.
> LLMs embody value systems, and American and European values are not the same
Your values and American values might not be the same, but to say even most of Europe feels the same way is a big stretch. And not even the US has very many shared values anymore.
If we’re being honest, Europe doesn’t actually have much of a common value system outside of whatever is momentarily trendy in the urban monoculture, which is why it refuses to work together on most things and is currently being torn apart at the seams (see the rise of nationalist far right parties in most states).
The EU are a collection of states where the average citizen can’t even communicate with their neighbor in a common language beyond the level of a 4 year old. How could they possibly be aligned on a value system in the same way the US is.
Mistral has solid OCR, STT and TTS models and I would love to support them by switching with all of our business workloads to Mistral... but their LLM models are sadly not competitive at all. In our business benchmarks their Mistral Medium 3.5 with reasoning is worse than Gemma 4 31B and Glimmer 30B. It's a 128B dense model that's priced accordingly! Mistral Small 4 is way worse than Gemma 4 26B A4B. I applaude the effort that they release those models as open weight but Gemma 4 models are currently way easier to run with more tok/s and less hardware. Their API pricing is just insane for what you get. But I guess enterprise customers don't care about it, this is why they are probably not lowering it.
Mistral just needs to good enough category think all those flash models or Qwen3.8 27b which they sadly aren't at the moment, that plus being European lab will mean that they will have very nice business. Even now these SOTA models feel too overkill for most tasks.
Last I saw job offers for mistral (engineering, Paris) it advertised 90k euros base salary. Not sure how they’re intending to compete with the US when even a top AI lab can’t afford to be competitive :-/
You dont even live in the Alps with 90K lmao. And if you expect top performance, you should expect top salary. Or at least something competitive. 90K for a company like Mistral is a bit embarrassing.
I actually live in Paris and get your point, but at some point you have to face the reality of the market/industry and really try your best to get good folks to join, especially in the current climate where US researchers may be looking for greener pastures.
US software companies (Google, Meta, MS, etc) have offices in Paris and they pay way more than that. Especially if you take into account total comp (Mistral is private so any stocks you get are "locked" until IPO or sell out)
Yes 90k is more than most French software companies pay, but the salary needs to be competitive with those US companies.
When I was in Germany a few years ago, I was browsing jobs available in France and most software development jobs were under $55K USD. I don't know how you're supposed to survive on that. You can probably make more money as a waiter.
Germany is not so great either in that sense though at least in crypto sector, the pay was reasonable... Though as a job it was total BS.
UK was pretty good at some point (though all the really good jobs were gate-kept by IR35 regulations), The Netherlands was pretty good (especially contract roles thanks to big tax discounts for foreigners). But IMO, one of the best countries outside of the US for software work has been Australia. I actually came back to Australia because of it. Though it's not quite as lucrative today but it's certainly still much better paid than France and has potential.
That said, if you want to do your own startup and you do not have special business connections or pedigree, then the US is the only real option IMO.
I think it comes down to easy money. Easy come, easy go. To be rich enough to invest in startups in France, you need to be born into old money or be a complete freak of nature. These kinds of people don't part with their money easily. They need a guaranteed return and they can't rely on any media channel to bring traffic. If you live outside of the US, all media channels are essentially working against you; you can literally feel that and so everything is riskier. It's like playing the same game with a massive handicap.
But you should be able to hire most of the people you want, not just some that happen to pick Paris.
Also, if given a choice of 90k eur in Paris vs let's say 2m usd in US I'm pretty sure suddenly most of the issues people have with US would suddenly disappear ;)
I actually believe the amount of $ is what was missing for them to improve their fundamental flaws. They're a very competent team, but were working with a tiny fraction of the budget of US/China teams.
If Mistral aren't going to distill other people's large models they obviously need to train their own large models. This obviously requires money for optimization, tuning and training hardware.
They've started hosting GLM-5.2, that should be very telling of their capabilities at the moment. Hopefully the investment will allow them to hire the right people to become competitive.
Chinese models are trained on dubiously collected model traces from Claude/OpenAI that are purchased from model routers. All the major Chinese models use this data. That's one reason they've been able to catch up with Anthropic/OpenAI so quickly, they have so much data.
Mistral can't train on that data, because this data would be illegal to purchase & train on in the EU.
Shame them... for not working their employees into a heart attack at 40?
I'm not sure if your hypothesis is correct. I personally don't think the crazy work hours are what make the US so competitive in tech. But even if it were, it's odd to call for shaming countries for having less insane work cultures.
Way to reduce a multi dimensional concept into a single scalar value. Care to explain what exactly make it inferior? Try using more than one word if you can.
there was "H" at paris at the time, they raised 200M or so, but never got so much visibility. don't know at which stage they are now or if they accomplished something
Honestly being only 1 year behind makes me an optimist. You’re telling me Europe can be slightly behind with 1000x less capex and way more sustainable economics? Awesome. The world moves slower than AI progresses, I can see a scenario where 1 year isn’t a problem.
Chinese models are open, available to distill, and they also publish papers about their research. Being one year behind is a skill issue.
I think the most of the money would go to purchase hardware, But I hope they can start making adequate compensation for AI engineers and researchers to move forward.
My feeling is that its a difference in how funding works in different places. The USA will go all in with the populations pensions on a gamble, the Chinese subsidize. This way of operating is typical for the EU.
> Being one year behind is a skill issue
E.g. if you are an AI researcher in Europe you can just go to USA and make generational wealth. This is not a criticism of the EU model, but rather insane American capex effectively monopolizing.
> I think the most of the money would go to purchase hardware,
> > I think the most of the money would go to purchase hardware,
> So it's not a skill issue?
I meant by offering sovereign cloud/inference, not for training. But even if it was for training, Chinese labs have limited supply of GPUs, look what they've done. So it is a skill issue.
Also to clarify, I didn't mean European engineers' skills, I meant "you get what you pay for" as a company, that's why I hope they start offering better compensation to retain talent.
Handing in a photo-copy of a photo-copy of a photo-copy of a photo-copy of a photo-copy of a photo-copy of someone else's homework might work once or twice, but it is no way to run a (non grift) business.
If you follow the ideology of the Silicon Valley billionaire types, based around Ayn Rand (eg “Atlas Shrugged”) and Curtis Yarvin, as well as what Thiel himself has said over the years, the Antichrist is governmental institutions, seen as the inhibitors of progress. I am not joking btw.
to be fair, there is no plausible explanation for Mistral's fall into irrelevance other than having to deal with regulatory burden. US and CN labs get to feed their models all the data in the world while Mistral has to comply with Let's-Smother-The-AI-Baby-In-The-Crib-To-Maybe-Prevent-It-From-Maybe-Becoming-A-Criminal Act that EU nepo babies had expedited as soon as AI became the current thing.
Mistral had open-weights SOTA when Chinese labs had no noteworthy LLMs at all, and now even its closed-weights offerings are behind half a hundred Chinese models -- most of them from labs with way less funding -- despite having had unhindered access to Nvidia hardware all this time.
I know AI has more regulations here in the EU, but what I don't see clearly explained is what impact that has on pure research: creating new comparable models. So I'm suspicious that it's really impactful. On market deployment, maybe - that's where regulations lie. On internal company research?
Most things in the EU AI Act are entirely reasonable. On its own it's a pretty sensible piece of legislation.
But the US and China are incredibly lenient on enforcing even existing laws against their AI companies. In the US the standard seems to be "no instructions to make bombs or drugs, and the ability to make porn of living people has to be restricted to paid users". Laws around cyber security and copyright that used to be heavily enforced are seemingly suspended for AI companies, or settled out of court. No company that tries to actually follow the law can compete with that
On its own, all regulations usually look reasonable.
It’s when you stack over 100,000 regulations (this is literally the number the EU has created) on top of each other and then combine them with country level regs that you create a system that’s impossible to grow in.
Regulation might have something to do with it but your example doesn't show that imo.
Multiple US labs were taken to court over copyright infringment for their training data. Mistral did the same data mining, but actually the copyright situation is weaker here so it's harder to build a case against them.
I think the capital situation in EU is just worse. A lot of is tied up in more conservative businesses that are reluctant to bet the house on some shiny new thing.
Well, sure, but on the other hand Peter Thiel thinks she might be the antichrist, and he has an awful lot of money which means he is an expert on everything.
He says she’s the Antichrist because she’s against progress and threatens the continued thriving of our civilization. Being antichrist is an archetype.
The whole "It is easier for a camel to go through the eye of a needle than for a rich man to enter the kingdom of God.” also hints at _something_.
But, I guess some very religious people did not got all the subtleties of "Thou shall not kill" yet, do, let's give everyone another 2000 years to process the idea.
(Also, maybe it's hard for a camel, but easier in ocaml ? I'll have to try marketing "Static typing as a way to reach heaven")
> He throwed the moneychangers out of the temple - so I guess he had a short temper for stupid frauds.
He also expelled demons from a man and conceded to their request to be installed in 2,000 pigs. That led to them drowning, causing substantial destruction of somebody else’s property for no obvious practical necessity. The local people become frightened and ask Jesus to leave their region.
This is why wealth should be less concentrated. A few people has their hands in all businesses and that will end badly.
Taxation is not just about paying for the costs of the country, it is also about making sure that nobody has enough power to subjugate the will of the people. Monopolies, billionaires, the accumulation of such power is making the average citizen powerless. And you can feel it.
> Taxation will not solve this before the global society collapses
Taxation can and must solve it before "society collapses". And many people are working to solve the problem. It is difficult and requires effort but nothing that has not been done before. We got rid of king (by the will of god) in the past so techno-power-hungry cringey businessmen should be easier.
Yup. I tried to say that conventional taxation is not enough. As the power is hidden in Panama bank accounts we need to figure out means beyond taxation to neutralize it as it being used to obliterate the future.
It makes no sense to train frontier models from scratch anymore. The best frontier models are only a half year ahead of Chinese open models. In this regard Anthropic and OpenAI are also in a bad spot when they waste so much compute on training models.
An important factor is that fine tuning existing open models is incredible cheap. You can easily change any cultural biases if you want a model to be 'sovereign'. And Mistral could combine that with their custom data sets for their enterprise customer needs. Mistral still trains their own models, but they also seem to offer fine tuning existing models.
With model weights being commoditized, another differentiator could be deploying efficient inference chips, especially if you combine it with a developer ecosystem for vendor lock-in. That is why it is interesting that both Samsung and ASML are investors, since they are companies that could make a difference in this area.
It only makes sense to train a frontier model if you are trying a different architecture to one that is available from an existing frontier model. This is because the different model architecture will learn the weights differently.
It may make sense to train a frontier model on an existing architecture if the base model is not available and the instruction trained version doesn't fit with what you want. There are techniques like ablation, but those could have other effects on the model, and there can still be lingering effects of the instruction training in the model that surface less frequently (e.g. on an input not covered by the ablation training).
Otherwise, fine tuning is definitely the way to go. However, you need to be careful not to over-tune the model such that it is only tuned to the data you are training it on.
Maybe true for frontier LLM, but there's plenty of space in the niches. For example, I think their TTS/STT models are pretty good, speaking from personal experience.
The only reason OpenAI and Anthropic are making a loss is due to training new models to keep up with competition - not because the industry is flawed in terms of business model.
> Is that not self evident by the insane revenue from frontier labs?
No...? Of course not?
Because revenue is only one side of the equation. Did you ever look at total cumulative OPEX and CAPEX, and how long it will take them to even just break even at current growth?
They're likely over $80b ARR by now. They'll be at $800b ARR next year at the same rate. Let's say their growth gets cut down to 3x instead of 10x - that's still $240b ARR by this time next year.
When you are growing so fast, you don't need to make a net profit. You just need to make sure your unit economics are good - which it seems like they are given reports that their gross margins are at 60-70%.
> They're likely over $80b ARR by now. They'll be at $800b ARR next year at the same rate.
And they will be 800 trilion ARR in a couple of years, following that same rate! 8 quadrillion by 2029!
> When you are growing so fast, you don't need to make a net profit. You just need to make sure your unit economics are good - which it seems like they are given reports that their gross margins are at 60-70%.
If their margins were anywhere near this good, they wouldn't need to raise so much money so often.
If you create a machine that turns 1 dollar into 3 dollars, you don't dillute your ownership of the machine, you use your fabulous profits to expand your machine's capabilities.
They’re in the EU. If they are a year or so behind and things start to plateau they’ll catch up. Americans perhaps don’t realize we can also tariff their digital goods to protect our own. There is a scenario where Americans and Chinese foot the bill and EU gets out cheap, e.g. similar to the Apple approach to AI.
I really do hope they can catch up with the American models, but I think setting the Chinese models as the goal would be best. They seems to be able to create great models with low cost that are probably useful for 90% of the day to day tasks. Mistral should not focus on competing with Claude Fable or OpenAI's Astra at first but have a good EU alternative to Opus or even Sonnet. The fact that it's European will be enough to be used by a lot of companies and governments in the EU that are (trying to) move away from US tech.
Nah. They are not, everyone is losing money, and just staying alive from gov subsidies. A claude PRO license is 20 usd/month, and for it to be profitable it should be somewhere around 200-300/month.
I think the Pro and other subs allow them to save big on ads (Claude Code use is required and it pushes their ads) and allow easy access to free data (CC occasionally solicits feedback and session sharing, which I'm pretty sure some users oblige). Also CC does massive prompt caching tailored to work in lockstep with their platform. With all that and who knows what else, I'd say it balances over time.
10-20x more still needed, and then somewhere to build a couple DCs. Fingers crossed they make it, neither US nor Chinese labs can be trusted, even with open weights.
"Mistral raises €3B to make sovereign, open-weight AI the technology frontier". Oh yeah, that sounds reasonable, indeed, one receives this much money just for sovereignty's sake.
I’ve heard many folks get rejected and leave with a bad taste in their mouth. Interviews should make you feel like you’ve failed due to you not being quite there while still recommending friends to apply (“I didn’t make it but you should try to apply!” Vs “i felt like I had the privilege of even talking to the dude and then got ghosted”).
We're also seeing behemoths like Google and Meta both fail to keep up with OpenAI and Anthropic. Microsoft has given up on SOTA training. Loads of once promising LLM training companies are no longer relevant at the SOTA stage such as Cohere, Mistral.
I've said this countless times but SOTA LLM market looks like a classic monopoly/duopoly market over time. Each generation requires magnitudes more resources to train and you can only compete if you've made enough money on your previous generation.
Stay out of the race sounds insane when AI is, and will be, a common commodity. China is getting great bang for their buck on AI training, why not the same for Europe?
The two companies have read the market the same.
It's always very dangerous for first movers and their investors, and the commodification of intelligence seems even more likely each time a chinese open model release. It's less exciting to do business that way, but if you're building to stand the test of time, it's wiser that way.
Apple's strategy seems to be "wait till real business shakes out" but Mistral's strategy seems to be "go after profitable niches and avoid unwinnable fights".
I think what Mistral is doing is smart within their financial constraints, but this comparison is misleading. Mistral is an LLM company; Apple is a consumer hardware and services company.
It's smart for Apple not to join the LLM arms race, because they can just pick the cheapest supplier and let other companies take the financial losses. Mistral is in a very different situation; they are the supplier.
Apple is a $4.7 trillion company selling computers and iPhones. How many computers and iPhones is Mistral selling?
In your comparison Apple is Apple while Mistral is orange.
Mistral is not Apple, and is not emulating their strategy. Please show me Mistral’s half a $Trillion in yearly revenue coming from consumer hardware/software.
Then I’ll agree with you that they’re taking the Apple strategy.
First of all, they release the models' weights, perdonally I don't consider any other option as viable (no OpenAI and definitely no Anthropic, thank you).
I especially like their Vibe Chat web offer, the allowed monthly usage with a free account is incredibly generous (still have to hit a limit) and the deep research feature (5/month for free) is also valuable.
I don't know anything about the alleged regulation maxxing problems, I don't perceive them as a problem for my causal/personal usage anyway.
The gap has only been increasing, though. Devstral 2 was obviously not great compared to Claude/GPT but kind of acceptable if you were willing to compromise. There has been no real progress since then and frontier labs are massively ahead.
But I do care about capability and so far only Anthropic and, very recently with Astra, OpenAI can deliver on coding quality. And capability matters immensely. There is a world of difference between being able to do something and not being able.
AFAIK: Mistral does not even try to compete in this field. There are other use cases for LLMs beside coding. As Mistral AI wrote:
> During the first wave of generative AI, the central question was who could build the most powerful model. Organizations and governments are now asking a different one: how to harness the power of AI for their mission-critical needs without surrendering control over the infrastructure and intelligence loop. Demand for that combination of performance with control, choice and independence is growing internationally, as enterprises and governments weigh the long-term technology dependencies, data governance requirements and deployment choices that come with any AI investment.
> Mistral is the only AI company in the world building the full stack required to answer that question: open-weight models, the infrastructure and the compute capacity they run on, and the products that bring them into production; ensuring that customers are never locked into a single vendor's roadmap, pricing or availability.
> Mistral’s full-stack and open approach also allows organizations to build on it without exposing their most valuable data, workflows and institutional knowledge to anyone outside their own walls. That's what makes Mistral’s stack the sovereign AI layer, meaning retaining control across four dimensions: data that stays inside the organization's boundaries, models that are controllable and customizable, compute that is private and predictable, and systems in production that are fully controllable and auditable.
They have released models speficially for programming that ”vibe coding” would be safer.
https://mistral.ai/news/leanstral/
It's definitely a lot slower, though
There are things GPT 6.0 can accomplish for me that 5.3 was not able to. But there are also things it still fails at, and it doesn't seem to be much better at the big picture. It does spam about 100x more tests though and I wonder if just RLHFing it to test everything constantly is carrying it more. 6.0 writes so many tests and spends so much time verifying it's work in python sandboxes. Slow as hell but it tends to get things right the first time more which is good, I guess. I don't love the thought of a 500loc feature adding +4000loc due to tests though.
Even those old llama models were ok for coding
Yes yes they won't be like Claude's fire and forget (until you see how many tokens you burned to write "Hello World")
You must care about good benchmarks (identify those that have relevance).
My American bank sold my credit card transactions to advertisers. My American mobile operator had insane fees for roaming and other features that are basic in Europe. Sending a bank transfer in America was unreliable and slow and expensive because there was nothing like SEPA instant 24/7 free transfer (I guess FedWire does that nowadays, I don’t know if consumers actually have guaranteed access at all banks like they do in EU).
When American AI companies become established Fortune 100 members, they’ll start abusing their customers just like all the others in that club, those banks and phone operators and Microsoft and the rest. Google once pretended to be different, now they have the corporate cancer. No reason to believe the same won’t happen to OpenAI and Anthropic.
Even a lot of their attempts to increase competitiveness like forcing Apple to allow alternative app stores have been halfassed and not very effective.
Edit: forgot, they also call for banning Chinese models and open weights models...
Markets, left to their own devices, do not end up automatically being competitive in favor of consumers.
But above all, if AI wasn’t hyped as a threat to humanity, to jobs, to security, regulation could have been cautious.
To add insult to injury, given how voters are tired of technology, expecting a different move from governments is a losing bet.
True. Does not mean all regulation is increases competition and marker efficiency by default. If anything EU has long abandoned the core tenets of Ordoliberalism and "Social Market Economy"
I heard there used to be dozens of plane-building companies when the field was unregulated, with new, more efficient models and lower prices appearing regularly. I wonder what happened.
I kept hearing the same about previous age tech and there also Americans rule the Internet: Google Search, Workspace or m365, AWS, Cloudflare... Even Linus moved to the US.
Consumer rights - 14 day cooling off period for online purchases
Sale of goods - 2 year seller liability for defective goods
Right to repair - likely doesn’t need explanation, require manufacturers to offer repairs and replacement parts
Unfair commercial practices - prevent misleading or deceptive advertising
Unfair contract terms directive - prevents disproportionately one-sided provisions in T&Cs (think Disney giving themselves legal immunity for a free Disney+ trial)
The EU’s problem seems to be in communicating its benefits to its citizens.
We somehow need to "want" a slower news cycle and simultaneously push back against the algorithmic outrage. The world is more predictable and safer than ever but the perception of the average person is not this, it's manufactured outrage at every possible opportunity and topic.
Despite the average european objectively being better off and safer than a decade ago, almost none feel that way and that is almost 100% due to the news algorithms being tuned to keep you in a constant dread cycle.
Let's not forget I am paying directly for those when I buy a product. Those benefits are not free, they cost money. The money has to come from somewhere. It can't come from profits (since those are the sacred reason the company exists in the first place) thus they will come from higher prices.
Higher product prices mean less competitive companies. It also means fewer startups which must implement the required regulation no matter the market says. Then you need protectionism against Chinese competition.
But the worst is that I must pay for those benefits whether if I want them or not. Thus my choice is reduced. I cannot vote with my wallet and send price signals in the market to tell sellers that I want those benefits or not and at what price. The market gets corrupted. The economy falters. Outside competitors gain. The Government needs to intervene again and again, growing stronger and more intrusive "for my own benefit".
It's the decay spiral we are witnessing here in the EU right now.
As part of the AI Act, an important decision on a person (such as a hiring decision) cannot be made without human involvement.
You think that's not in favor of "consumers"? (also note the implicitly submissive framing of humans as receptacles for consumption instead of subjects with rights!)
which is a market Chinese labs won't get into.
they only other company they compete with is probably palantir in that regard.
So not only outpriced, outperformed, but also outregulated.
Idk i wish there was some bright light in the future, but i dont see it
We need a lot more hardware for universal LLMs (probably 10x, if not more), and we need it to be cheaper.
If anything, I would fund a DRAM maker in Europe.
I wouldn't. I can think of at least 10 regulations (labor and environmental chef amongst them) that would render such a maker uncompetitive from the start. Chinese makers would eat its lunch.
And with half the score of the top models, that's where Mistral is at.
As it is now, no tech company in the EU will use it, if they were just 80% from the top, that would be more feasible.
I’m surprised that there’s no domestic chip production either, we don’t have our own CPUs or GPUs, meanwhile China is spinning up manufacturing so they don’t have to work around the Nvidia export restrictions as much.
I like Mistral and there’s cool stuff going on like how EuroLLM models know Latvian language and all the other EU ones better than way bigger models, but we don’t have anything frontier.
At the very least, they should be distilling Kimi K3 and GLM 5.3 as much as possible and working on MoE models like ~35B and ~120B versions to match Qwen.
EU needs to meet their own demand, the US wants to be the global provider.
The second one is about which model to run and who controls and oversees quality control. OpenAI and Anthropic seem to insist that only they can do that. But of course here in the EU we see that a bit differently. The big US based hyper-scalers are neither liked nor trusted here at this point. We don't trust the Chinese model makers much either. But with open weight models, we can at least pick different models and run them on our own terms.
Also, what most companies need is not necessarily the latest fashionable model straight from the Silicon Valley cat walk but something that will work reliably and predictably for years. Factories are not going to install the latest model in their production lines every few weeks. Same with most banks, insurers, etc. I actually know people that do business with those in relation to AI development in Germany. Companies like that are very much obsessing about self hosting their models. Sending customer data off premises is a big concern for them. They are building stuff that will be used for many years. In five years, nobody will care which model was best in autumn of 2026. But a lot of software built this year that uses AI might still be running.
You have to see Mistral's investment in that context. They could make a lot of money in the EU if they do a decent enough job. Lots of conservative companies here that are going to pick something that's good enough and then they'll be using that for many years.
Every newly released model is paraded as SOTA showing peak or near peak performance on cherry-picked bendmarks the model was either fine-tuned on, or tested under specific conditions optimal for that model.
You crazy to think that Fable and Astra capabilities is fake
Whether AI is hosted by the USA, Europe or China - they all are awful and eliminating real jobs while also driving up RAM prices etc... Why should I want to support any of these?
Where in the fuck did we stall out?
Especially as Chinese models are getting better Mistral is getting less and less relevant.
It pains me because I want them to succeed, but despite them denying it I believe they'll end up restrict their activity to (1) selling hosting for Chinese models (they're already hosting GLM) and (2) selling AI-related consultant service (they're also doing that already).
So they’ll probably make more money creating PowerPoints with ChatGPT than they will trying to compete with the US and China.
Which would be the most European outcome ever.
Which is interesting since who are the investors and what exactly are their roles? Samsung - European? BlackRock - European? Salesforce Ventures - European? Etc.
So yes it might well be
> [...] the largest equity fundraising round ever completed by a European technology company, three years after the company's launch.
but the money isn't European.
Also, €3B is nothing in this market, especially when you're competing with more efficient competitors. €3B in Europe is probably the same as €10B in the USA and €20B or €30B in China.
I disagree on that point. Models are getting good enough that you can switch them and barely notice. I'm switching between Opus, GPT Codex and GLM 5.3 for coding and I can barely tell the difference.
I think they'll become more like telcos than anything, selling a commodity. It's even truer when any provider can host open weight models like GLM-5.3.
Basically a world with dozens of Baseten, with AI labs having a hard time monetizing, just like editors of open source software.
Mistral does not offer a model that makes sense to use. I understand they now host the already outdated GLM 5.2, courtesy of China providing the weights. And Mistral offers this for 2x-3x the price of other providers.
This is supposed to be a success story?
> so if you don't have the most intelligent or cheapest model in the world, you're losing
So how does this match up to the fact that there is currently OpenAI and Anthropic, both raking in money? They can't both have the smartest model at the same time, can they? And all those inference companies selling API access to open weight models on OpenRouter, which are apparently also earning billions already? While the former are probably bound to have much higher cost for research and training than they are currently earning, which may be called "losing", the latter don't have that problem, they can simply price their API access such that the money earned covers their costs, no training and practically no research necessary. In your theory these companies shouldn't have a cent of earnings.
They are simultaneously first: the two leapfrog each other with regularity, and are meaningfully ahead of the competition.
There is no close second, because the AA Index points are expentially harder to get as you get closer to the first.
They have been focusing largely on government and business not consumer, which is fine. Perhaps coding is not something they want to achieve, but they do provide Codestral. It's a signal it's a market of interest to them.
You'd have to be the producer of the machine producing the mask. Or, even better, the software that produces the plan according to which a machine produces a mask.
LLMs embody value systems, and American and European values are not the same (yes, there are overlaps, but also key differences).
More nefariously, I can also imagine LLMs that silently degrade their reasoning when used in a national security context of a non-US country.
So, Mistral may not be competitive with OpenAI and Anthropic, but in many contexts that doesn’t matter. And, perhaps this gap could be closed with more funding (the three billion funding figure is a rounding error next to US labs). I’m sort of surprised that the EU isn’t stepping in to support them.
It is! The second largest investor in this round is Scaleup Europe Fund:
https://eic.ec.europa.eu/eic-fund/scaleup-europe-fund_en
On a similar note: Why does it have to be the EU to step up?
Why doesn't EU rather speed up making VC investments more attractive, so that EU and banks don't do the majority of investing?
*I don't have answers to these questions. It just frustrates me how many investments here come from politicians and banks, rather than from investors, people, and companies.
There are also some areas where values across Europe are quite divergent. Think LGBT rights and social acceptance, religion & secularism, immigration & multiculturalism.
And countries don’t fall into neat “liberal west vs conservative east”. Spain is exceptionally liberal on LGBT issues while remaining more religious than some northern countries; Denmark is socially liberal but has adopted relatively restrictive immigration policies.
that is being replaced with Pax Silica
https://www.state.gov/pax-silica
Your values and American values might not be the same, but to say even most of Europe feels the same way is a big stretch. And not even the US has very many shared values anymore.
If we’re being honest, Europe doesn’t actually have much of a common value system outside of whatever is momentarily trendy in the urban monoculture, which is why it refuses to work together on most things and is currently being torn apart at the seams (see the rise of nationalist far right parties in most states).
The EU are a collection of states where the average citizen can’t even communicate with their neighbor in a common language beyond the level of a 4 year old. How could they possibly be aligned on a value system in the same way the US is.
Mistral's revenue is mostly B2B, and that's much more difficult to move.
Thats quite competitive for a base salary, as high as max ICT5 Staff base at Apple Munich.
Is that still a thing? Or have those would leave have left, after all, it's been ~6 years of Trump already.
Yes 90k is more than most French software companies pay, but the salary needs to be competitive with those US companies.
Germany is not so great either in that sense though at least in crypto sector, the pay was reasonable... Though as a job it was total BS.
UK was pretty good at some point (though all the really good jobs were gate-kept by IR35 regulations), The Netherlands was pretty good (especially contract roles thanks to big tax discounts for foreigners). But IMO, one of the best countries outside of the US for software work has been Australia. I actually came back to Australia because of it. Though it's not quite as lucrative today but it's certainly still much better paid than France and has potential.
That said, if you want to do your own startup and you do not have special business connections or pedigree, then the US is the only real option IMO.
I think it comes down to easy money. Easy come, easy go. To be rich enough to invest in startups in France, you need to be born into old money or be a complete freak of nature. These kinds of people don't part with their money easily. They need a guaranteed return and they can't rely on any media channel to bring traffic. If you live outside of the US, all media channels are essentially working against you; you can literally feel that and so everything is riskier. It's like playing the same game with a massive handicap.
Also, if given a choice of 90k eur in Paris vs let's say 2m usd in US I'm pretty sure suddenly most of the issues people have with US would suddenly disappear ;)
Mistral can't train on that data, because this data would be illegal to purchase & train on in the EU.
Culture.
In terms of output, EU working culture is inferior to American/Chinese working culture.
Work life balance is irrelevant for frontier/sovereign AI.
If you live in France, Germany, Estonia, UK, just walk outside and look at employees leaving their offices?
We need to start shaming individual countries in addition to blanket EU.
I'm not sure if your hypothesis is correct. I personally don't think the crazy work hours are what make the US so competitive in tech. But even if it were, it's odd to call for shaming countries for having less insane work cultures.
My US colleagues, like German one,just bullshit on reality of what they produce.
Real problem is more on VC, autonomy and small companies.
Way to reduce a multi dimensional concept into a single scalar value. Care to explain what exactly make it inferior? Try using more than one word if you can.
I think the most of the money would go to purchase hardware, But I hope they can start making adequate compensation for AI engineers and researchers to move forward.
> Being one year behind is a skill issue
E.g. if you are an AI researcher in Europe you can just go to USA and make generational wealth. This is not a criticism of the EU model, but rather insane American capex effectively monopolizing.
> I think the most of the money would go to purchase hardware,
So it's not a skill issue?
> So it's not a skill issue?
I meant by offering sovereign cloud/inference, not for training. But even if it was for training, Chinese labs have limited supply of GPUs, look what they've done. So it is a skill issue.
Also to clarify, I didn't mean European engineers' skills, I meant "you get what you pay for" as a company, that's why I hope they start offering better compensation to retain talent.
Handing in a photo-copy of a photo-copy of a photo-copy of a photo-copy of a photo-copy of a photo-copy of someone else's homework might work once or twice, but it is no way to run a (non grift) business.
I don't trust anything the antichrist invests in
Mistral had open-weights SOTA when Chinese labs had no noteworthy LLMs at all, and now even its closed-weights offerings are behind half a hundred Chinese models -- most of them from labs with way less funding -- despite having had unhindered access to Nvidia hardware all this time.
I know AI has more regulations here in the EU, but what I don't see clearly explained is what impact that has on pure research: creating new comparable models. So I'm suspicious that it's really impactful. On market deployment, maybe - that's where regulations lie. On internal company research?
But the US and China are incredibly lenient on enforcing even existing laws against their AI companies. In the US the standard seems to be "no instructions to make bombs or drugs, and the ability to make porn of living people has to be restricted to paid users". Laws around cyber security and copyright that used to be heavily enforced are seemingly suspended for AI companies, or settled out of court. No company that tries to actually follow the law can compete with that
It’s when you stack over 100,000 regulations (this is literally the number the EU has created) on top of each other and then combine them with country level regs that you create a system that’s impossible to grow in.
Multiple US labs were taken to court over copyright infringment for their training data. Mistral did the same data mining, but actually the copyright situation is weaker here so it's harder to build a case against them.
I think the capital situation in EU is just worse. A lot of is tied up in more conservative businesses that are reluctant to bet the house on some shiny new thing.
I miss when "reality is stranger than fiction" was a cool phrase because it didn't happen all the time.
Here's a nice summary. Haven't read the book yet though it's on my list.
a16z is Humpty Dumpty and Temu Lex Luthor.
The antichrist is Greta Thunberg.
How exactly does she threaten the continued thriving of our civilization?
It is fine for people to disagree with her worldviews, but labeling her a threat to civilization is an absurd claim.
A civilization that can be threatened by a single person's acts of peaceful protest might be doing something, but it sure isn't thriving.
Thiel called one of his companies Palantir which is weird too.
But, I guess some very religious people did not got all the subtleties of "Thou shall not kill" yet, do, let's give everyone another 2000 years to process the idea.
(Also, maybe it's hard for a camel, but easier in ocaml ? I'll have to try marketing "Static typing as a way to reach heaven")
He also expelled demons from a man and conceded to their request to be installed in 2,000 pigs. That led to them drowning, causing substantial destruction of somebody else’s property for no obvious practical necessity. The local people become frightened and ask Jesus to leave their region.
Taxation is not just about paying for the costs of the country, it is also about making sure that nobody has enough power to subjugate the will of the people. Monopolies, billionaires, the accumulation of such power is making the average citizen powerless. And you can feel it.
Taxation will not solve this before the global society collapses so the question is whether we figure out a way to level the power.
This is not evil, but a trait of human nature of corruption or addiction to power.
Taxation can and must solve it before "society collapses". And many people are working to solve the problem. It is difficult and requires effort but nothing that has not been done before. We got rid of king (by the will of god) in the past so techno-power-hungry cringey businessmen should be easier.
there is so such thing.
if there is it sure is very flexible and forgetful
An important factor is that fine tuning existing open models is incredible cheap. You can easily change any cultural biases if you want a model to be 'sovereign'. And Mistral could combine that with their custom data sets for their enterprise customer needs. Mistral still trains their own models, but they also seem to offer fine tuning existing models.
With model weights being commoditized, another differentiator could be deploying efficient inference chips, especially if you combine it with a developer ecosystem for vendor lock-in. That is why it is interesting that both Samsung and ASML are investors, since they are companies that could make a difference in this area.
It may make sense to train a frontier model on an existing architecture if the base model is not available and the instruction trained version doesn't fit with what you want. There are techniques like ablation, but those could have other effects on the model, and there can still be lingering effects of the instruction training in the model that surface less frequently (e.g. on an input not covered by the ablation training).
Otherwise, fine tuning is definitely the way to go. However, you need to be careful not to over-tune the model such that it is only tuned to the data you are training it on.
In the EU it seems we are very much at risk of being cut off from frontier AI if the US government should decide to do so.
Is that not self evident by the insane revenue from frontier labs?
https://isaiprofitable.com/
The only reason OpenAI and Anthropic are making a loss is due to training new models to keep up with competition - not because the industry is flawed in terms of business model.
Always sell shovels in a gold rush I guess
No...? Of course not?
Because revenue is only one side of the equation. Did you ever look at total cumulative OPEX and CAPEX, and how long it will take them to even just break even at current growth?
They're likely over $80b ARR by now. They'll be at $800b ARR next year at the same rate. Let's say their growth gets cut down to 3x instead of 10x - that's still $240b ARR by this time next year.
When you are growing so fast, you don't need to make a net profit. You just need to make sure your unit economics are good - which it seems like they are given reports that their gross margins are at 60-70%.
And they will be 800 trilion ARR in a couple of years, following that same rate! 8 quadrillion by 2029!
> When you are growing so fast, you don't need to make a net profit. You just need to make sure your unit economics are good - which it seems like they are given reports that their gross margins are at 60-70%.
If their margins were anywhere near this good, they wouldn't need to raise so much money so often.
If you create a machine that turns 1 dollar into 3 dollars, you don't dillute your ownership of the machine, you use your fabulous profits to expand your machine's capabilities.
Perhaps Mistral has too high moral standards to keep up?
They would if they could, which means they can‘t, even though they want to.
Seriously "thought to be" is such a baseless statement. Thought to be by whom? And on what basis?
The bubble is going to burst soon.
Maybe we need to abolish copyright?
I want my money back.
Not even a phone call :(
Anyway, the interview process was so janky it decreased my confidence in their success.
Not easy as it sounds yes, but it is better than throwing your CV into the ATS void 1000 times (the wrong way to apply for a job btw)
Mistral have no moat as it seems except some 'regulatory' moat.
The moat doesn't seem to exist. Probably safer to stay out of the race for now.
That paper released by a lone Google employee 3 years ago still misguiding people it seems.
We're also seeing behemoths like Google and Meta both fail to keep up with OpenAI and Anthropic. Microsoft has given up on SOTA training. Loads of once promising LLM training companies are no longer relevant at the SOTA stage such as Cohere, Mistral.
I've said this countless times but SOTA LLM market looks like a classic monopoly/duopoly market over time. Each generation requires magnitudes more resources to train and you can only compete if you've made enough money on your previous generation.