134 comments

[ 5.0 ms ] story [ 214 ms ] thread
Love it as a consumer because Mistral is probably not doing any more open weights if they get tangled with Microsoft.

But man, as a business owner in EU, this would just make me not want to build an AI startup in EU

Personally, growing a startup enough to catch the eye of EU regulators is a problem that I’d like to have.
Doing all the work and taking all the stress only to have the EU remove a chance of growth and exit is a problem that I never want to have. I went through some failed startups, and at least that was our own fault, but still I never want to do that again. I don't know what I'd do in this situation.
Why is "exit" always the goal? Why not go on growing?
Perhaps the founders want to remain there and grow, perhaps they want to focus on something else or retire. Perhaps the plan always was to sell after X years - I don't see anything wrong about that, still better to innovate&sell than just buy stock&sell IMHO.

You can't attract investors if they can't sell in case the investment is successful. This is severely limiting and a terrible situation to be in - especially if you already have investors, who might demand their money back from you - way before your profits could cover it.

And the worst thing that might happen - you design a startup to be acquired by Microsoft/Apple/Google and then EU comes and says no. WTF?

(comment deleted)
(comment deleted)
Why would competition law and/or national security laws not apply to a start-up or its acquirers? Why would they or their acquirers get exceptions? Would those exceptions cover patents as well on the national security? What else and why?

Also, this isn't the EU saying "no" at the moment, but looking into it, as it does with many things.

They shouldn't get exceptions. There shouldn't be so many laws and regulations that result in EU being so behind economy-, success- and investment-wise.
Not sure why you think it is the number of laws and regulations that lead to that. The US has huge amounts of those, too.

Some structural differences are in part actually from the US having been less permissive than Europe in banking.

I think that’s the wrong question too

Even growth should be questioned holistically when you can literally make a few million dollars once and be done with it

instead of every quarter while pretending that making the same or lesser amounts are disasters

Few million dollars split by how many people? A small startup that provided a decent exit for me was sold for 50M EUR. I got 100K from that - the company had over 1000 shareholders. That's not really life changing amount of money, was barely enough to get a mortgage for my small apartment (30% of the price).

BTW the company was failing. If we didn't sell we'd have 0 very soon. It's really not as easy as "well let's just grow I guess".

Ideally just 1

If thats not your circumstance then it doesn't apply to you

I’m not the person that thinks there is anything wrong with selling shares, or failure, by the way

I would just rule out very many ideas to pursue at all, and encourage others to do the same.

They're not mutually exclusive.

But founders and employees at some point want to leave and do other things. They want to take their stake in the company and sell it to buy a house or retire or pay for their kids' college or whatever.

That requires an exit, either as an acquisition or an IPO.

(In theory, dividends or profit distributions are an alternative, but they're much harder to do anything with because they're mostly in the future and not guaranteed. You can't pay for your kids' college with future profit-sharing that may or may not materialize. And, dividends or profit distributions are the opposite of growth, because you're no longer re-investing in the company.)

Why is "growing" always the goal? Why not go on sustaining?
Because sustaining is a job, and it's a very hard job. The shareholders might not like that job or want to do a different one, or they just can't do it. Or the market is going down, but someone could use the IP/assets/team.
> Because sustaining is a job, and it's a very hard job.

... and infinite growth is neither of those?

Who said infinite? Startup founders that got investments have designed the exit points based on market conditions, and nobody would trust "infinite". It's just like stock investments (unless you "invest" based on feelings, of course) - you calculate the risk, the potential, and then you decide. Nobody reasonable puts "infinite growth" into their Excel sheets.

And yes, growing until that exit point is a lot of work - so it makes total sense people want an exit at some point.

The market expects infinite growth. Imagine what would happen if a publicly-traded company announced that they would stop pursuing growth.
No, that's not true. There are plenty of stable market capitalization, dividend-generating public companies. Their cap-to-EBITDA ratio is much smaller, but that doesn't change the fact.

Of course everybody expects the company to take advantage of opportunities - but it has to make economical sense. Nobody wants their company to try to grow so hard the whole company crashes.

Of course it doesn't expect infinite growth. Companies _aim_ for it, but nobody expects them to do it.
> Why is "growing" always the goal? Why not go on sustaining?

We're talking about a nascent field. Sustaining means holding the status quo while others lap you in capability and cost.

Because speculative startups require a lot of investment/venture capital, but the VCs need a defined return at some point. They need an exit with a sale, an acquisition or an IPO in order for the whole process to make financial sense. If companies just 'grow' but never sell out, there's not enough return for VCs, so they don't invest. No investment = way less startups to begin with
Especially important if you want to compete in training base models - if they were using 1500 H100s like the quote somewhere above implies, the market rate for that right now is something like $7500-$10k per hour (retail, on demand, so probably lower in reality, but still, it adds up incredibly quickly if you’re training 24 hours per day).
> Why is "exit" always the goal?

The goal should be whatever the creator wants.

(comment deleted)
(comment deleted)
It's tough, it's also part of the reason so little is done here, just compare the size of market capitalization of say DAX40 (German top 40) or CAC40 (French top 40):

DAX40 $1.9T

CAC40 $2.8T

There are single companies in US that exceed both of these values.

EURO STOXX 50 would probably be a better benchmark, which has a total market cap of around €4.0T, which is more than any single company. (your general point stands though, it's nowhere near the US market)
Part of that is that you're taking a fixed number of companies. So if Europe has more, but smaller companies, the value of the top-40 would be lower than if they had fewer but larger companies. To me the first sounds healthier, everything else being equal.
Also, lots of large European companies are privately held. For example: IKEA, Schwartz group (Kaufland, Lidl), Bosch, Aldi...
All 70 year old companies is not helping your point.

US has new upstarts like Tesla and Nvidia, which at their prime, add like 1 IKEA per fortnight to their market cap.

I'm not sure if "the market doesn't know how to price Tesla and is rapidly fluctuating up and down" is really saying much.

In terms of revenue Bosch is comparable to Nvidia and Tesla. IKEA is about half that and Schwarz is about Nvidia and Tesla combined, but of course those two are bad comparisons because they are in the rather unexciting retail market.

But generally I agree that exciting upstarts that reach fortune 500 size tend to be US companies.

That's because literally the entire world is betting on their technology right now. I can't imagine that being located in the US is the cause for Nvidia being in the right place at the right time with CUDA.
>That's because literally the entire world is betting on their technology right now. I can't imagine that being located in the US is the cause for Nvidia being in the right place at the right time with CUDA.

Ah yes, feel free to name all the other European startups that are in the right place at the right time with their own technology. Maybe one of these days you'll acknowledge that there may be a reason why such companies keep appearing in the US and East Asia, and not across the Atlantic.

Yes, I know that Nvidia is 30 years old. That does not change my point. Where were the European GPU startups back in the 1990s, let alone today?

(This is where you inevitably cite ASML. Psst: Look into where and who developed ASML's lithography technology, and what companies were not allowed to build off it.)

> feel free to name all the other European startups that are in the right place at the right time with their own technology.

For starters, there's ARM and STMicroelectronics. Then all the classical household-name businesses like SAAB, Phillips, Nokia, Siemens et. al. However, there's also more recent "SF-style" startups like Spotify, TransferWise and Klarna. Take your pick?

> Where were the European GPU startups back in the 1990s, let alone today?

Imagination Technologies was a pretty big name for a while: https://en.wikipedia.org/wiki/List_of_PowerVR_products

Most of ASML's technology was developed at ASML. Yes, not all, and there are US patents. But there was an insane amount of work they did in making it reality (together with Zeiss).
No, that's a good point. And no, I was not going to cite ASML.

My point is that maybe the US isn't better for causing wildly successful ideas to be rewarded by the market.

The US probably is better for starting tech companies. More rolls of the dice, more chances for success.

But it's not like Nvidia's founders were shopping around for a country to start Nvidia in either. If the Wikipedia history is to be believed, they were already accomplished engineers working at established computing companies in the US, so why would they found Nvidia anywhere else? The web of causality is deeply tangled here. Was the US in the 90s onward actually better for tech startups, or was it a momentum effect carried forward from the US already being a tech and computing leader in the 80s? Maybe a little of both.

Nvidia was founded 30 years ago so not exactly new either.
>All 70 year old companies is not helping your point.

I read an interesting point recently: Austria's 100 wealthiest families have two thirds of the country's wealth, and zero have earned their money from technology; they've all inherited it.

How different would the same list be for Germany? One, perhaps two families earned their fortunes from tech (and by "tech" I mean "SAP")?

Because the European capital markets are very different with far less local money going into it, more bank funding etc. The divergence between the US and Europe there is more than 100 years old.
I imagine there has to be some kind of security implication here, the same reason why the US is restricting the export of AI chipsets to china. It's not ideal if US megacorps are heavily investing in or buying out EU AI startups.
EU has to tech scene, so it doesn't want a potentially successful project to be transferred to US.

What is the most biggest EU software company? SAP?

That China has a more vibrant tech scene than Europe is a real shame.
And yet China's relative success is due to protectionism (or failed antitrust in USA/EU).

In a "winner takes all" situation, how else can the EU compete ? (The US did protectionism too while their industries were fledgling, to avoid getting crushed by the British...)

This is not a winner takes all situation at all.
Tech is like film. Big unicorns with big exits are like the blockbuster tent pole superhero films that net hundreds of millions of dollars. You don’t have those the whole industry collapses. No art flicks or risky experiments with new ideas without Guardians of the Galaxy.

It’s not just investment capital which usually grows out of exits but also the whole ecosystem of services and consultants and experts and law firms specializing in the sector. All of it.

The EU doesn’t have enough routes to exit due to both regulation and a way more conservative business climate. As a result you don’t get a big tech ecosystem and most exits result in the whole thing being pulled into the US orbit.

Regulations are not a problem, it's the lack of a big common market. Expanding from let's say Belgium to Germany is almost as difficult as expanding to the USA. That's why the two countries that have a big tech scenes (US/China) are the ones that have big unified markets.
That argues for more U in EU. Hard sell against a rising populist current though.
Regulation is what makes that hard.
Not really. Regulation serves the interest of the consumer, rather than the kind of deregulation that serves the interest of the shareholder.

In the US you build your app and everything is in English. You can launch that across all of the states as-is.

In Europe (not even the EU specifically) you’re gonna need i18n, particularly with B2B. If you want to launch in Germany, for example, you’ll be expected to provide an interface in German. You can’t presume everyone speaks English, because they don’t. Expect multilingual support too.

Then you have payment. Different countries have different cultural norms, so what people are used to using in the Netherlands is unlikely to be supported generically elsewhere. It’s not gonna be enough to just support Stripe and card payments or PayPal for B2C.

Regulation doesn’t mandate any of this. It’s effectively conducting business internationally.

SAP and Spotify come to mind, more recent promising startups are this one and Stable Diffusion.
Top companies from Stoxx 50:

  LVMH Moët Hennessy Louis Vuitton (422B Euro)
  ASML (347B)
  L'Oréal (241B)
  Hermès (240B)
  SAP  (203B)
  Totalenergies (142B)
  Siemens (140B)
For comparison, Spotify is only 50B.
These are top companies, not top software companies (even if we include Spotify).

So it's really SAP.

I mean, you do realize that there is more to an economy than software? Which is basically either retail or ad selling anyway?
The original comment was that EU sucks at software as a business that makes money.

The response was a list all companies, i.e. completely unrelated to original comment.

Edit: i see a typo in original comment 's/to/no/g'

isn't Stoxx 50 just the EU? i think you're missing some huge players (Shell, Novo Nordisk etc)

but to get back to the general point, that list looks kind of bleak to me. bags, makeup and oil. at least there's ASML in there (i hope they won't have a reason to move their listing to NYC).

> isn't Stoxx 50 just the EU? i

Not even that, it's just the Eurozone. So it's missing most of Scandinavia and the UK, which are probably the most 'techy' economies in Europe.

As someone who has spent the last 30 years in European tech companies, mostly startups, that is news to me.

Europe does not tend to produce as large tech startups as the US, in part because a lot of promising companies leave for Silicon Valley first chance they get because funding opportunities are better there. But then neither does most of the US, partly for the same reason.

We see the same all the way down - when I worked in a UK based technology VC, we saw both people decamping from the regions to London, and from London to SV for these kinds of reasons, as well as an influx of tech companies from all over elsewhere in Europe who saw London as one of a short list of European cities to move to in order to make things easier - some of whom undoubtedly would move on to SV.

I myself moved to the UK 24 years ago because I co-founded a startup in Norway, and the VC's that invested in us adviced us to move the company to London because there was insufficient capital for tech/software companies in Norway at the time.

Europe has plenty of tech. Tech is not all, or even mostly, the tech giants. It could have more, and it could put more effort into keeping more here. But regulations will not be enough for that - it'll take more capital infusion.

> But man, as a business owner in EU, this would just make me not want to build an AI startup in EU

Like, purely on the basis that MS might not be allowed to take you over?

TBH I'd assume the FTC would at least be asking questions here, too. EC and FTC thinking on competition is generally not all _that_ different, but the EC tends to be quicker to act these days (this was not always the case; the FTC was very much out in front on Intel and Internet Explorer back in the day).

You don't own what you can't sell. Any asset value will be massively reduced if its untransferable.

Mistral isn't exactly some decade-long state funded national champion. Its a 1 year old startup whose core staff all come from American tech companies. With 0 revenue and minimal market share, there's like no conceivable metric in traditional antitrust theory to block this. Yet the EU prevents it from taking foreign investment all the same.

Look at TSMC in comparison. It IS the national champion of Taiwan, yet its majority foreign owned, doesn't stop it from underpinning the Taiwan economy and spawning massive ecosystems in its local areas.. Taiwan's government understands its place in the international economy, so prefers to cooperate and trust its international partners.

It's not about antitrust. It's about national security.
This is the second problem with the EU.

They can't defend their own backyard, refuses to pay for their 2%, and free-rides off of US protection.

They tried to court China as their partner, only to now face a flood of Chinese car imports that outcompete their once prided industries.

They seem to only care about national security when it comes to fleecing US companies, rather than say defending against Russia, or strategizing against China.

> They seem to only care about national security when it comes to fleecing US companies

Stop with that nonsense. Every single country that imports American goods has regulations on how it is imported. Your expectations are misaligned if you think the concept of a Free Market extends to other countries carte-blanche. Not even the United States can stop itself from regulating it's own businesses, and usually it's for the better.

These comments pop-up all the time but never have a reasonable alternative to suggest. This is what governance looks like; I'm sorry if you're offended on Apple or Microsoft's behalf.

You would have an anti-trust problem if somebody went out and bought all of the promising AI startups.

So at what point in their buying spree do you cut them off? Do you wait until they've bought 90% of AI startups? 50%? Does it matter if there are 4 AI startups or 40?

It starts to matter when the broad public is adversely affected by the lack of AI diversity. I don’t think we’re there yet, let them buy.
From my perspective, the only name and products people (who are not here) know are ChatGPT/OpenAI API based.

Do you have a different experience?

The problem here is that, if you can see the adverse effects on the public, you're too late. So you need to anticipate them. That's not an easy thing to do.
So, pre-crime?

It's not illegal to be a monopoly[0] (at least in the US) but abusing your monopoly is what gets you in trouble. Copyright and patents are examples of gov't granted monopolies.

[0] Fully conceding the fact that MS is a convicted monopolist and has to play by different rules than everyone else.

(comment deleted)
> Yet the EU prevents it from taking foreign investment all the same.

No they haven't.

True. In Kafkaesque sense they haven't. They are just checking things are in order.
I mean, they're looking into the situation where MS keeps ~taking over AI startups. I'd be quite surprised if the FTC is not at some level also looking into this. Like, it would be weird if they weren't! At this point, no-one's telling MS it's not allowed to do this, but it would be strange if no-one was taking note of it.
This move from MSFT is clearly perceived as aggressive from EU given how fragile EU sovereignty in AI is at this point.

If EU wants to own its AI future, that’s not a great way of getting started.

I thought EU had money a lot more than €15M (Microsoft investment), actually a lot more than €1.8B (Mistral valuation). Why didn't EU make an offer?

If EU wants to own its AI future, it has enough money to buy one. They decided not to.

The EU generally doesn't act on the market like that. This isn't China or Soviet Russia after all, we are still operating under a form of capitalism. The EU does offer grants, but those are generally meant for projects that wouldn't get money on the free market
Is it just me that think that’s a huge catch-22? The project isn’t eligible for grants because it gets money on the free market, but is then limited by who it can come from or in what amount
Isn't Norway operating under a form of capitalism as well as owning controlling share of Statoil? Why mention China or Russia?

I guess Norway is not a member of EU.

There is also the German state of Lower Saxony holding 12% of Volkswagen, and of course lots of countries fully or partially owning technically private organizations in infrastructure, transportation and communication. So "generally" is doing some lifting here. It's not that this never happens, but it is the exception. It would be highly unusual for the EU to just directly invest in companies like that.
Norway retaining a controlling share in Statoil, now Equinor, is not quite the same as buying a startup or indeed buying any other company.

Statoil was created by the state not bought by it. If anything Statoil is an example of the very opposite idea and that is that Norway has sold shares in a company set up by the state rather than bought shares in an already existing company.

The EU is not a state. There is no possible way of having something like nationalisation at the EU level. So whilst nationalised companies exist in different member-states (almost all of them, I would think), the EU still cannot do that.
> EU generally doesn't act on the market like that

It absolutely does. The problem is it must then partition the work across the entire continent. America has this problem, too, e.g. the F-35, but the consensus model of EU politics makes it particularly insidious.

No AI capable EU native cloud provider exists. Everything in Europe runs on Azure or Amazon.

Also, EU doesn't but things for their consumers, do you even know what the EU is?

Time will be the ultimate judge but Mistral's Co-founder/CEO is still committing to be a leading provider of open weight models [1]:

> Clarifying a couple of things since we’re reading creative interpretations of our latest announcements:

- We’re still committed to leading open-weight models! We ask for a little patience, 1.5k H100s only got us that far.

- We have a reselling agreement with Microsoft, that we’re very excited about. Alongside similar partnerships, it will accelerate our growth.

- Microsoft invested in a small convertible note alongside many other companies, as a distribution partner. We are an independent European company with global ambitions, that part is not changing either.

> We’re seeing some interest for Le Chat and Mistral Large, on both la Plateforme and Azure, and we’ll be iterating fast!

[1] https://twitter.com/arthurmensch/status/1762818733016322168

Statements like these are only a momentary snapshot. As soon as sufficiently money is at stake the board might change their tune overnight. The whole Altman-fired-from-OpenAI-then-reinstated drama should be warning enough.
What do you think they got Sam to bend the knee on, out of curiosity?
In my opinion, Altman's permanent departure would have been the death of OpenAI. After the majority of the staff threatened their resignation from OpenAI, Altman became the one holding all the leverage. I'd rather guess that OpenAI's board made him a once-in-a-lifetime offer to persuade him to come back.
I mean, the FTC would absolutely look into this agreement if you did it in the US. Regulators "looking into" things is not a thing we should be upset about, that's why they exist.
And not only this, everyone loses if the FTC and the likes don't do their job properly. I'm still amazed how much "small" people argue for big entities like microsoft. You won't have these problems in your lifetime.
> But man, as a business owner in EU, this would just make me not want to build an AI startup in EU

If a Chinese company with massive government contracts had invested in OpenAI there is no way that the US would not start an investigation.

Microsoft has massive government contracts in literally every EU country. US is an ally, not antagonist/enemy.
until they elect a president that throws allies under the bus (I.e. NATO, NAFTA, etc) and gives state secrets to enemies
>until they elect a president that throws allies under the bus (I.e. NATO, NAFTA, etc)

Based on recent news, Trump has done more for NATO rearmament than any US president in history. Gee, maybe sometimes drastic steps are needed when 20 years of three other presidents doing everything but stand on their heads while begging other NATO members to spend more on their own defense failed.

Trump was lambasted by the usual bien-pensant suspects when he told NATO this in 2018 <https://www.youtube.com/watch?v=Vpwkdmwui3k>. He was right when at the UN he warned against Western European dependence on Russian energy, for which the German foreign minister and others laughed at him <https://www.youtube.com/watch?v=FfJv9QYrlwg>.

And remember, Trump isn't even in office now; he's already done more for Western collective defense than a) the putative current president and b) an actual land war in Europe!

>and gives state secrets to enemies

Repeat this often enough and maybe it'll come true one day!

> Trump has done more for NATO rearmament than any US president in history

Putin did that. To the degree Trump's statements have had an effect, it's in deharmonizing trans-Atlantic military coöperation.

(EDIT: Trump and Putin had similar effects. See [1].)

[1] https://news.ycombinator.com/item?id=39540275

>Putin did that.

No. As I said, not even an actual land war in Europe has been as effective as Trump's mean words in getting Western Europe to actually get off their butts regarding defense.

(comment deleted)
You said that, but it does not make it true. Trump rocked the boat but did fuck all.

The truth is that for many European countries military defence is was a bit of an abstract concept, with no real enemy nearby and no really interventionist foreign policies. Letting NATO rot was not a big problem in that context. What changed everything was when an enemy willing to use its military in a war in Europe materialised.

> not even an actual land war in Europe has been as effective as Trump's mean words in getting Western Europe to actually get off their butts regarding defense

Putin and Trump appeared to have played similar roles in increasing European defence spending, with Putin beating out Trump (but not by much).

Let's look at Ex-US NATO defence expenditure as a fraction of GDP from 2017 to 2020 in comparison with 2014 to 2016 [1]. On average it rose +13 bps, 1.43% to 1.56%, (p = 0.09, unpaired t-test). Not statisfically significant. Same for '17 to '20 versus '21 to '23: +12 bps, 1.56% to 1.69% (p = 0.13).

Looking broader, '14 to '16 versus '17 to '23, we get +18.6 bps (p = 0.02) for Trump. Significant! But if we do the same for Putin, '14 to '20 versus '21 to '23, we get +18.1 bps (p = 0.02). Also significant! (And uncannily similar.)

Seeing the above, I get why there is an impasse. The data support an insignificant conclusion for giving the entire effect to Trump or Putin with low confidence. They more significantly support giving half the effect to each. Given we're "crediting" (in the high-confidence analysis) '21 to '23 to both, and Putin's effects have manifested over just 2 years, that makes the claim that Trump did more to boost ex-US NATO defence spending than Putin technically incorrect. But only technically.

(Would note that a t-test for such data isn't really correct, so would take the above as casual calculations.)

[1] https://www.nato.int/nato_static_fl2014/assets/pdf/2023/7/pd... Table 3

"There are no permanent friends, only permanent interests."
Didn't US/EU have a trade war just ~3 years ago?
They’ve had trade wars for much of recent history, but don’t read too much into that. Allies can still disagree on matters of trade.
Maybe ally/enemy is a spectrum and also not one-dimensional. "Not ally" does not imply enemy and my military ally might not be my ally in trade. Also most US allies have learned to not read too much into the word ally. The US makes for a rather fickle and sometimes insidious partner.

Trust in the US on military matters is rather high (despite some recent blunders), but in every other arena it is basically nonexistent. I don't see why the US should receive any less scrutiny than China in this instance.

I entirely agree.

I’d add that as long as NATO was uncontroversial in American politics, allies could count on being able to work with each successive president, regardless of party, which definitely did help. There was the occasional bickering (like the French being vocal in their opposition to the war in Iraq), but nothing that would compromise the alliance in the long run. Things like “friendly” intelligence gathering, or indeed trade disputes, never stopped, but did not preclude cooperation at the highest levels.

In that context, it’s fine to argue every now and then because everyone knows that neither side is likely to escalate things out of control. As you said, it did not affect trust in the fundamentals.

Is it? From some points of view MS is "compromised" and potentially not an ally of the West, any more.
… for now. Trump is very much antagonistic and has a good shot ant being the next president. Even if he is not, considering current American politics, whoever gets elected in 2028 could very well be worse than him in terms of working with allies. Recent history taught us to not take some things for granted.
> history taught us to not take some things for granted

Huge difference between not taking someone for granted and being antagonistic towards them.

I don’t see any country in the EU being antagonistic to the US, really. Even nationalist parties are enamoured with American white nationalists.
You think there even would be an investigation/probe into it? Perhaps a really quick one.
> If a Chinese company with massive government contracts had invested in OpenAI there is no way that the US would not start an investigation

Sure, but if a European did we wouldn't.

So if you start in America, you can access both continents. If you start in China, you can access Europe and Asia. If you start in Europe, you're limited to one.

Granted, this is just a proble. But if it amounts to something, it's an asymmetrically chilling move.

Pretty sure the US explicitly restricts investment in Chinese AI companies now. The US has learnt its lesson investing in China too.
> the US explicitly restricts investment in Chinese AI companies now

Congress seeks to curb it but doesn't in practice. That may change. In any case, I was discounting access to American capital market for Chinese AI start-ups: they can tap both China and Europe.

The current lack of anti-monopoly enforcement and corporate oversight in the US is a race to the bottom. Should the rest of the world compete by letting corporations run amok, acquire all competition and build monopolies? Or should they try to regulate, but risk driving away founders for whom the American legal environment is extremely lucrative?

It's similar to the debate France had about raising the progressive tax rates - some ultrarich people would rather give up their French residency, than pay higher taxes.

There's clearly stuff the EU could do to be more competitive, but I think the core issue is that American investors can expect less risk and bigger payoffs, because the environment is friendlier to them, but the environment being friendlier to them leads to regular people getting screwed by giant companies that operate with no oversight. Fast forward the 2020s and you literally have planes falling out of the sky.

I don't think there are any great options.

I'm also aware this is a fringe view on HN, kinda by definition. It's a forum for founders, mostly in the US, and the median opinion on what you should be able to do with your company on here is extreme by most peoples' standards. But the echo chamber needs some dissent on this point.

> Love it as a consumer because Mistral is probably not doing any more open weights if they get tangled with Microsoft

The solution is to mandate open weights. Not kneecap your flagship by putting it at a permanent capital disadvantage.

Totally agree with this sentiment. I love open-source and will be on it's side to the end. But I'm not sure I want it forced upon people either. Companies should believe in the model and developers should choose companies that believe in that model.
> But man, as a business owner in EU, this would just make me not want to build an AI startup in EU

I am not an expert but does this kind of probes not help businesses against things like hostile take overs and against VC buying up all your competitors to crush you with scale?

15 million in Mistral vs 13 billion in OpenAI. There must be a different reason at play here.
if you can handicap a competitor for 15 million to save your 13 billion investment, go for it

There is a reason EU is investigating this, and this is exactly it. They are sick of the Microsoft extinguish aspect

To be frank, I am more worried about the fact that they raised $385M from US investors at the seed stage. That speaks volume about how much it really is an EU company. Good luck to EU for retaking control of this ship.
Damn, that’s an interesting point
385M USD is way too much for EU investors. They usually invest 10x-100x less. So the alternative without US investors would be to have no startup of this kind - great job, that will really help EU grow!
What makes US investors so uniquely ambitious and generous? I'm starting to appreciate the US financial sector more and more.
The US investors just have so much more disposable cash they can risk. EU investors are looking for safety because losing 100M EUR means they lost literally everything. Their "risky volatile investments" part of the portfolio is 1-5M EUR in total, usually even less - and split that between at least 3-10 startups. And there's much less people with money overall, so you can't raise large enough VC funds.

The usual EU investor is simply going to buy some real estate instead of participating in VC funds - still great returns, very low risk. The more progressive ones will maybe try some real estate development projects. No need to think about this software thing. The angel investors I met with that had portfolios of 5-20M EUR wanted to invest 20-50K EUR into our startup, definitely not more.

The US is a very rich country.
Pension funds. European pensions are run as insurances and/or financed via taxes. Thus, there is no investment capital from pensions. The U.S. is quite the opposite, and that fact alone probably explains > 60% of the gap in VC financing.
some people just use the governmental pensions (like in Germany) which aren't necessarily from a fund so they aren't funds at the beginning. I heard France just aloud pensions funds to use for VC financing.

My opinion this is one of the big reasons for the lack of strong startup investments. The other one is the lacking financial union.

They have lots more money. Which probably originates to climatic and geographic reasons, but more recently due to the _massive_ destruction of wealth in the world wars. It's very hard to build up a wealthy middle class when every few decades millions die on the battlefield.
Let’s fix the root cause instead of leaning in to a subpar EU investment funnel.

EU has billions to invest in AI. I don’t buy that EU has “much less money than the US”.

The EU doesn't invest, the EU provides subsidies with many strings attached. One of the strings usually is "can't be used to pay shareholders" (as in wages, not just dividends). Also, the subsidies are usually provided after the fact (e.g. a project is completed) and you need a loan to actually do the project. That's a risk, and banks don't like providing loans to small companies without predictable cashflow; they would do it only if a shareholder with enough assets personally underwrites it - another huge risk, now a startup can destroy your whole life.

If you meant the investors of EU - please show me where you got the data, because the official EU data suggest otherwise. People simply aren't that rich around here, which is a downstream effect of the much slower, smaller economy.

Though both areas are very rich, there’s a lot less total wealth in the EU than in the US. Due to taxes, this is especially true for rich Europeans vs rich Americans.

In the below report, it looks like Europe with it’s much larger population clocks in at a total wealth of 110T, while the USA + Canada has 150T. That’s 600k per adult in N.A. vs 200k per adult in Europe.

More specifically, it looks like 1%ers in the US are worth USD 7m, but in Europe the 1%ers are worth USD 1.5m.

I’m not denigrating Europe or the European way, just pointing out differences in the capital environment.

[page 142, 144 of https://www.ubs.com/global/en/family-office-uhnw/reports/glo...]

Microsoft only invested $16M....lol.....I thought Mistral would demand more but I guess partnership with Microsoft is more valuable than cash.
To be honest. $16M is a huge amount of money for most companies based in Europe.