Sure, there is an AI bubble. But 13B valuation for a startup that is the best positioned to have access to most EU entreprise needs of AI, this does not feel overvalued to me. Mistral does not even need to produce leading models, they just need to produce good enough models that do not lag too much behind the top ones.
greatgib · 17h ago
That annoys me a lot that ASML that is quite a good stock with a good business dilute themselves with that.
Mistral might or might not be a good business in the future but that puts ASML subject to the potential AI bubble.
If they have too much cash at hand, and they wouldn't see any good use to invest on staying ahead in the technological race, it would be better to give back the money to shareholders. Then individual shareholders could decide if they want to put their extra money in Mistral or not.
Behind this, you can easily guess that some politician pressured the company to support the European AI unicorn. That is the European way, ruining what works well...
Ianjit · 5h ago
$1.5bn investment represents 0.5% of ASML's $300bn market capitalisation. The investment in Mistral is immaterial to ASML's investors if the investment goes to 0.
swores · 16h ago
Maybe the thinking is that if by investing in AI companies they can either increase the size/duration of the bubble (if it's a bubble), or speed up the growth of the industry (if it's not a bubble), that can lead to more investment in chips not just from companies they've directly invested in but from the wider scene, leading to enough extra profit for ASML to make the investments indirectly profitable?
If the company who sells the machines for making shovels to the shovel companies can invest in a single gold digging company to help convince twenty other gold diggers to buy more shovels...
Fade_Dance · 9h ago
My first thought was like OP - disappointingly off-brand from a purebred silicon lithography keystone.
Your scenario seems like a bit of a stretch, but maybe it could be distilled a bit. It could be an anchor of outright vertical integration. You have Open AI spending billions in Capex... and remember, they even tried to secure funding for fabs. NVIDIA has also done quite a bit of synergistic vertical integration to get to their trillion dollar status (one has to look within the data center and within the software stacks. It goes way back, back to things like the Infiniband acquisition). Taking all that into account, it's arguably not that absurd for ASML to push outside of the castle walls a bit with at least some feeler tentacles. Maybe someday ASML cutting edge equipment will be supplying a mega- European AI factory, where in an alternate reality they use slightly more economical single pattern EUV machines instead of double patterning, and ASML's stakeholder status helped make it so.
That almost seems distasteful to consider, but it almost seems like the way the world is going, with corporate interests quite obviously at the forefront of decision making. One would think ASML might even have some insight into this, considering their position in the world...
Sure, there is an AI bubble. But 13B valuation for a startup that is the best positioned to have access to most EU entreprise needs of AI, this does not feel overvalued to me. Mistral does not even need to produce leading models, they just need to produce good enough models that do not lag too much behind the top ones.
Mistral might or might not be a good business in the future but that puts ASML subject to the potential AI bubble.
If they have too much cash at hand, and they wouldn't see any good use to invest on staying ahead in the technological race, it would be better to give back the money to shareholders. Then individual shareholders could decide if they want to put their extra money in Mistral or not.
Behind this, you can easily guess that some politician pressured the company to support the European AI unicorn. That is the European way, ruining what works well...
If the company who sells the machines for making shovels to the shovel companies can invest in a single gold digging company to help convince twenty other gold diggers to buy more shovels...
Your scenario seems like a bit of a stretch, but maybe it could be distilled a bit. It could be an anchor of outright vertical integration. You have Open AI spending billions in Capex... and remember, they even tried to secure funding for fabs. NVIDIA has also done quite a bit of synergistic vertical integration to get to their trillion dollar status (one has to look within the data center and within the software stacks. It goes way back, back to things like the Infiniband acquisition). Taking all that into account, it's arguably not that absurd for ASML to push outside of the castle walls a bit with at least some feeler tentacles. Maybe someday ASML cutting edge equipment will be supplying a mega- European AI factory, where in an alternate reality they use slightly more economical single pattern EUV machines instead of double patterning, and ASML's stakeholder status helped make it so.
That almost seems distasteful to consider, but it almost seems like the way the world is going, with corporate interests quite obviously at the forefront of decision making. One would think ASML might even have some insight into this, considering their position in the world...
I thought the duty of a company is to maximise the return to shareholders.
Anyway, totally not a bubble.