Torsten Slok, chief economist at Apollo Global Management, recently argued that the stock market currently overvalues a handful of tech giants – including Nvidia and Microsoft –...
The problem with this take is that most of the “AI boom” research and development is happening in house at firms that do generate unfathomable amounts of money. Google is developing its own AI funded by its insane ad profits, Facebook is doing the same thing. OpenAI has massive investors with fuck tons of money to burn, including Masayoshi Son’s Softbank Investment Fund that is really just a front for the Saudi sovereign wealth fund, and Claude/Anthropic has huge Amazon investments. The bubble won’t “pop” because unlike the dot-com bubble these companies aren’t propped up by massive valuations on the stock market that can be destroyed at the slightest shift in investor sentiment. Instead, their propped up by cash hoards of some of the largest and most profitable firms in the history of humanity. Two of the biggest players in this field are even private, entirely unconnected from stock valuations.
Bigger players will obviously survive, and that was the case during dotCom bubble as well. However, there are tons of small startups funded by VC money that are grifting all kinds of AI based solutions right now. That’s where the actual bubble is. Also, it’s not at all clear that the business model the large players have is going to make sense long term. Google, OpenAI, and Anthropic are all banking on AI as a service business model. However, open source players are rapidly catching up and are now even outperforming them as seen with recent Kimi 2 release https://venturebeat.com/ai/moonshot-ais-kimi-k2-outperforms-gpt-4-in-key-benchmarks-and-its-free/
Chinese companies took a completely different approach to AI, where models are released as open source and treated as a common foundation for building things on top of. I expect that Chinese approach will win in the end because it invites collaboration while amortizing research and development costs. Open models will inevitably be setting the standards globally, making it increasingly harder for closed models to compete.
Ah I see what you mean, yeah basically every AI company that is not developing an actual model is 100% fucked in the long wrong, because their entire business model is essentially “we make a prompt around ChatGPT API calls” which is not sustainable as soon as pricing for APIs starts going way up. I also think that the larger players banking on AI as a service business model is not going to work out for the reasons you mentioned (Deepseek R1 alone, which I can run on like a normal laptop, is already fantastic) BUT even if none of that doesn’t work out Google/Facebook/Microsoft/Apple et al. are going to be fine, because their revenue is not actually from AI at all. There’s no bubble to pop there.
The problem with this take is that most of the “AI boom” research and development is happening in house at firms that do generate unfathomable amounts of money. Google is developing its own AI funded by its insane ad profits, Facebook is doing the same thing. OpenAI has massive investors with fuck tons of money to burn, including Masayoshi Son’s Softbank Investment Fund that is really just a front for the Saudi sovereign wealth fund, and Claude/Anthropic has huge Amazon investments. The bubble won’t “pop” because unlike the dot-com bubble these companies aren’t propped up by massive valuations on the stock market that can be destroyed at the slightest shift in investor sentiment. Instead, their propped up by cash hoards of some of the largest and most profitable firms in the history of humanity. Two of the biggest players in this field are even private, entirely unconnected from stock valuations.
Bigger players will obviously survive, and that was the case during dotCom bubble as well. However, there are tons of small startups funded by VC money that are grifting all kinds of AI based solutions right now. That’s where the actual bubble is. Also, it’s not at all clear that the business model the large players have is going to make sense long term. Google, OpenAI, and Anthropic are all banking on AI as a service business model. However, open source players are rapidly catching up and are now even outperforming them as seen with recent Kimi 2 release https://venturebeat.com/ai/moonshot-ais-kimi-k2-outperforms-gpt-4-in-key-benchmarks-and-its-free/
Chinese companies took a completely different approach to AI, where models are released as open source and treated as a common foundation for building things on top of. I expect that Chinese approach will win in the end because it invites collaboration while amortizing research and development costs. Open models will inevitably be setting the standards globally, making it increasingly harder for closed models to compete.
Ah I see what you mean, yeah basically every AI company that is not developing an actual model is 100% fucked in the long wrong, because their entire business model is essentially “we make a prompt around ChatGPT API calls” which is not sustainable as soon as pricing for APIs starts going way up. I also think that the larger players banking on AI as a service business model is not going to work out for the reasons you mentioned (Deepseek R1 alone, which I can run on like a normal laptop, is already fantastic) BUT even if none of that doesn’t work out Google/Facebook/Microsoft/Apple et al. are going to be fine, because their revenue is not actually from AI at all. There’s no bubble to pop there.
I agree, the big corps are gonna be fine in the end. They’ll probably also gobble up any startups that end up producing genuinely useful stuff.