Oct 26, 2026 · 50 min
Listen from 37:31
Listen at 37:31
AI prices fall as compute costs keep valuations under pressure
Can the AI Boom Survive a Margin Squeeze?
The episode tests whether frontier AI companies can sustain enormous infrastructure spending when model prices are dropping and demand remains uncertain.
- 1Falling token prices and elevated GPU rental costs could compress the margins supporting frontier-lab valuations.
- 2Forecast data-center shortages depend on future AI adoption, making infrastructure spending vulnerable to uncertain demand and agentic products.
- 3AI may transform society even if the current investment cycle ends in a boom, bust, or prolonged shakeout.
Don't miss
Steve Hou separates the question of whether AI is a bubble from the more consequential question of how long its boom-and-bust investment cycle will last.
The brief
Steve Hou of Silicon Data explains why AI infrastructure has moved from specialist concern to everyday conversation—and why its economics deserve closer scrutiny.
The central tension is stark: model prices are falling while GPU rental and compute costs remain high, putting revenue growth and frontier-lab valuations under pressure.
Hou describes tokens as the units models process, showing how longer conversations and retained context increase the computational bill behind apparently cheap AI.
The expected data-center shortage is partly a bet on future demand, especially from AI agents, whose adoption could reshape workloads without guaranteeing profitable usage.
The episode’s broader conclusion is less about calling a bubble than timing the investment cycle: AI may be transformative even if today’s capital spending disappoints.
The discussion ends with AI’s social costs, including job displacement, and the uneasy question of whether philanthropy can offset harms created by powerful technology.
What was said on this episode
23 statements · 9 positive · 10 negative · 2 mixed · 2 neutral
AI will transform society regardless of the investment cycle’s outcome.
“regardless of how this whole investment cycle transpires, going to be a transformative technology that's going to shape and change society and our world beyond recognition.”
Listen at 5:54
Compute futures could become one of the largest commodity futures markets.
“we have to believe that this will be one of the biggest, if not the biggest commodity rivaling oil futures.”
Listen at 9:09
AI model prices have nearly halved since midyear.
“since about the end of May, the token index or the expenditure-weighted effective... price paid for the models has been coming down pretty significantly, almost half of where it was back in the middle of the year.”
Listen at 10:09
GPU rental rates have risen rapidly and continuously.
“rental rates for these chips have been going up rapidly, continuously.”
Listen at 10:25
Slowing growth may raise questions about Anthropic and OpenAI IPO valuations.
“I suspect this is going to create some questions surrounding their upcoming IPOs and the, I think, high, shall we say, valuations, relatively high valuations.”
Listen at 13:06
Slower Anthropic and OpenAI growth could undermine their valuations and long-term expectations.
“there could be structural reasons for why, for competitive reasons that it may not grow as rapidly as people have previously thought, and that will put into question some of the valuations and rosier hopes for their long-term value.”
Listen at 14:04
AI has driven exponential demand for memory and data storage.
“the demand for memory and demand for storage of data has just gone exponential”
Listen at 23:03
Expected AI demand growth will exceed data-center supply growth.
“that would imply a level of demand that is not currently matched by the growth in supply. So to that extent, yes, there is a shortage in data centers.”
Listen at 25:09
People will use more AI as better-fit products proliferate.
“I do think people are going to be using more AI and it's going to proliferate”
Listen at 25:30
AI-agent overload problems will emerge in coming months.
“some of the questions and the pain points you're pointing out are going to actually surface in the coming months.”
Listen at 32:25
Restaurants may need AI-native operations to handle agentic requests.
“you've got to basically force a new type of restaurants that is AI native that can actually handle these things.”
Listen at 32:52
Investment and capital cycles invariably pass through boom and bust phases.
“every single investment cycle, every capital cycle has always gone through the boom and bust.”
Listen at 37:23
The AI market reached a local peak in midyear.
“The market has already... topped in a local fashion in the middle of the year”
Listen at 37:31
The AI market is feverish, with sophisticated investors reducing exposure.
“we are probably in a pretty feverish moment of the market that some of the smarter money has already either taken chips off the table”
Listen at 39:00
Even a bubble’s end could include violent counter-rallies.
“nothing ever happens in a straight line. Even if this was the end of the bubble, we can have extremely violent counter-rallies”
Listen at 39:19
The claim that AI is useless vaporware is fundamentally false.
“I think that is fundamentally untrue.”
Listen at 39:55
AI is transformative and will reshape society.
“I do think this is a fundamentally transformative technology. I do think it's going to reshape society.”
Listen at 40:08
AI demand may materialize more slowly than bullish forecasts assume.
“the demand could show up slower than some of the most bullish people in SF seem to think.”
Listen at 41:10
Supply may arrive before enterprise and consumer AI demand accelerates sufficiently.
“the supply could actually show up or show up in 2027, 2028. And then you find yourself not having as much demand because enterprise adoption wasn't as rapid and maybe consumer adoption is not as rapid”
Listen at 41:32
AI markets face a real risk of an adoption-driven air pocket.
“This risk of an air pocket is very much real”
Listen at 41:56
Strong AI outcomes could produce a violent market melt-up.
“there's also a risk of a right-tail risk by things that really do go right, in which case we can actually see a pretty violent melt-up as well.”
Listen at 42:02
Energy and oil companies provide investments outside directly AI-related technology.
“there are plenty of things that are not AI, directly AI related. You know, energy, you know, like traditional fossil energy stocks and, you know, oil companies, they are not AI”
Listen at 43:29
AI will transform society in beneficial and potentially harmful ways.
“I think AI is going to be hugely transformative in good and potentially bad ways as well.”
Listen at 45:34
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.