May 1, 2026 · 1h 21m
Hyperscalers risk overbuilding as OpenAI misses revenue targets
OpenAI Misses Targets, Codex vs Claude, Elon vs Sam Trial, Big Hyperscaler Beats, Peptide Craze
A massive mismatch is growing between capital expenditure on artificial intelligence and the actual revenue generated by leading AI startups.
- 1Hyperscalers are committing historic capital to AI infrastructure despite emerging signs of softening short-term demand.
- 2Vibecoding is lowering the barrier to software creation while introducing unpredictable risks for tech startups.
- 3The mainstream expansion of weight-loss peptides like Retatrutide is poised to reshape healthcare and consumer markets.
Don't miss
The debate over whether the 725 billion dollar hyperscaler spend will yield viable economic returns or end in a historic bubble.
The brief
Tech giants are pouring an unprecedented 725 billion dollars into capital expenditures, yet OpenAI is already missing its revenue targets, raising critical questions about the immediate financial returns on artificial intelligence.
The rise of AI-assisted programming, or vibecoding, allows founders to build software rapidly without deep technical skills, shifting the startup landscape but introducing new structural risks.
At the same time, the legal battle between Elon Musk and Sam Altman intensifies, exposing deep philosophical and operational rifts over the commercialization of generative AI.
Beyond software, the physical world is shifting as weight-loss peptides like Retatrutide enter the mainstream, signaling a massive looming disruption to global food and healthcare industries.
What was said on this episode
40 statements · 21 positive · 13 negative · 2 mixed · 4 neutral
ChatGPT 5.5 has recently outperformed Claude Opus 4.7 at the product level.
“if you just compare ChatGPT 5.5 to Opus 4.7, it does appear that OpenAI has had a better couple of weeks at a product level”
Listen at 5:54
Coding has become the most important AI application sector.
“coding has become the all important sector of AI”
Listen at 7:03
OpenAI currently has more compute capacity than Anthropic.
“they have more compute than anthropic right now”
Listen at 7:12
OpenAI's enterprise business is growing extremely rapidly.
“enterprise is going gangbusters”
Listen at 7:37
OpenAI and Anthropic are each worth multiple trillions of dollars.
“I think this is a multi trillion dollar company. I think Anthropic is a multi trillion dollar company.”
Listen at 8:30
Power availability constrains the entire AI market.
“Everything in this market is power constrained.”
Listen at 9:29
AI companies miss forecasts because power supply, not demand, limits token generation.
“The reason that these folks may miss a number or a forecast have nothing to do with demand. It is entirely 100% due to the supply of the power necessary to generate the output token.”
Listen at 9:32
AI power constraints will benefit major hyperscalers.
“It will benefit the hyperscalers, specifically Oracle, Amazon, Meta, Microsoft and Google.”
Listen at 10:46
Grok and SpaceX can exploit AI capacity shortages because they have excess capacity.
“that's a huge lane for Grok to just run through and SpaceX to run through because they have a ton of excess capacity”
Listen at 11:10
Forty percent of announced AI infrastructure projects will be canceled.
“40% of that is going to get canceled because they've done such a poor job of creating a good positive halo around AI”
Listen at 12:12
Pruning can shrink neural networks 90% without reducing accuracy.
“you could actually reduce the size of these networks by 90% and get the same accuracy out by pruning very large models down to smaller models”
Listen at 15:09
Neural-network pruning can reduce inference costs tenfold.
“by doing this you can actually reduce inference costs by 10x”
Listen at 15:23
Pruning could deliver ten times more AI usage from existing data-center energy capacity.
“you can make much more use, call it 10 times the use on data center and energy capacity than we can today”
Listen at 16:30
Google's consumer share gains caused OpenAI to miss user and revenue targets.
“if there is a single reason why OpenAI did not hit its user targets and its revenue targets certainly around consumer, you'd have to say it's because Google managed to take meaningful share”
Listen at 18:40
OpenAI Codex is currently gaining share of coding-token usage.
“I do think that Codex is taking share in coding tokens right now.”
Listen at 19:18
All frontier AI models will reach Mythos-level cyber capability within about six months.
“within six months or so, all the frontier models are going to have Mythos level cyber capability”
Listen at 20:57
Mythos and GPT-5.5 discover existing vulnerabilities rather than creating them.
“mythos or GPT 5.5, it doesn't create the vulnerabilities, it just discovers them”
Listen at 22:23
Organizations should use AI cyber capabilities to harden systems before attackers obtain them.
“As long as everyone does what they're supposed to do, which is use the new capabilities to harden their code bases and infrastructure and security before the hackers get a hold of these capabilities.”
Listen at 23:17
Machine-driven cyber operations will fundamentally change cybersecurity within five or six years.
“the nature of cyber is going to completely change probably in the next five or six years”
Listen at 25:30
Most operational software will eventually be rewritten, increasingly by machines.
“all the operational software that runs the world will get rewritten”
Listen at 25:53
Machine-generated software will increase cyber threats.
“the cyber threat actually will only increase”
Listen at 26:02
Hyperscalers will increase leverage and financial engineering as infrastructure spending rises.
“These companies will now get levered, they're going to get highly sophisticated around the financial engineering.”
Listen at 45:24
Investors should follow hyperscaler spending and buy the recipient companies because they are underpriced.
“I think it may be simpler and this is what I tweeted to just follow the dollars, like a trillion dollars a year going out of the hyperscalers. Where is it going? Just follow those dollars and buy those companies because those companies are already underpriced.”
Listen at 45:48
Hyperscaler earnings validated the bullish AI investment thesis.
“the bull thesis for AI just got validated in a single afternoon”
Listen at 47:25
AI agents will not replace humans entirely.
“I don't Believe that this is going to replace humans.”
Listen at 49:03
AI agents require human supervision.
“The agents have to be supervised.”
Listen at 49:29
The AI productivity boom will not cause massive job losses.
“I don't think there's going to be this huge job loss associated with this productivity boom that we're going to get.”
Listen at 50:07
AI systems remain unable to recognize the limits of their knowledge.
“AI still doesn't know what it doesn't know”
Listen at 55:53
Longer autonomous AI tasks are more likely to fail or drift.
“The longer the time horizon for a task, the more likely it is to go off the rails and adrift.”
Listen at 56:29
Predictions that AI will eliminate all software developers were excessively optimistic.
“this idea of eliminating all software developers was the peak of inflated expectations”
Listen at 56:40
Agentic coding is poorly suited to casually building and maintaining complex software.
“What it's less great for is casually building complex software that you have to maintain on an ongoing basis and take all the risks for upgrades, maintenance, keeping up to date with latest security issues.”
Listen at 57:27
Irresponsible vibe coding will cause firings and destroy some public companies' enterprise value.
“People will get fired because there will be some public companies where some goofball tries to vibe code their way out of something and they're going to torch the enterprise value.”
Listen at 57:49
Retatrutide reduced average A1C from 7.9% to 6% over 40 weeks.
“A1C drops from 7.9% to 6% in 40 weeks”
Listen at 1:00:02
Retatrutide users lost 37 pounds versus six pounds for placebo over 40 weeks.
“The average user in this phase three trial saw their weight decline from 214 pounds. They lost 37 pounds. That's compared to 6 pounds on placebo in 40 weeks.”
Listen at 1:00:15
A 2 mg retatrutide dose may preserve anti-inflammatory benefits.
“you could probably get this thing dosed down to 2mg and still see a lot of the anti inflammatory maintenance and other benefits”
Listen at 1:01:10
Retatrutide could receive FDA approval later in 2026 or sooner.
“could happen sooner, could happen sometime later this year”
Listen at 1:01:39
Retatrutide's glucagon activity favors fat burning over muscle loss.
“this glucagon agonist that they put into this retatrutide is it favors fat burning over muscle burning”
Listen at 1:03:06
The EPA determined that Roundup does not cause cancer.
“the EPA and their regulatory authority determined that Roundup does not cause cancer”
Listen at 1:10:30
Bayer has paid or reserved $20 billion for Roundup lawsuits, with 90,000 cases pending.
“To date, Bayer, which now owns Monsanto, has paid out $10 billion in these lawsuits, and they have reserved 10 billion on their balance sheet. And they have 90,000 cases still outstanding in the courts.”
Listen at 1:11:13
The Supreme Court Roundup case appeared evenly balanced between the parties.
“this could be a 50, 50 coin flip, 5, 4, either way”
Listen at 1:13:01
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.