Aug 27, 2026 · 2h 28m
Tech critic Ed Zitron warns of a catastrophic AI market collapse by 2027
The Man Who Calls BS On AI: AI Is The World’s Greatest SCAM, And They All Know It! | Ed Zitron
As trillions of dollars pour into artificial intelligence, this episode challenges the industry's growth narrative and outlines the severe economic and environmental risks of a looming tech collapse.
- 1Generative AI business models are fundamentally unprofitable due to the massive computing costs of data centers and unsustainable flat-rate subscriptions.
- 2The rapid integration of AI into everyday software is creating buggy code, declining software stability, and increased tech industry downtime.
- 3A major financial correction is looming by 2027 as tech startups deplete their subsidized funding and trigger a broader market downturn.
Don't miss
Ed Zitron predicts a cascading global tech depression starting by 2027, triggered by OpenAI running out of cash.
The brief
Tech critic Ed Zitron argues that the generative AI boom is an unsustainable financial bubble, labeling the massive industry push as a non-consensual technology rollout driven by disingenuous marketing rather than real consumer demand.
Zitron points out that major tech firms are burning billions of dollars on data centers to subsidize unprofitable startups, warning that flat-rate subscription models cannot cover the massive computing costs required to run these models.
Unlike the dot-com bubble, which built foundational internet infrastructure, Zitron contends the current AI trend is creating a rot economy characterized by buggy software, increased tech downtime, and severe environmental strains on local energy grids.
The critical turning point will arrive by 2027, Zitron predicts, when major players like OpenAI run out of cash, triggering a cascading tech depression that will heavily impact retail investors, retirements, and the broader global economy.
To prepare for the impending market correction, Zitron advises everyday investors to take a highly conservative financial approach, avoid overhyped tech stocks, hold cash, and focus on human community over digital noise.
What was said on this episode
31 statements · 2 positive · 27 negative · 1 mixed · 1 neutral
Generative AI is fundamentally deceptive.
“I think generative AI is at its heart, con.”
Listen at 0:00
AI companies overstate capabilities, future performance, and financials.
“I think that they overstate both what it can do, what it will do, and the underlying financials to the point that they are misleading the entire world”
Listen at 2:39
Generative AI is expensive, unprofitable, and fundamentally unreliable cloud software.
“it's boring cloud software that's extremely expensive and unprofitable and also unreliable at its core.”
Listen at 3:10
OpenAI and Anthropic generate 70% of AI revenues across the cited companies.
“70% of all AI revenues across those three companies are from OpenAI and Anthropic.”
Listen at 4:25
OpenAI and Anthropic cannot survive without funding from major technology companies.
“Two unprofitable, unsustainable companies that literally cannot afford to exist without these very same companies giving them money.”
Listen at 4:30
AI companies have spent over one trillion dollars and plan another trillion next year.
“they've spent over a trillion dollars so far and they want to spend another trillion dollars next year.”
Listen at 7:30
Generative AI's current value does not justify trillion-dollar investment.
“That's not worth a trillion dollars. None of it is.”
Listen at 9:13
AI companies operate at very large losses.
“all of these companies run at a horrifying loss.”
Listen at 13:04
AI subscriptions sell usage costing substantially more than customers pay.
“they would get demand selling $20 or $40 for a dollar.”
Listen at 15:06
AI inference is becoming more expensive, and inference providers appear unprofitable.
“It seems to be getting more expensive. In fact, everyone inference providers don't seem to be profitable.”
Listen at 16:28
Software quality is declining as LLM use and mandated adoption increase.
“The quality of software is going down, weirdly enough, as More people use LLMs and more businesses demand, and I really do mean demand, that people use these services.”
Listen at 23:53
AI developers have not eliminated hallucinations.
“And they've certainly not succeeded.”
Listen at 26:05
A cited AI task benchmark shows only 50% successful completion.
“successfully completing them 50% of the time.”
Listen at 26:31
Generative AI would be more defensible if sold at its actual cost.
“If this was sold at its honest cost, yeah, I would actually.”
Listen at 39:26
Major AI hardware efforts have not made generative AI profitable or cheaper.
“no one seems to magically be able to make this profitable or indeed even less costly.”
Listen at 46:57
Autonomous vehicles should be deployed slowly and treated as unsafe until proven otherwise.
“we need to be so, so, so careful and treat them as guilty till proven innocent”
Listen at 1:03:22
There is no evidence that generative AI produces productivity gains.
“there is no evidence of, like, productivity gains.”
Listen at 1:06:56
Agentic AI is primarily LLMs operating with additional orchestration, not autonomous intelligence.
“Agentic AI is one of the bigger lies they tell. It's like when you hear agent, you're meant to think autonomous AI can do what you want. It's still LLMs.”
Listen at 1:23:11
LLMs are particularly useful for troubleshooting from technical logs.
“The troubleshooting thing is awesome for it. It's the. The one weakness I have. It's like genuinely being able to drop a log into it. That's awesome.”
Listen at 1:31:30
Generative AI improvement faces hard technical limits.
“I actually don't think it. I think that there are hard limits that we're going to.”
Listen at 1:36:15
Generative AI has already reached diminishing returns.
“We've kind of already hit the diminishing returns level”
Listen at 1:36:21
A roughly thousandfold cost reduction would be needed to change his assessment of AI.
“There would need to be a hardware breakthrough that reduced the cost by like a thousand.”
Listen at 1:57:36
AI data-center power infrastructure is reckless and damages local communities.
“I think the use of gas turbines is fucking disgraceful. The water situation, I'm not super well read on. So I'm not going to wade into it. But the use of gas turbines and behind the beat of power is reckless and damaging to communities.”
Listen at 1:58:04
Claims that AI will become conscious or generally superintelligent lack proof.
“These are theories. Anyone saying this stuff will become. This is just guessing and does not have proof.”
Listen at 1:59:59
AI boosters should be restricted to claims about the present or immediate future.
“you gotta talk about today, mate. You get two weeks in the future, max.”
Listen at 2:06:26
OpenAI will need at least $100 billion annually to survive.
“they are going to need to raise at least $100 billion a year just to survive.”
Listen at 2:09:40
The AI bubble will begin losing momentum around 2027.
“I think sometime in 2027 things are going to start running out of steam.”
Listen at 2:10:25
OpenAI will collapse financially.
“I think OpenAI collapses.”
Listen at 2:11:13
An AI-bubble collapse would trigger a technology-sector depression.
“I think we enter a tech depression”
Listen at 2:12:12
Investors should avoid OpenAI and maintain cash amid market uncertainty.
“Oh, God, no. I honestly, I live in cash right now.”
Listen at 2:18:47
Investors should be skeptical of technology companies’ promises.
“be suspicious of tech. Like that's actually the biggest thing. It's like, be suspicious of what they're promising.”
Listen at 2:19:09
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.
Books & mentions
Ed Zitron's Newsletter
To read more of his sharp, ongoing written commentary and analysis on the tech industry