Tech critic Ed Zitron has made a provocative claim: OpenAI and Anthropic are running an unsustainable financial model that he describes as a "con," burning through over a trillion dollars in infrastructure spending while generating revenues that don't justify the investment. In a recent interview on The Diary of a CEO podcast, Zitron laid out a detailed critique of the generative AI boom, arguing that the two companies are unprofitable, dependent on subsidies from larger tech firms, and selling a product that overpromises on capabilities while underdelivering on real-world value.

What's the Financial Reality Behind the AI Boom?

According to Zitron's analysis, the numbers tell a troubling story. Amazon invested 50 billion dollars in OpenAI this year, while Google sent 10 billion dollars to Anthropic. Despite these massive infusions of capital, OpenAI and Anthropic remain unprofitable. Zitron points out that 70 percent of all AI revenues across the six leading AI companies (Anthropic, Amazon, Nvidia, Microsoft, OpenAI, and Google) come from just these two unprofitable firms.

The infrastructure costs are staggering. Zitron notes that the AI industry has already spent over 1 trillion dollars in capital expenditures, or CapEx, which refers to long-term investments like data centers and specialized computer chips called GPUs. The industry is planning to spend another trillion dollars next year. To illustrate the scale, Zitron describes OpenAI and Oracle's Stargate Abilene data center in Texas, which will consume 1.2 gigawatts of power, condensing more electricity than the entire city of Bristol uses annually into a space roughly 1,172 times smaller.

How Are These Companies Justifying Their Spending?

The companies argue that they are investing ahead of monetization, betting that user adoption will eventually translate into profits. OpenAI and Anthropic have indeed achieved rapid user growth, with hundreds of millions of people using these tools daily. However, Zitron contends that this growth doesn't justify the financial model. He argues that the companies are being deliberately opaque about their actual revenues, using vague metrics like "annualized run rates" that are never clearly defined and can mean different things at different times.

Zitron

"I think generative AI is at its heart a con. I don't think it is sold as honest software. 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," said Ed Zitron.

Ed Zitron, Tech Critic and Host of Better Offline Podcast

Zitron emphasizes that when these companies do disclose AI revenues, they use metrics that lack transparency. Public companies typically provide clear financial disclosures when they have good news, but the silence around AI profitability speaks volumes, he argues.

What Are the Core Problems With Current AI Tools?

Beyond the financial critique, Zitron questions whether the technology itself delivers on its promises. He argues that despite the hype around artificial intelligence replacing jobs and curing diseases, the actual tools are unreliable and require complex workarounds to function properly. Users must carefully select which model to use for different tasks, adjust prompts in specific ways, and chain multiple models together to achieve results. This contradicts the fundamental promise of AI as autonomous, intelligent software that works reliably out of the box.

Zitron's core criticisms include:

• Overstated Capabilities: Companies have marketed generative AI as a transformative technology that will replace jobs and solve major problems, but the tools struggle with basic tasks like search and require extensive manual configuration.
• Unprofitable Business Model: OpenAI and Anthropic cannot sustain themselves without continuous subsidies from larger tech companies, raising questions about whether the business model is viable long-term.
• Lack of Financial Transparency: Public companies avoid clearly disclosing AI revenues and use undefined metrics like annualized run rates, preventing investors and the public from understanding the true financial picture.
• Massive Infrastructure Costs: The trillion-dollar investment in data centers and GPUs has not yet generated proportional returns, and the industry plans to spend another trillion dollars despite unclear profitability.

When Might This Bubble Burst?

Zitron's interview suggests that the sustainability of the current AI investment model is in question. He argues that the financial structure is fundamentally broken, with companies spending far more on infrastructure than they are generating in revenue. The expectation that OpenAI and Anthropic will generate 400 billion dollars or more in revenue over the next 3.5 years relies on assumptions that Zitron views as unrealistic given the current trajectory and profitability challenges.

The critique raises important questions for investors, policymakers, and technology companies about whether the current generative AI boom is sustainable or whether it represents a speculative bubble that could collapse, potentially with significant economic consequences. Zitron's argument challenges the prevailing narrative in Silicon Valley and among tech investors that the AI revolution is inevitable and will justify the massive capital expenditures currently underway.