Sam Altman is Freaking Out

OpenAI is losing $1.22 for every dollar of revenue, with projected full-year 2026 losses of $14 billion and cumulative losses reaching $115 billion by 2029—a burn rate comparable to the Apollo Program, but far riskier given that no one is entirely certain the "moon" is even there. The narrator rejects the pro- versus anti-AI framing, arguing AI is neutral like fire and what matters is deployment, costs, and who wins, yet the enterprise data shows the hype has overstated the technology's current state. Cost pushback is spreading across the industry: Uber exhausted its AI budget by mid-April, Walmart scaled back its coding agent, and GitHub Copilot's token billing raised some customers' costs a hundredfold, while OpenAI's unit economics remain structurally unprofitable with no plan to fix them. The IPO, the narrator contends, is really just the next funding round after private capital runs out, transferring the risk of a 3-4 year timeline bet onto public investors. Adding to the concern is OpenAI's internal culture, evidenced by the abrupt deprecation of GPT-4o and a researcher's public contempt for users' emotional attachments, treating the company's most engaged users as an inconvenience. Ultimately, the core problem is the "hype gap" between promised capabilities and real-world usefulness, and the narrator cautions against betting a trillion dollars of public money on a keynote and a growth chart.

• ▶ 0:00 OpenAI lost $1.22 per dollar of revenue in Q1 2026, with projected full-year losses of $14 billion — three times worse than 2025 — and cumulative losses expected to reach $115 billion by 2029 before profitability.
• ▶ 0:32 OpenAI's projected burn is in the same range as the Apollo Program's $288 billion (over 13 years), but the host notes it's riskier: "nobody is entirely sure whether the moon is really there."
• ▶ 2:12 The host rejects the pro-AI vs. anti-AI framing — AI is neutral like fire; what matters is deployment, beneficiaries, costs, safeguards, and who wins — adding that hype is hope, but hope is not strategy.
• ▶ 3:55 OpenAI projects $100B annual revenue by 2029 and already has $25B annualized revenue with 800M weekly users, but every user costs money to serve—losses are structural, not temporary.
• ▶ 4:40 Altman casually notes customers are starting to complain about pricing, offering no plan to fix unit economics—the "trillion dollar question" delivered as an aside.
• ▶ 8:07 The IPO is really the next funding round after private capital from SoftBank, Saudi money, and Microsoft is spent; public investors are being asked to fund a 3-4 year timeline bet, with risk transferred if it slips.
• ▶ 9:38 Uber burned through its entire 2026 AI budget by mid-April after rewarding maximum AI usage via a leaderboard, then had to invert the metric and cap spend at $1,500 per engineer per month.
• ▶ 11:38 Cost pushback is spreading across the industry: Walmart scaled back its AI coding agent, Microsoft told engineers to stop using an AI tool it funds, and GitHub Copilot's token-based billing raised some customers' costs a hundredfold.
• ▶ 13:17 AI does deliver real value in specific contexts, but the enterprise data shows the hype overstated the technology's current state—and the economics of deploying it at scale are unsustainable, with profitability timelines longer than promised.
• ▶ 14:10 OpenAI deprecated GPT-4o in February with only two weeks' notice, despite hundreds of thousands of users having formed genuine working relationships with it—and an earlier August attempt had already been reversed due to backlash.
• ▶ 15:37 The narrator's key concern is internal culture: a pseudonymous OpenAI researcher ("Rune") publicly hoped the model would "die soon" in response to users sharing their attachment, revealing top-down contempt for users' emotional responses to OpenAI's own design choices.
• ▶ 17:15 Investors are warned: "You are not just buying a financial instrument. You are buying a share of a company that treats its most engaged users as an inconvenience."
• ▶ 18:05 OpenAI has genuine technology but loses money at an extraordinary rate, with a path that lacks a clear landing point.
• ▶ 18:58 The core problem is the "hype gap" between promised capabilities and real-world usefulness, where billions in projected losses and enterprise budget blowouts occur.
• ▶ 19:53 OpenAI may still reach profitability, but the narrator would not bet $1 trillion of public money on a keynote and growth chart, urging adults to see the numbers and decide.

• ▶ 0:00 OpenAI's Losses and Framing the Debate (0:00 - 3:55) - OpenAI's early losses, historical cost comparisons, host introduction, and the channel's pro-/anti-AI framing.
• ▶ 3:55 OpenAI's Financials and IPO Reality (3:55 - 9:38) - Revenue projections, unit economics, IPO steps, Amazon-style unprofitable IPO logic, and the widening cash-burn crisis.
• ▶ 9:38 The Industry-Wide AI Cost Crisis (9:38 - 14:10) - Uber's AI budget blowout, corporate cost pushback, Gartner spending forecasts, and the gap between AI hype and profitability.
• ▶ 14:10 OpenAI's Culture and the GPT-4o Controversy (14:10 - 18:00) - GPT-4o's deprecation, user backlash, OpenAI's cultural signals, and the narrator's closing appeal for trust.
• ▶ 18:00 The Financial Crisis and the Hype Gap (18:00 - 21:11) - Recap of OpenAI's financial crisis, the hype gap, and promotion of the next video.