Will the AI bubble burst soon? Why Big Tech CEOs are asking the US government for protection
- What to look for
- You can check figures cited as evidence of a bubble, including Anthropic's and OpenAI's losses, paying households at 2.2%, China's token shares, S&P concentration, and the Buffett indicator.
- Why it matters
- This helps you learn how to assess the bubble debate using figures such as investment prospectuses, token shares, and concentration, and develop a critical viewing perspective that distinguishes partisan reasoning.
AI summary
Ben Norton characterizes the contradiction between Anthropic's warnings of existential risk and its pursuit of a 2 trillion dollar IPO as marketing hypocrisy. He interprets the September (month 9) Trump–Big Tech luncheon and the agreement on voluntary regulation without legal force as a show to protect Silicon Valley. As evidence of a bubble, he cites the WSJ's assessment of the biggest bet in American history, data centers accounting for 3.6% of GDP, and IPOs by three companies totaling 5.2 trillion dollars. He reports on China's open-source catch-up efforts (token shares of 54% in the US, 48% in Europe, and 61% in the Global South), lobbying for bans, and attempts at monopolization. Citing Anthropic's 2025 net loss of 420 hundred-million and revenue of 46 hundred-million, OpenAI's 2024 loss of 50 hundred-million, and paying households at 2.2%, he describes a cash-burning structure. He identifies a vicious debt cycle in AI capital expenditure by 4 companies exceeding 1 trillion dollars over 3 years, the yield on 10-year bonds surging from 4% to 5.2%, and issuance of 2,360 hundred-million dollars during months 1~5 of year 26. Citing the S&P's top 10 companies accounting for 40% and a Buffett indicator of 235%, he argues that the level exceeds those of 1929 and 2000. All directional outlooks are the speaker's conditional judgments, not definitive short-term price predictions.