Key Highlights

  • USD 420 billion to close to USD 6 trillion is the climb a market commentator ascribes to Nvidia (NASDAQ:NVDA) between ChatGPT's launch in November 2022 and early October 2026, a gain he says rests on one premise.
  • More than USD 1 trillion of value was lost in the week of October 15 by about twenty of the largest AI and chip stocks, on worries about how AI gets funded, which he calls a rehearsal.
  • 75% of the S&P 500's returns between November 2022 and September 2025 came from AI stocks, according to J.P. Morgan figures cited in the commentary, so a break in the chain would matter to index holders.
  • 40 times a year is how fast Epoch AI found the price of GPT-4-level answers to PhD-level science questions fell, the commentator's main evidence that price per token is the chain's weak link.

The Value Climb

A market commentator argued on October 26, 2026 that the AI trade is a chain of companies, each selling to the next, and that Nvidia sits near its head. In the commentary, Nvidia (NASDAQ:NVDA) was worth about USD 420 billion when ChatGPT launched in November 2022 and close to USD 6 trillion in early October 2026. He says that in the week of October 15, about twenty of the largest AI and chip stocks lost more than USD 1 trillion of value on worries about how AI gets funded, and he calls it a rehearsal. He also cites J.P. Morgan on AI stocks and the S&P 500's returns. These are his opinions and press-reported figures.

How the Chain Works

The commentary lists the links. Nvidia, AMD and Broadcom design the chips, TSMC makes them, and SK Hynix, Samsung and Micron Supply memory. Google, Amazon, Microsoft, Oracle and Meta buy the chips and Fill data centres, OpenAI and Anthropic rent that capacity, and users pay the labs at the end. Because one company's sales are the next company's spending, he argues the chain rests on the price per token, the fee for each small chunk of text a model reads or writes. If that price falls while the labs' computing bills stay fixed in dollars, the labs earn less, and the expected Revenue behind chip and cloud share prices weakens.

Risks and Counterpoints

The commentator says he believes in the technology and that his scenarios rest on his own assumptions. His evidence for falling prices includes the Epoch AI estimate and a corporate card company's index showing Business spending on OpenAI and Anthropic flat in August and September while token use rose. Against that, press reports of Anthropic's draft prospectus put first half 2026 revenue at USD 16.3 billion against USD 1.1 billion a year earlier. For Nvidia (NASDAQ:NVDA), the open question is whether chip Demand tracks lab revenue or lab pricing. He says the final Anthropic prospectus will confirm or correct his numbers, and that is the next data point to watch.

Where Things Stand

Nvidia (NASDAQ:NVDA) rose from about USD 420 billion to close to USD 6 trillion, in the commentator's account, because every link of the AI chain was priced on the premise that demand grows forever. His bearish view is that the price per token is the weakest link and that the Anthropic listing could expose it. Other readers may weigh the same data differently, since usage and lab revenue are rising. None of this is advice.