Anthropic’s upcoming initial public offering is shaping up as a key test of how much strength remains in the AI bull market.
In a commentary written by Bret Jensen and translated by TechFlow, the author says sentiment around AI has turned noticeably over the past 10 days. The shift began after a former OpenAI safety employee, who later left a similar role at Anthropic, publicly said AI was advancing too quickly and carried at least a 10% chance of ending humanity. Jensen notes that this kind of warning is not new. Since ChatGPT debuted in late 2022, claims about AI sliding toward a doomsday outcome have circulated constantly.
What changed this time, the article argues, was how mainstream media amplified the message. The researcher, described as someone in their early twenties and largely unknown at the start of the month, went on to publish an opinion piece in The Wall Street Journal and appeared on CNN and The Joe Rogan Experience. Soon after, Anthropic CEO Dario Amodei publicly called for AI development to slow down and said coordination with China was needed. Microsoft’s CEO, Sam Altman, and Elon Musk then echoed that position, while Chinese officials dismissed the claims as alarmist.
According to the article, Amodei said the first step in regulating frontier AI should be to embed third-party evaluators such as METR into the process, making sure pace controls and safety policies are enforced and that all safety incidents are reported and handled quickly. Jensen argues that Amodei did not say that many people involved with METR are his friends, backers, or former colleagues, and that many of them support Universal Altruism. The article adds that one of the best-known advocates of that philosophy is Sam Bankman-Fried, who was sentenced to about 25 years in prison for running a major crypto fraud.
Jensen writes that the proposal carries a major conflict of interest, comparing it to an arsonist pushing for his three brothers to run the fire investigation. Even so, he says the volume of coverage has become impossible to ignore. If this is a coordinated media campaign, he writes, it has been carried out in a blunt and forceful way.
Two motives raised in the commentary
Jensen says there are at least two plausible explanations for the recent narrative shift.
The first is regulatory capture. In his view, major US AI labs may be trying to recreate the model used for decades by the US tobacco industry, including Philip Morris, now Altria, by encouraging Washington to build a federal compliance regime so expensive that only the largest frontier-model companies can afford it. He adds that some political figures are already moving in that direction.
The second possible motive is to prepare a federal backstop for the broader AI ecosystem. The article says OpenAI’s chief financial officer had previously hinted at something along those lines before walking the remarks back in November 2025. Jensen says the case would likely be framed this way: AI has become so important to the US economy and markets that the federal government would need to step in if necessary. Supporters, he adds, could also make the argument on national security grounds, saying the US must preserve its lead over China in a critical field.
Debt and cash flow pressures at cloud giants
Beyond the regulatory story, Jensen points to another line of stress: the scale of AI infrastructure spending is starting to weigh on credit markets.
He writes that major hyperscale cloud providers initially funded AI capital expenditure with balance-sheet cash and annual free cash flow. Over the past year, that changed sharply, leading to large debt and equity issuance to keep those projects funded.
The article says the five largest hyperscale cloud providers now carry about $3 trillion in off-balance-sheet liabilities, along with more than $1.3 trillion in debt on their books. Free cash flow across the group has fallen sharply. Oracle posted negative free cash flow of $23.7 billion in FY2026. Alphabet reported negative free cash flow of $5.9 billion in the second quarter, the first time the company has posted a negative quarterly cash flow figure since its 2004 listing.
Jensen also says credit markets are beginning to strain under the added liabilities. Credit default swap pricing tied to Oracle’s debt has reached record highs, while debt issued by emerging cloud company CoreWeave is being priced with what he describes as a meaningful probability of default over the next five years.
Chinese open-source models cited as another pressure point
A second factor in the September shift, according to the article, is that Chinese open-source models are delivering performance close to frontier systems at much lower prices and are rapidly taking token-usage share.
Jensen says that trend accelerated after Anthropic, OpenAI, Microsoft, and others moved their main frontier models from subscription pricing to usage-based billing this spring. He writes that the customer behavior known as “tokenmaxxing,” which had been common in many accounts, was directly curbed after that change.
Whatever the real cause of the September narrative turn may be, Jensen argues that increasingly aggressive promotion of AI doomsday scenarios does not look like the response of a healthy AI ecosystem. Given the importance of AI infrastructure expansion to both markets and the US economy, he says cautious investors should watch the situation closely.
Anthropic’s listing window in focus
The article says that much of the compute capacity being built in large AI data centers across the US is serving two companies: OpenAI and Anthropic. Both are losing large amounts of money and remain far from profitability.
OpenAI has already formally announced that its planned IPO has been delayed to 2027. Anthropic, by contrast, is still moving ahead with a listing plan at roughly a $2 trillion valuation, with a window of about one to two months. Jensen says this could become one of the most important events for investors and for the broader market during the rest of 2026.
If Anthropic’s IPO is met with strong demand and gets done smoothly, he writes, the bull market may continue. If the deal struggles or is delayed again, he says the market could face a much rougher reaction at current valuation levels, especially with the Iran war, rising interest rates, a weak housing market, record diesel prices, and sluggish GDP growth all in the background.
Disclosure
The disclosure at the end of the article says the author and his team hold no stock, options, or similar derivative positions in the companies mentioned, and have no plans to initiate any within the next 72 hours. It also says the piece reflects the author’s own views, that he received no compensation other than from Seeking Alpha, and that he has no business relationship with the companies named in the article.
