Anthropic's $2T IPO Would Need Double Nvidia's Profits, Analyst Says

Anthropic would need to record twice the trailing-year profit of Nvidia—the world's most valuable tech company—to justify the $2 trillion valuation it is reportedly targeting for a Nasdaq listing, according to a Tuesday report from independent financial researcher New Constructs.

The firm valued Anthropic at $150 billion and labeled the planned offering the "most ridiculous IPO of 2026," warning that the deal amounts to an "unprecedented test of investor gullibility." The math that follows from those two numbers frames the entire bear case.

How do Anthropic's numbers stack up against the $2 trillion target?

Nvidia cleared $190 billion in net income over the past four quarters. Anthropic generated $4.6 billion in 2025 revenue and posted a $42 billion net loss in the same period, according to a leaked copy of the company's prospectus reported by Reuters.

To reach a $2 trillion market capitalization, New Constructs calculated, Anthropic would have to deliver roughly twice what Nvidia earned in its trailing year. That puts the AI lab's target valuation at about 13 times Nvidia's annual profit, even as Anthropic loses more than nine times its own revenue.

New Constructs framed the gap in two sentences: "We don't think Anthropic has a viable business."

Why does New Constructs doubt closed AI models?

Trainer's firm tied the skepticism to competition from open-weight AI. The research note argues that proprietary labs will struggle to earn durable profit as freely available models close the performance gap on coding, reasoning, and agentic tasks.

"Since the arrival of open-source models, it's been clear that the closed models would struggle to generate profits," the firm wrote.

The claim matters because Anthropic sells access to Claude, a closed family of large language models, to enterprises and developers. Meta's Llama line, Mistral, DeepSeek, and an expanding roster of Chinese open-weight labs have matched or neared closed-model performance in successive benchmarks over 2024 and 2025. Pricing power compresses as the substitutes improve.

Has David Trainer's track record held up?

Trainer, the founder and CEO of New Constructs, has built a reputation as one of Wall Street's most prominent IPO bears. The firm called WeWork "the most ridiculous IPO of 2019" before the office-sharing company shelved its offering and filed for bankruptcy in 2023.

WeWork had been valued privately at $47 billion. Six weeks after the New Constructs note, the company pulled its IPO amid weak demand and intense scrutiny of its finances.

"While Anthropic offers more to society than WeWork ever did, at a $2 trillion valuation, its IPO presents far bigger risks and is positioned to be a far bigger rip off of U.S. capital markets," New Constructs wrote, framing the deal as a liquidity event for private-market backers rather than a wealth opportunity for public investors.

The firm flagged Allbirds' 2021 IPO as well. Allbirds debuted on the Nasdaq at a $4.1 billion valuation before selling its assets to American Exchange Group for an estimated $39 million earlier this year, pivoting to AI in the process.

New Constructs has missed too. Its "most ridiculous" 2020 pick was DoorDash, which it compared to WeWork as "similarly disadvantaged." DoorDash's stock rose on its first trading day in December 2020 and pushed the company's market cap above $60 billion. The valuation has since climbed to $83 billion.

Trainer acknowledged the misses in a 2021 CNBC interview. "I can't let that bother me," he said. "I have to stay true to what I think is right."

What do the leaked financials actually show?

Anthropic has not made its prospectus public. New Constructs built its case on reported figures, citing the New York Times, which reported in September that the company was on pace to generate $100 billion in annualized revenue by the end of 2026.

Anthropic's own claim, issued at the end of July, put its annualized revenue run rate up sevenfold year-over-year at $65 billion.

Both numbers sit well below the trajectory implied by a $2 trillion valuation, even after applying aggressive revenue multiples. New Constructs stressed that it has not seen the underlying filing.

"While we were not fortunate enough to be one of the few to whom Anthropic's S-1 was selectively disclosed, the reports of the leaked financials reveal more than enough to assess the gargantuan risks of investing in this IPO," the firm wrote.

Why is the existential-risk language part of the bear case?

The report also flagged Anthropic's public framing of AI's downside. New Constructs pointed to the company's stated position that AI could pose "a catastrophic or existential risk to humanity" as another reason public-market buyers should stay away.

Anthropic did not respond to a request for comment.

The remark lands as U.S. policymakers weigh new disclosure rules for frontier-model developers, including Anthropic, OpenAI, Google DeepMind, and xAI. A company asking public investors for capital while publicly warning that its own technology could end civilization presents a disclosure problem regulators have not resolved.

What happens next?

The reported $2 trillion figure remains an aspiration tied to private-market pricing rather than a disclosed range. Until Anthropic publishes its S-1, investors are pricing the deal on leaked revenue, leaked losses, and the company's own run-rate statements.

If the offering prices near the reported valuation, New Structures expects the public market to test that price within weeks—the same window in which WeWork's deal collapsed in 2019. If it prices closer to $150 billion, the gap between Wall Street's private marks and public-market discipline will narrow, and the "liquidity event" framing will lose its central justification.