
Tech • AI • Robotics • Game
⏱ ≈ 3 min read · 2 h 3 min video — 2 h saved
Anthropic fell sharply on its public-market debut even as an AI-assisted claim of a proof for the Riemann Hypothesis stirred excitement, highlighting the widening gap between Silicon Valley’s expectations for AI and Wall Street’s focus on near-term revenue, risk and financing.
Anthropic traded at about $87, down roughly 10% on the day, for an active market value near $1.14 trillion, still modestly above its $83 IPO price. The selloff came even as discussion intensified around a claim that Claude helped produce a proof of the Riemann Hypothesis, one of mathematics’ most important unsolved problems. Investors showed little interest in tying that claim to near-term earnings.
The reported proof has not been validated by the mathematical community, and the Clay Mathematics Institute still lists the Riemann Hypothesis as an unsolved Millennium Prize Problem. Review could take a long time, especially because the raw material was described as about 250,000 pages before trimming. Even if the result holds, the direct financial reward is only $1 million, underscoring why traders are not pricing it in as a business catalyst.
The Riemann Hypothesis concerns the distribution of prime numbers, which underpin wide areas of number theory and have practical links to cryptography. Many mathematical results are proved only conditionally, assuming the hypothesis is true. A confirmed proof would strengthen a large body of existing work, but its effect on commercial AI valuations is far less immediate.
The episode exposed a deeper split between AI believers and public-market investors. In much of Silicon Valley, advances in automated research are viewed as evidence for recursive self-improvement and accelerating model capability. In equities, many investors still treat leading AI firms more like high-capex software businesses, valuing them on revenue durability, margins and infrastructure costs rather than breakthrough narratives.
The decline in Anthropic came amid a wider pullback in technology shares, with the Nasdaq and S&P 500 also falling. Market participants pointed to geopolitical stress, cyberattacks and volatility around major AI and semiconductor names including Nvidia. Even after a multiyear boom in chips, data centers and energy infrastructure, tech benchmarks were described as lagging the return on short-term U.S. Treasury bills this year.
A separate financing dispute showed how quickly AI optimism runs into old-fashioned credit concerns. Lenders tied to a roughly $40 billion SpaceX financing backed by Nvidia GPUs are reportedly wary of plans to send those chips into orbit as part of proposed space-based data centers. If the chips become collateral in space, repossessing them after a default becomes far more complicated.
The financing structure depends on SpaceX having cash-generating operations outside launch services, chiefly Starlink. Its annualized revenue run rate was cited at about $21 billion based on third-quarter 2026 figures, giving the company a stronger basis for large debt deals than many AI startups. That mirrors a broader market reality: access to big financing increasingly favors companies with existing cash flow, such as Meta, Google and Amazon.
AI buildout is also running into physical constraints. Land, power rights, water rights, permits and construction contracts can all change hands before a data center is completed, and even the chips allocated to a site may be redirected. Some older facilities reportedly cannot support the weight of newer Nvidia Blackwell racks without structural retrofits, adding another cost layer to AI expansion.
The policy backdrop is shifting as well. A bipartisan Senate vote of roughly 63-33 advanced the Bipartisan American Affordability and Jobs Act, framed by supporters as a “freedom to build” measure. The effort reflects growing recognition that slow permitting has become a serious obstacle to energy projects, housing, factories and the infrastructure needed for AI.
The trading slump in Anthropic showed that spectacular AI-adjacent headlines are not enough to overcome investor concern about costs, financing and execution. For now, public markets are rewarding cash flow and build capacity more than promises of revolutionary mathematics.
Ask a question