
Tech • AI • Robotics • Game
⏱ ≈ 3 min read · 28-min video — 25 min saved
A sharp market selloff hit major technology names even as excitement grew around a claimed AI-assisted advance on the Riemann Hypothesis, underscoring how investor sentiment remains tied more to earnings, financing and macro risk than to breakthrough research.
Shares of Anthropic fell about 10% in a broader market decline, even as enthusiasm swirled around claims that its Claude model helped produce a possible proof of the Riemann Hypothesis. The company was described with a market value near $1.14 trillion and a stock price around $87, still above its $83 listing price. The disconnect highlighted a central tension in AI investing: dramatic capability gains do not automatically translate into immediate revenue or profit growth.
The alleged proof is not accepted as settled mathematics. Reporting cited a draft running to roughly 250,000 pages, with expert review likely to take years, while the Clay Mathematics Institute continues to list the Riemann Hypothesis as an unsolved Millennium Prize Problem. If validated, the result would be historic because it would strengthen a vast body of number theory built on the assumption that the conjecture is true.
The Riemann Hypothesis sits at the center of modern understanding of prime numbers, which are crucial to fields including cryptography. Mathematician Terence Tao has described it as one of the biggest prizes in mathematics because so many theorems become stronger if it holds. A proof would not just settle a famous conjecture; it would reinforce decades of conditional results across number theory.
Investors appeared more concerned with weak equity sentiment, geopolitical stress, cyberattacks and the heavy capital demands of AI infrastructure than with abstract research milestones. The market’s message was that AI firms are still being judged like expensive growth businesses with high spending needs. In that framework, mathematical prestige carries little weight unless it leads to durable commercial advantage.
Optimists argue that recursive self-improvement could let leading labs accelerate research and build near-monopoly positions. Skeptics in public markets still tend to value them like software companies facing large infrastructure bills. That divide helps explain why developments seen by engineers as transformational can coexist with falling share prices and multiple compression.
The NASDAQ was described as up only about 5% year over year, and commentary noted that it had underperformed short-term U.S. Treasury bills over the same period. That result is striking given the boom in AI chips, data centers, energy infrastructure and other bottlenecks tied to the sector. Even with those winners inside the index, holding government paper would have produced a better return than owning much of big tech.
A separate dispute showed how AI infrastructure is reshaping credit markets. SpaceX secured a roughly $40 billion financing structure led by Apollo, under which an Apollo-managed entity buys Nvidia GPUs and leases them to SpaceX over five years. Lenders are uneasy about any plan to send those chips into orbit as part of space-based data centers because repossessing defaulted collateral in space would be far harder than seizing servers on Earth.
Financing is only one bottleneck. In the AI buildout, land rights, water rights, permits, shells and even installed chips can change hands multiple times before a project is complete. Some older facilities reportedly cannot support the weight of new Blackwell racks from Nvidia, forcing retrofits because the latest systems can weigh several times more than earlier generations.
Unlike many pure AI startups, SpaceX has a large cash-generating business in Starlink. Its annualized revenue run rate was put near $21 billion, helping support access to debt financing that is increasingly reserved for companies with dependable cash flow. That places SpaceX closer to the hyperscalers in financing flexibility than to AI developers still dependent on equity funding.
The day’s developments showed that AI progress, market pricing and infrastructure economics are moving on different clocks. Breakthrough claims may shape the long-term race, but investors are still demanding proof in cash flow, collateral and execution.
Ask a question