AI listings and valuations fall as funding market toughens

The funding squeeze seems likely to widen the divide inside the AI industry. Hyperscalers with solid cash flow and veteran suppliers can self-finance their spending, while newer participants that rely on debt or equity are facing higher expenses and a shorter runway. Credit markets are the key signal: the spreads on high-yield AI infrastructure debt, particularly with CoreWeave leading, are the earliest indicator of whether stress is spreading or staying confined. In Asia, chip stocks in South Korea and Taiwan are acutely sensitive to any drop in AI capital spending, a risk already evident in SK hynix's market value falling by about half since June. Oil is part of the cost-of-capital equation: Brent crude above $110 a barrel keeps inflation and rate expectations elevated, holding up bond yields and raising the barrier for capital-intensive projects with payoffs years away.

---

AI hardware demand is still strong, but investors are growing much choosier in how they fund it, and the companies that depend most on outside financing are feeling the pinch first.

Summary:

• This month, three AI-related listings have been cancelled, halted or postponed: Firmus Grid's roughly $5 billion IPO, Airtrunk's Singapore listing and SK hynix's Solidigm IPO.
• OpenAI is reportedly on the verge of raising up to $30 billion at a pre-money valuation of $1.25 trillion, lower than the roughly $1.4 trillion it initially sought.
• Debt markets are flashing strain: yields on some of CoreWeave's unsecured bonds are around 15%.
• On Thursday, the Philadelphia semiconductor index fell 6.7% for its worst day since June, as Meta's capital spending reignited worries about returns on AI investment.
• Counter-evidence is also strong: JPMorgan says positioning is largely unwound and the tech sector generates about $400 billion more cash than it spends, while Western Digital reported surging demand and prices.
• Key events ahead include megacap earnings, the Fed decision and SK hynix results on October 28, plus OpenAI's funding round and Anthropic's expected listing in mid-November.

The AI boom still has strong demand, but it is hitting a funding market where caution is rising. In a matter of weeks, three AI infrastructure listings have been withdrawn, halted, or pushed back, private valuations have been marked down from earlier targets, and some debt markets are reflecting actual distress. Taken together, these signs suggest that investors who once financed the build-out with few questions are now growing more selective.

The newest indication arrived on Friday, when SK hynix said its Solidigm unit had abandoned a review of a US IPO intended to help fund a new fabrication plant in the United States. The South Korean chipmaker cited market conditions and share price trends, adding that the decision did not reflect any change in Solidigm's growth prospects. Reports from earlier in the autumn had indicated the unit could be valued at up to $150 billion in the listing.

Solidigm's decision followed two similar moves in Asia-Pacific. Earlier this month, Australian data centre operator Firmus Grid cancelled an IPO worth almost $5 billion that would have been the country's largest in nearly 30 years, blaming volatility. Days later, Blackstone-backed Airtrunk delayed a planned Singapore listing to the first half of next year, The Wall Street Journal reported.

The repricing also extends to private markets. OpenAI is close to finalising a funding round of up to $30 billion at a pre-money valuation of $1.25 trillion, led by investors from the United Arab Emirates, according to recent reports. Earlier this month, Bloomberg reported that the company had offered investors a fixed valuation of about $1.4 trillion. The new figure is still far above the roughly $850 billion valuation from March, but it shows that investors have pushed back on price. Reuters reported that OpenAI is rushing to lock in commitments before rival Anthropic's expected market debut in mid-November, as both tap a similar pool of capital.

The tension is most apparent in credit. CoreWeave, the biggest of the neoclouds that lease computing power for AI tasks, has experienced yields on some unsecured debt rise to around 15%, a level typically associated with trouble, after heavy borrowing to build data centres. Analysts are split. HSBC has flagged structural concerns across the neocloud sector, while Goldman Sachs argues that CoreWeave's use of asset-backed project financing limits the increase in its borrowing costs.

Public markets mirror the same doubts. The Philadelphia semiconductor index fell 6.7% on Thursday for its largest single-day drop since June, and the Nasdaq fell more than 2.5% after tech giant earnings, including Meta's spending plans, revived worries that AI returns may not justify the investment. SK hynix's market capitalisation has dropped by about half since its June high.

Several factors are driving this shift. Interest rates are climbing: the Federal Reserve resumed increases in September and the Bank of Japan has pushed its policy rate to a 31-year high. Higher rates increase the cost of funding projects whose revenue comes years after the outlay. Geopolitical risk has also grown, with the war involving Iran keeping Brent crude above $110 a barrel and inflation expectations high. As the size of AI spending has expanded, investors have concentrated more on when it will pay off.

There is, however, a strong counter-argument. This week, JPMorgan strategists stated that hedge funds have already scaled back their exposure to chip stocks, which could limit further declines. They also noted that neocloud capital spending of about $80 billion this year is modest next to roughly $1.8 trillion of total technology investment, and that the broader tech sector is producing about $400 billion more cash than it uses, leaving bigger companies capacity to cover any funding gaps. Demand also appears real: Western Digital said this week that revenue rose about 55% as AI customers drove a 28% increase in storage prices, and that it has demand visibility to 2030 or 2031 from its largest customers. OpenAI has told investors that its annualised revenue has reached about $50 billion.

When combined, the evidence points to a divergence rather than a crash. Companies that can fund AI investments out of their own cash flow stay well positioned. Those that depend on external capital, especially debt-funded newcomers and new listings, face higher costs and stricter terms. This points to healthier discipline than an unfettered boom, but it implies that growth may decelerate at the edge for the most leveraged parts of the supply chain.

Several tests are ahead. Megacap earnings will show whether spending plans keep up. The Fed decides on October 28, the same day SK hynix releases results. OpenAI's round has not yet closed, and Anthropic's listing next month will reveal whether public investors are still willing to back AI at scale. A warm reception would suggest the window is reopening. Additional withdrawn deals or wider spreads on AI-related debt would indicate the squeeze is deepening.