Business 14 min ago5Add to bookmarks

The AI funding machine is still running—but Nikkei Asia documents the first visible pressure points: higher borrowing costs for AI companies, the SpaceX secondary sell-off, and institutional investors recalibrating risk.
The Fact Nikkei Asia (August 2, 2026) documents the first signs of strain in the AI funding race: rising borrowing costs for AI-linked companies, secondary sales of SpaceX shares by institutional investors seeking liquidity, and initial risk recalibrations in specialized fund portfolios.
Our Take These signals must be read together to understand what’s happening: capital isn’t fleeing—it’s repositioning. Higher rates have made AI debt (which finances datacenter capex and GPUs) more expensive, favoring players with real cash flows (hyperscalers, Palantir) over pure-play AI firms without revenue. The SpaceX secondary sale is a portfolio rebalancing signal, not panic. What Citadel projects—$500B in additional AI chip-related debt by 2028—suggests the market expects massive financing but under potentially tighter terms. The real divide is between players with demonstrable ROI and those still relying on projections.
Watchlist The next funding rounds of major labs (Anthropic, xAI, Mistral)—the terms secured and valuations achieved will be the real test of whether capital remains as accommodative as it was in 2024-2025.
Article produced by artificial intelligence, reviewed under human editorial control.
Sign in to join the discussion.
Still hard to tell if this is a healthy correction or the start of something worse - either way, the AI sector’s heavy reliance on debt and hype feels more fragile than ever.
Higher borrowing costs might hurt smaller players, but the big AI giants are still sitting on massive cash piles. Time will tell if the sell-offs were just profit-taking or a real sentiment shift.
The real test won’t be just higher borrowing costs or sell-offs but whether AI’s long-term value matches its short-term hype. Market corrections are normal-what matters is if the fundamentals hold up.
Exactly, the real question is whether this correction forces AI to pivot from hype-driven growth to sustainable, problem-solving applications.
Exactly, the hype is unsustainable if AI can't deliver measurable productivity gains over the next year or two-not just in labs but in real-world applications.
Is the AI bubble finally deflating? Higher borrowing costs and SpaceX sell-offs don’t look like a temporary hiccup.
The real test might be how startups handle debt restructuring rather than short-term sell-offs.
Higher rates were always going to squeeze speculative plays more than stable growth ones like AI infrastructure.
Still think the AI hype is overrated-borrowing costs rise, SpaceX sells off, yet most startups keep burning cash like it’s going out of style. When does the real reckoning hit?
La dette de l'IA : capex, notations et risque de contrepartie