The Next Bubble

AI software has the awkward profile of a market that is both obviously real and easy to overpay for. OpenAI annualized revenue has been reported crossing the $20 billion mark and later prints in the mid-twenties. Cursor went from a November 2025 raise—$2.3 billion at about a $29.3 billion valuation on more than $1 billion of revenue—to reports of $2 billion-plus ARR and chatter about a much higher next round within months. Gartner 2026 forecasts put worldwide AI spending growth around 47% and AI software near $453 billion, up from roughly $283 billion the prior year. Those are not empty slideware numbers. The narrative wrapping them is bigger still. Agents become the interface. Software stops assisting knowledge work and starts doing it. Today AI vendors capture a meaningful slice of white-collar economic value. That leap—from revenue is exploding to these specific companies own the residual—is where bubble math lives. Unit economics do not look like classic SaaS. Inference costs scale with usage. Price cuts show up fast when models look interchangeable; Reuters reported OpenAI cutting developer pricing on a frontier model by more than 20% in August 2026 under competitive pressure from Anthropic and Chinese labs. Gross margin quality is therefore a moving target. Investors who import SaaS multiples onto metered intelligence are making a category error unless switching costs and distribution prove sticky. There is a second knife edge. Gartner has estimated on the order of $234 billion of enterprise application spend at risk from agentic AI by around 2030—roughly a fifth of enterprise SaaS in that framing. AI can create a huge new category while chewing through the pricing power of the old one. Value might land with model labs, application-layer startups, data owners, or customers who simply pay less for the same work. AI wins does not specify the ticker. Adoption surveys still show the gap between demos and P&Ls. McKinsey mid-decade State of AI work found most organizations stuck in experiment or pilot mode; a minority reported enterprise-level EBIT impact from AI, and for many of those the impact was under 5% of EBIT. Markets can capitalize the end-state years before CFOs can find it in earnings. Spending and pilots are not the same as durable productivity capture. Frontier valuations make the discount rate explicit. Reuters reporting in August 2026 described Anthropic IPO chatter tied in part to roughly $190–200 billion of 2028 revenue against a then-current run rate near $47 billion. Maybe they get there. The price still embeds a narrow path on competition, cost, share, and pricing power. When markets price one future instead of a range, disappointments do not require the technology to fail—only to land in a merely-very-good scenario. Reflexivity does the rest. Capital funds distribution, talent wars, and price subsidies. Growth validates the narrative. The narrative raises the next round. At first that can be productive: it accelerates real product development. Later high valuations become their own evidence that the future is inevitable. When financing tightens, the same loop runs backward and sorts durable products from capital-fueled ones. None of this requires revenue to go to zero for investors to get hurt. A classic late-cycle path is slower growth, harder price competition, softer margins, and a re-rating of the terminal share story—while AI keeps spreading through the economy. That is the dot-com lesson in a new wrapper: the technology can remake workflows while investors who paid for perpetual scarcity and winner-take-most margins still lose money. There is a clean bull path too. If agents become the default interface for knowledge work and a few platforms keep distribution and data advantages, today revenue leaders can grow into aggressive valuations. Spending forecasts already imply a very large industry. The open question is not whether AI software matters. It is how much of the surplus accrues to vendors versus customers once competition and switching costs do their work. The software bubble case is not that AI is fake. It is that competition, commodity pressure on models, and early ROI evidence can transfer a lot of the surplus to customers—exactly while valuations assume the opposite.

Private credit's rise makes sense on paper. After 2008, banks faced tighter capital rules and pulled back from middle-market leveraged lending. Private equity sponsors needed financing. Institutional investors—pensions, insurers, endowments—wanted yield that didn't live in the public high-yield market. Non-bank lenders stepped in with bilateral loans, often floating-rate, often with more flexible structures than a broadly syndicated deal. That story is real. So is the size. The IMF's April 2024 Global Financial Stability Report put the market at roughly $2.1 trillion in assets and committed capital, with about three-quarters in the United States. AIMA's Financing the Economy 2025 work estimated global private credit AUM near $3.5 trillion by end-2024. The FSB's 2026 vulnerability report, drawing on member data, put outstanding private credit in a $1.5–2.0 trillion band—already comparable to the institutional leveraged-loan market. Definitions differ, which is itself a tell: when you can't agree on the denominator, opacity is part of the product. The bubble case is not that private credit is fake. It's that the boom has outrun the market's ability to show stress clearly. Start with borrowers. Private credit companies tend to be smaller and more levered than public-bond peers. The IMF noted that more than a third of firms with private-credit-like size and leverage had interest costs exceeding earnings after the rate shock. Floating-rate loans transmit higher policy rates quickly. One response has been payment-in-kind (PIK): instead of cash interest, the unpaid coupon compounds into the loan balance. In BDCs, the share of interest income coming from PIK roughly doubled since 2019 to around 9% in IMF figures; some portfolios have seen far higher PIK shares. PIK can be a bridge. It can also be a way to defer the moment when a loan is recognized as broken. Then there's valuation. Private loans rarely trade. They're often marked to model with limited price discovery. That softens reported volatility in good times and can delay loss recognition in bad ones. Ratings are frequently private or absent; the FSB has flagged growing use of lesser-known private ratings to open doors for rating-reliant investors such as insurers. When prices don't move until they have to, stable NAVs can be a feature of the accounting, not proof of credit quality. Competition is doing the usual late-cycle work. Dry powder rose sharply over the past decade. Managers have fixed fundraising clocks and return targets. Banks have come back into larger deals. The result, according to both the Fed and the IMF, is pressure on spreads, underwriting standards, and covenants—including more covenant-lite structures that look a lot like the public leveraged-loan market private credit once claimed to be stricter than. The Fed has also documented bank credit lines to private-credit vehicles growing from roughly $8 billion in early 2013 to about $95 billion by 2024-Q4. That is still small relative to the banking system, but it is the kind of plumbing that matters when funds need liquidity at the same time. Liquidity mismatches are creeping in from another direction. Evergreen and retail-accessible private credit vehicles offer more frequent redemption features than classic closed-end funds. In a calm market that sells diversification. In a stressed market it can force sales or gates just when marks are least trustworthy. None of this means the sector is worthless. Private credit finances real companies. Long-term capital reduces classic bank-style maturity transformation. Default rates have stayed relatively low through the first big hiking cycle—partly because lenders can amend and extend with a small creditor group. The industry has not yet been stress-tested at today's size in a deep recession. That's the point of treating it as a bubble candidate. The danger isn't that every loan blows up tomorrow. It's that a large, opaque, interconnected credit system can look fine—until PIK stops being a temporary fix, until marks have to catch up with reality, until banks tighten lines, or until retail vehicles face redemption pressure. Useful credit can still be a bad investment if too much capital chased too little underwriting discipline at the wrong point in the rate cycle.