While financing AI chips is a growing market, Goodwin warns against taking junior residual risk. His team consulted top Silicon Valley and big tech experts about the value of AI chips in 3-7 years and found that "none of them have a clue." This fundamental uncertainty makes junior positions a dangerous gamble, not a sound investment.
Related Insights
The call for a "federal backstop" isn't about saving a failing company, but de-risking loans for data centers filled with expensive GPUs that quickly become obsolete. Unlike durable infrastructure like railroads, the short shelf-life of chips makes lenders hesitant without government guarantees on the financing.
Unlike prior tech revolutions funded mainly by equity, the AI infrastructure build-out is increasingly reliant on debt. This blurs the line between speculative growth capital (equity) and financing for predictable cash flows (debt), magnifying potential losses and increasing systemic failure risk if the AI boom falters.
In a new, high-risk category, betting on infrastructure ('shovels') isn't necessarily safer. If the category fails, both app and infra lose. But if it succeeds, the application layer captures disproportionately more value, making the infrastructure a lower-upside bet for the same level of existential risk.
A new risk is entering the AI capital stack: leverage. Entities are being created with high-debt financing (80% debt, 20% equity), creating 'leverage upon leverage.' This structure, combined with circular investments between major players, echoes the telecom bust of the late 90s and requires close monitoring.
AI data center financing is built on a dangerous "temporal mismatch." The core collateral—GPUs—has a useful life of just 18-24 months due to intense use, while being financed by long-term debt. This creates a constant, high-stakes refinancing risk.
Underwriting debt for AI data centers is more challenging than for oil extraction. While oil is a predictable commodity, the value of GPUs depreciates rapidly and their long-term worth is uncertain, making it harder for lenders to gauge the risk of these tech-heavy assets over time.
Tech giants are no longer funding AI capital expenditures solely with their massive free cash flow. They are increasingly turning to debt issuance, which fundamentally alters their risk profile. This introduces default risk and requires a repricing of their credit spreads and equity valuations.
Unlike the dot-com bubble's weak issuers, the current AI debt boom is driven by investment-grade giants. However, the risk is that these stable companies are using debt to finance speculative, 'equity-like' technology ventures, a concerning trend for credit investors.
In the current AI hype cycle, a common mistake is valuing startups as if they've already achieved massive growth, rather than basing valuation on actual, demonstrated traction. This "paying ahead of growth" leads to inflated valuations and high risk, a lesson from previous tech booms and busts.
Companies like CoreWeave collateralize massive loans with NVIDIA GPUs to fund their build-out. This creates a critical timeline problem: the industry must generate highly profitable AI workloads before the GPUs, which have a limited lifespan and depreciate quickly, wear out. The business model fails if valuable applications don't scale fast enough.
