OpenAI, Nvidia fuel $1T AI market with web of circular deals

Original Article ↗ Hacker News Discussion ↗

AI Bubble & “Circular Deals”

  • Many see the Nvidia–OpenAI–Oracle–AMD arrangements as evidence of an AI bubble being inflated by the industry itself, not by organic demand.
  • Critics highlight conditional, forward-looking commitments (rather than present cashflow) that markets treat as if they were guaranteed revenue, comparing this to Enron-style accounting, 2000-era ad swaps, capacity swaps in telecom, and Global Crossing.
  • Others argue these are just large-scale vendor-financing and strategic equity deals, not Ponzi schemes: hardware vendors front risk and take equity upside; AI firms get discounted or prioritized access to compute.

What “Circular” Means (and Disagreements About It)

  • One camp: “Circular” = money or stock effectively going in a loop (e.g., vendor invests in customer, customer spends with vendor, both book big headline numbers). This can obscure how much real, third-party demand exists.
  • Another camp: sees straightforward one-way flows (chips and cloud capacity sold to OpenAI) and insists circularity would require back‑and‑forth trading of the same asset purely to pump prices; they view current structures as normal business/credit risk.

Valuations, Accounting, and Systemic Risk

  • Several point out that these structures can:
    • Inflate revenue while true profit remains unclear.
    • Encourage double‑counting and investor FOMO.
    • Concentrate risk: if OpenAI or AI demand stumbles, knock‑on effects could hit Nvidia, Oracle, AMD, broader tech indices, and pension-heavy index funds.
  • Comparisons are drawn to dot‑com, 2008 housing, subprime, and AOL–Time Warner; some think fewer companies/people are involved now but far more capital is at stake.

Real Economic Value vs. Hype

  • Supporters: Nvidia really makes chips; AI usage and token consumption are exploding; OpenAI’s revenue growth is rapid even if unit economics are poor.
  • Skeptics: current demand is heavily subsidy- and hype-driven; products often overpromise (hallucinations, refunds); valuations assume ever-increasing LLM scale that may not materialize.

Investor Reactions & Personal Strategies

  • Some commenters are cutting exposure to US mega-cap AI names and market-cap-weighted indices, shifting into small caps, non-US markets, gold, or real assets (e.g., land).
  • Others warn that timing crashes is nearly impossible; advocate diversification, factor investing, and not overreacting to headlines.
  • Underneath is a broader worry that “incestuous” AI financing plus policy complacency could amplify a future correction well beyond the AI sector.