Microsoft needs to open up more about its OpenAI dealings

Original Article ↗ Hacker News Discussion ↗

OpenAI’s Economics and Business Model

  • Commenters note OpenAI has never been profitable and suggest losses are enormous; some predict it may never reach profitability and could ultimately be absorbed by Microsoft or crash the broader AI market.
  • Skepticism that ads or “erotic” content will rescue the business: ad markets are already dominated by highly optimized incumbents, and it’s unclear ChatGPT data will translate into a strong ad product.
  • Some see erotic/romantic chat as a potentially huge revenue source (citing existing “sexy chatbot” demand and porn’s historic role in monetizing new tech), but others worry about revenge porn, deepfakes, minors, and card processors cutting them off.
  • There’s concern OpenAI still lacks a coherent business model, with jokes about “ask the AGI for a business plan.”

Commoditization and Competitive Pressure

  • OpenAI’s pricing is seen as trapped: raise prices and users switch to improving cheap/open models; keep prices low and losses continue.
  • Several argue pure-LLM companies (OpenAI, Anthropic) are structurally weaker than platforms (Microsoft, Google, Meta) that can bundle LLMs into existing products.

Microsoft’s Strategic Rationale

  • Many think Microsoft’s real win is defensive/strategic: integrating GPT into Office, Windows, Azure, etc., to preserve and enhance its core franchises and prevent churn.
  • For a company with hundreds of billions in revenue, a multi‑billion loss on OpenAI is framed by some as “only money” and effectively R&D spend.
  • Others counter that if the numbers were good, Microsoft would be showcasing them instead of burying them and using vague “Copilot” metrics.

Accounting, Disclosure, and “Enron Vibes”

  • Debate centers on whether Microsoft should treat OpenAI as a related party and provide detailed disclosures of transactions (cloud credits, revenue sharing).
  • Some posters argue the equity-method stake should trigger related‑party disclosure; others insist under US GAAP it need not be presented that way.
  • The opacity, special-purpose vehicles for datacenters, and large equity-method losses remind some of past accounting scandals, even if they stop short of calling this fraud.
  • There is pushback that this is normal long‑term investment behavior, analogous to early Amazon/Google, and that critics are overreacting.

AI Bubble, Systemic Risk, and Hype

  • Mixed views on whether the AI boom is a classic bubble:
    • One side: “AI bubble is different” and less poppable because megacorps, not retail, hold most risk.
    • Other side: dot‑com is cited as precedent—technology can be transformative while prices still massively overshoot and later crash.
  • Concern that if OpenAI collapses, funding could dry up for much of the AI sector.
  • Commenters see parallels to crypto: same people, same hype dynamics, but with tech that is genuinely useful and simultaneously over‑ and under‑hyped.