Artificial Intelligencer: OpenAI’s $852 billion problem: finding focus

OpenAI raised $122 billion in a record private round, valuing the company at about $852 billion. Under competitive pressure from Google’s Gemini and Anthropic’s Claude Code, OpenAI is reallocating resources to its Codex coding product and enterprise tools and has canceled high-profile projects like the Sora AI video app (impacting a reported $1 billion Disney-linked deal). The moves highlight intensified competition, a strategic refocus on revenue-generating offerings, and broader implications for AI infrastructure demand (Morgan Stanley projects roughly $630 billion of spending in 2026).

Analysis

OpenAI's tactical retreat to core revenue-generating developer and enterprise tooling is a classic crisis-induced concentration: scarce compute, talent and go-to-market capacity will now be reallocated toward high-ARPU, engineering-centric products. That reallocation magnifies demand for GPUs and hyperscale cloud capacity in the near term (quarters) and pushes the real bottleneck out the stack — power, site availability and long-lead mechanical infrastructure — which amplifies pricing power for scarce inputs more than software competition does.

Second-order winners are therefore upstream suppliers and operators: GPU designers/owners capture margin through constrained silicon, and hyperscalers capture outsized attach rates for cloud services and enterprise contracts; second-order losers are media and consumer partners that require bespoke product cycles and long co-development timelines. The dynamic also creates asymmetric competitive pressure between platform partners who both collaborate and compete with OpenAI (notably MSFT) — a multi-quarter tussle that can swing revenue mix and incremental margins for those partners depending on which enterprise integrations stick.

Key risks: (1) product-market risk if developer monetization lags (6-12 months), (2) supply-side reversals if Anthropic or Google deliver materially cheaper/less-power-hungry models within months, and (3) regulatory/contractual tail risk from partners left stranded by pivots (legal/settlement drag over 12–24 months). Watch near-term product cadence and cloud bookings as binary catalysts and GPU supply/pricing and datacenter power rollouts as multi-quarter gating factors.

Key Decisions for Investors

• Long NVDA via a defined-risk call spread (buy 6-month ATM call, sell 6-month 20–30% OTM call). Rationale: capture continued GPU demand and pricing power while capping theta decay. Entry: within 10 trading days. Risk/reward: limited premium outlay, target 2x+ payoff if pricing stays tight over 3–6 months.
• Pair trade: long MSFT (or AMZN) equity / short DIS equity sized 1:1. Rationale: express cloud/platform capture of enterprise AI spend vs exposed media/partner revenue loss from cancelled co-development. Timeframe: 3–9 months. Risk: MSFT/AMZN product cannibalization or DIS content wins; set stop at 8–10% adverse move.
• Buy puts on DIS (6–9 month) as a hedge or directional short. Rationale: media partners are exposed to stranded investments and renegotiations; downside realized over multiple quarters as contracts unwind. Position size: tactical (max 1–2% portfolio).
• Long GOOGL (or GOOG) 3–6 month call/stock versus short-term volatility hedge. Rationale: Google can monetize enterprise-facing models and pressure rivals, and cloud demand tailwind supports upside; catalyst windows are product rollouts and enterprise deal announcements over next 1–3 quarters.