Anthropic vs OpenAI: An Investor Shift

Key event: skepticism about OpenAI's reported $122B funding round, which Ken Smythe warns may not be as real as presented. He says OpenAI shares are flooding the market while Anthropic is emerging as a stronger financial bet, signaling a rotation in AI investing that could weigh on OpenAI valuation and shift investor flows toward Anthropic.

Analysis

A surge of secondary supply and a reallocation inside late-stage AI capital is creating two simultaneous forces: transient downward pressure on narrative-heavy public comps and a durable concentration of demand for hyperscaler compute, specialized accelerators, and elite engineering talent. Over the next 3–12 months this should compress multiples on small/mid-cap “AI story” stocks while lifting revenue and pricing power for providers that sell GPU-hours, inference-as-a-service, or large-scale model hosting.

Second-order winners include cloud infra vendors, datacenter OS and orchestration plays, and vendors of IP/ops that lock in customers for large-model deployment; losers are mid-tier SaaS firms that rely on branding rather than differentiated model performance, and startups squeezed by rising talent costs and elongated fundraising cycles. Expect M&A pick-ups as strategic buyers (hyperscalers and incumbents) mop up lower-tier model teams and IP at discounted prices within 6–18 months.

Key tail risks: a high-profile model-safety incident or a visible funding rollback could reverse sentiment inside days and re-price both private secondaries and public peers; conversely, an anchor capital commitment or hyperscaler strategic partnership announced publicly would restore confidence and tighten multiples quickly. Monitor secondary transaction size, hyperscaler hosting deals, and option-implied skew in GPU-centric equities as high-frequency indicators of the narrative swing.

From a positioning standpoint, prefer concentrated exposure to firms with direct, monetizable capture of compute spend and long-term enterprise contracts, hedge narrative exposure via short/put protection on small-cap AI names, and keep dry powder to bid on distressed private paper that will surface if headline funding evaporates in the next 3–9 months.

Key Decisions for Investors

• Long NVDA via a 6–12 month call spread (buy ATM call, sell ~+25% OTM) sized 1–2% NAV. Entry on a 5–10% pullback or IV compression; expected payoff if datacenter GPU demand re-accelerates (target 2.5–4x notional upside), capped downside = premium paid.
• Overweight MSFT and GOOG (equal-weight, 6–12 month horizon) via LEAP calls or stock — size 3–5% NAV combined. Rationale: capture paid hosting and enterprise AI monetization; stop-loss: 10% from entry; scenario IRR upside if enterprise spend shifts to hyperscalers over 12 months.
• Short narrative small-cap AI exposure by buying 3–6 month puts or shorting ARKK-sized 1–2% NAV as a hedge against re-rating. Entry when sector ETF inflows spike or when a major secondary sale is announced; target 20–40% downside capture vs risk of whipsaw on broad market rallies.
• Allocate 2–4% NAV to opportunistic private-secondary bids (structured paper) with target entry discounts of 30–50% and hold 12–36 months. Execute only if a material late-stage financing drops or anchor commitments fail; target net IRR 20–30% with downside protected by downside-adjusted valuation.