3 Ways to Invest in Anthropic Before It Goes Public

Anthropic has confidentially filed for a U.S. IPO after raising $65 billion at a near-$965 billion valuation, making it one of the most closely watched AI listings on Wall Street. The article highlights three indirect public-market ways to gain exposure: Alphabet, with up to $40 billion in potential investment and $200 billion of Anthropic cloud spending commitments; Amazon, with up to $33 billion of potential cash commitments and nearly $100 billion of AWS-related spending; and Salesforce, whose Anthropic stake is valued at about $5 billion. The piece is constructive on AI infrastructure and enterprise software names, but it is primarily a thematic investment overview rather than new company-specific operating results.

Analysis

The market is effectively being offered a three-way monetization of the same scarcity asset: frontier-model compute. The less obvious winner is the infrastructure layer with contractual lock-in, because hyperscaler capex is no longer just defensive spend; it becomes a toll road into model adoption, creating a longer-duration annuity than the model company itself. That favors GOOGL and AMZN more than pure AI application names, since the real economic moat is shifting toward power, chips, networking, and cloud scheduling efficiency rather than model branding.

The second-order effect is competitive entanglement. GOOGL gains upside from Anthropic while simultaneously funding a rival to Gemini, which means the stock’s Anthropic exposure is partially diluted by internal cannibalization risk; any Anthropic share gains could pressure Google’s own model pricing and search-adjacent monetization over 12–24 months. AMZN has cleaner economics because AWS monetizes usage regardless of which model wins, and the hardware stack (Trainium/Graviton) gives Amazon a chance to compress customer acquisition costs for AI workloads while widening the moat against Azure on unit economics.

CRM is the most interesting contrarian setup: it has the smallest direct AI infrastructure claim, but the highest strategic optionality if enterprise agent workflows become procurement-driven rather than app-driven. The risk is that Anthropic/other agents disintermediate traditional SaaS seats faster than CRM can monetize automation, so the name can work only if AI becomes an add-on to workflow software rather than a substitute. Near term, the IPO is a sentiment catalyst, but the real move in these names likely unfolds over quarters as capex, cloud backlog, and enterprise adoption data confirm whether AI demand is converting into durable spend or just pre-IPO narrative inflation.

Key Decisions for Investors

• Long AMZN vs. GOOGL into the Anthropic IPO window (1-3 months): AMZN has cleaner upside from compute consumption and less internal product cannibalization risk; use this as a relative-value expression rather than an outright AI beta trade.
• Buy GOOGL only on pullbacks, not into strength (4-8 weeks): upside exists from cloud and equity appreciation, but the stock is already the most exposed to a 'benefit but also compete' narrative; prefer selling upside via covered calls if holding core.
• Initiate a small tactical long CRM / short software basket (e.g., CRM vs. broad SaaS ETF) for 3-6 months: if enterprise agents are monetized through workflow platforms, CRM’s embedded Claude relationship could re-rate faster than peers; keep size modest because displacement risk is real.
• Avoid chasing NVDA on this headline alone: the article is a demand-duration story, not a new supply-shock story; wait for evidence that Anthropic capex is incremental to broader AI spend before adding exposure.
• If implied vol rises into the IPO filing cycle, consider short-dated call spreads on AMZN and GOOGL: the setup is a slow-burn fundamental winner rather than a one-day repricing event, so options can monetize sentiment without overpaying for downside protection.