3 Ways to Invest in Anthropic Pre-IPO

Anthropic is portrayed as a potential high-profile IPO candidate, with investors able to gain indirect exposure through Alphabet, Amazon and Salesforce, all of which hold stakes in the AI company. Anthropic has committed more than $100 billion to AWS and signed a 20-year, 401-megawatt data-center lease with TeraWulf, highlighting substantial AI infrastructure demand. The article notes that pre-IPO allocations may eventually be available through select brokers, though access could require minimum balances or purchases and is not guaranteed.

Analysis

The investable read-through is materially stronger for AMZN than for strategic Anthropic shareholders generally: incremental frontier-model workloads can improve AWS utilization and support higher-margin managed AI services, while the value of a minority stake is unlikely to move consolidated earnings for GOOG or CRM. GOOG faces a more complex outcome because an ecosystem partner’s model success can simultaneously validate AI demand and intensify competitive pressure on Gemini, Search monetization, and cloud workloads. CRM’s exposure is primarily optionality rather than a near-term earnings driver.

WULF is the highest-beta expression but should be treated as a project-finance/security-of-power trade, not a clean AI demand proxy. A long-duration data-center agreement only becomes equity-value accretive if construction milestones, utility interconnection, customer credit support, and the financing stack preserve residual cash flow for common shareholders. The second-order constraint is power: large AI campus commitments raise the value of secured, deliverable generation and transmission capacity, but they can also inflate capex, borrowing costs, and equipment lead-time risk across smaller data-center operators.

Consensus is likely to overvalue indirect ownership as a substitute for direct exposure ahead of any IPO. Public strategic investors already trade on their core cloud, advertising, and software fundamentals; a private-asset mark-up may be economically immaterial relative to daily moves in those businesses. Over the next 1-3 months, verified capacity deployment and cloud consumption matter more than private-market enthusiasm; over 6-18 months, the differentiator will be whether AI demand converts into durable inference revenue rather than pre-committed training capacity.

Key Decisions for Investors

• Prefer AMZN over GOOG and CRM for a 6-18 month AI-infrastructure allocation: AWS has the clearest pathway from incremental model demand to recurring consumption revenue. Size as a core long rather than an IPO proxy; reassess if AWS growth decelerates or management signals AI capex is outrunning incremental cloud revenue.
• Do not initiate WULF solely on announced AI-campus economics. Place a buy watch on independently confirmed construction financing, interconnection timing, customer collateral, and expected equity dilution; absent those data, the risk/reward is dominated by funding and execution rather than AI demand.
• For a higher-conviction relative-value expression, consider long AMZN / short CRM over 3-6 months, sized modestly. The thesis is that infrastructure consumption has more direct operating leverage than a minority investment and application-layer optionality; cover if CRM demonstrates AI-driven subscription reacceleration materially above consensus or AWS margins compress despite revenue growth.
• Avoid treating pre-IPO access as an actionable catalyst until allocation terms, transfer restrictions, valuation, and liquidity are disclosed. A rich private-market clearing price would be a contrarian warning signal for public AI infrastructure names rather than confirmation of broad upside.