Anthropic’s dual-class share structure borrows ideas from Mark Zuckerberg, Elon Musk—and the New York Times

Anthropic’s S-1 describes five share classes: a single Class F share gives the seven co-founders 50.1% of voting power without economic rights, while Class T lets a trust elect a board majority; public Class A holders are capped at 49.9% of votes. The filing warns the structure may limit public investors’ influence, reduce demand or increase volatility, and raise governance-litigation uncertainty. It also discloses an agreement to spend up to $120.5 billion on Nvidia-powered compute from SpaceX.AI, cancellable on 90 days’ notice after early 2027.

Analysis

The key pricing issue is not simply founder control: economic exposure and voting control can diverge for years, while the trust can shape board composition without holding equity. That combination may protect long-horizon AI investment from quarterly pressure, but it weakens public shareholders’ ability to respond to poor execution, related-party decisions, or a sale process. Expect a potential governance discount and a narrower marginal buyer base, particularly among funds with governance screens; whether that discount matters depends on IPO demand, float, and index eligibility, none of which are established here.

The control structure may also make litigation and trustee succession material valuation variables rather than background risks. Watch for challenges to the structure, changes in trust membership, or founder departures that could alter the voting regime. Over 6–18 months, the harder fundamental question is whether concentrated compute commitments create durable capacity and competitive advantage or leave Anthropic exposed to supplier dependence and cancellation risk. The disclosed arrangement with SpaceX.AI could support SPCX-affiliated demand, but the maximum commitment is not evidence of recognized revenue or take-or-pay economics; NVIDIA’s exposure is likewise not quantifiable from the article.

Contrarian view: a governance discount may be overestimated if investors value mission continuity and the IPO is priced to compensate. Conversely, historical outperformance of dual-class IPOs is not evidence that this unusually separated control/economic structure will outperform. Meta is a poor direct precedent: its founders’ voting power is tied to economic shares, unlike Anthropic’s durable Class F mechanism. No trade is justified solely by the filing; the IPO price, float, lockups, contract terms, and investor demand need verification.

Key Decisions for Investors

• Do not chase ANT at launch on the dual-class IPO performance statistic. Wait for price discovery and verify the public float, valuation, lockup schedule, and index treatment; consider exposure only if the entry price compensates for limited shareholder influence.
• Set a post-IPO governance watch: reassess if litigation is filed, trust composition changes, or founders leave the LLC. Those events could change the control discount; a favorable court outcome or credible governance safeguards would weaken the bearish case.
• Treat SPCX’s potential benefit from the compute arrangement as optionality, not booked earnings. Before taking a directional position, verify contract duration, minimum purchase obligations, cancellation terms, and revenue recognition; a cancellation notice or reduced commitment would falsify the demand-support thesis.
• Do not infer a material NVDA earnings catalyst from the compute disclosure. Monitor confirmed Anthropic capacity deployment and NVIDIA-related purchase disclosures; absent evidence of incremental orders, avoid trading NVDA on this headline.