The cleaner cap table. Why Anthropic’s public-benefit structure dodges OpenAI’s charitable-trust problem — and trades it for a governance question of its own.

📊 Full opportunity report: The cleaner cap table. Why Anthropic’s public-benefit structure dodges OpenAI’s charitable-trust problem — and trades it for a governance question of its own. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

TL;DR

Anthropic’s structure, featuring a Long-Term Benefit Trust, sidesteps the legal and regulatory issues faced by OpenAI’s charitable trust conversion. However, both companies face market skepticism over governance and mission protection, impacting their valuations.

Anthropic’s corporate structure, featuring a legally independent Long-Term Benefit Trust, avoids the legal and regulatory issues surrounding OpenAI’s charitable trust conversion, making it potentially more attractive for public markets.

Founded in April 2021 by former OpenAI researchers Dario and Daniela Amodei, Anthropic is structured as a Public Benefit Corporation with a Long-Term Benefit Trust layered on top. Unlike OpenAI, which converted a nonprofit into a for-profit, Anthropic’s structure was designed from inception to avoid such legal complications, as confirmed by sources familiar with its formation.

This Trust holds a special class of voting stock, with trustees empowered to control a majority of the company’s board and prioritize safety and mission over shareholder returns. This setup means no investor can override the Trust, including major stakeholders like Google, Amazon, or institutional investors involved in Series G funding, which was announced at a $380 billion valuation.

When Anthropic files its S-1, the Trust will be a central feature of disclosure, and market participants will scrutinize how it influences governance and valuation. The structure effectively relocates the governance discount that public markets typically apply to mission-driven companies, raising the question of whether it is more or less favorable than OpenAI’s historical conversion overhang.

The Cleaner Cap Table — Thorsten Meyer AI
CHARTER
● DISPATCH / MAY 2026
THORSTEN MEYER AI · AI GOVERNANCE · § 02
AI GOVERNANCE · 02
ANTHROPIC / STRUCTURAL MIRROR
Essay · Structural-Mirror Reading · 2026-05-20

The cleaner cap table.
Why Anthropic’s public-benefit
structure dodges OpenAI’s
charitable-trust problem —
and trades it for a governance
question of its own.

Anthropic never converted a charity. So it never has OpenAI’s problem. It has a different one.
Founded April 2021 as a Public Benefit Corporation from inception — no nonprofit to convert, no charitable assets to value, no AG charitable-trust oversight, no Musk-style theory available. On the dimension that dominated three weeks of OpenAI’s trial, Anthropic simply does not present the question. That is the clean side. The other side: the Long-Term Benefit Trust — five financially disinterested trustees holding Class T voting stock, with authority escalating to a board majority within ~four years and a mandate to put mission over shareholder returns. No investor can override it — not Google’s ~14%, not Amazon, not the GIC/Coatue syndicate behind the $30B Series G at $380B post-money. When Anthropic files, that Trust becomes the single most-debated feature of the S-1. The structural argument: Anthropic did not eliminate the governance discount. It relocated it. OpenAI’s question is whether the conversion lawfully extracted charitable value. Anthropic’s is whether the mission trust subordinates returns, and by how much. Both are governance discounts. The cleaner cap table is not the cleaner valuation.
2021
PBC from inception · no nonprofit
to convert · no charitable trust
5 / majority
LTBT trustees · escalating to a
board majority within ~4 years
$380B
Series G post-money · Feb 2026
$30B raise · GIC + Coatue led
$8-12B
2026 burn vs OpenAI ~$17B
breakeven 2027-28 vs 2030s
ANTHROPIC · PBC FROM INCEPTION 2021· LONG-TERM BENEFIT TRUST· 5 FINANCIALLY DISINTERESTED TRUSTEES· CLASS T VOTING STOCK· ESCALATES TO BOARD MAJORITY· NO CONVERSION TO CONTEST· SERIES G $30B AT $380B· GIC + COATUE LED· ARR $9B → $30B EARLY 2026· 80% ENTERPRISE· 8 OF FORTUNE 10· GOOGLE ~14% · AMAZON SECOND· WILSON SONSINI ENGAGED· NO S-1 ON FILE· SNAP / LYFT GOVERNANCE PRECEDENT· SPACEX 300MW / 220,000 GPUS· MISSION OVER MARGIN· THE DISCOUNT IS RELOCATED· ANTHROPIC · PBC FROM INCEPTION 2021· LONG-TERM BENEFIT TRUST· 5 FINANCIALLY DISINTERESTED TRUSTEES· CLASS T VOTING STOCK· ESCALATES TO BOARD MAJORITY· NO CONVERSION TO CONTEST· SERIES G $30B AT $380B· GIC + COATUE LED· ARR $9B → $30B EARLY 2026· 80% ENTERPRISE· 8 OF FORTUNE 10· GOOGLE ~14% · AMAZON SECOND· WILSON SONSINI ENGAGED· NO S-1 ON FILE· SNAP / LYFT GOVERNANCE PRECEDENT· SPACEX 300MW / 220,000 GPUS· MISSION OVER MARGIN· THE DISCOUNT IS RELOCATED·
FIG. 01 — TWO STRUCTURES, SIDE BY SIDE
Structural opposites that arrive at the same place
OpenAI built commercial capacity on a charitable foundation · Anthropic built mission protection on a commercial corporation
OpenAI · the conversion path
Converted into existence
2015 · Nonprofit founding
2019 · Capped-profit subsidiary (OpenAI LP)
Oct 2025 · PBC recapitalization · Foundation retains $130B equity + control
Asks the market: trust that the conversion was lawful and will not be unwound
Anthropic · the inception path
Incorporated as one
April 2021 · Public Benefit Corporation from day one
Sept 2023 · Long-Term Benefit Trust layered on top
Never · no nonprofit · no charitable assets · no conversion
Asks the market: trust that the mission trust will not subordinate your returns
Neither company offers the public market the default reassurance — a founder-or-board-controlled company whose directors owe undivided fiduciary duty to maximize shareholder value. OpenAI’s directors sit under a Foundation with a charitable mission. Anthropic’s directors sit under a Trust with a safety mission. The Musk verdict cleared one specific challenge to OpenAI’s path. It said nothing about Anthropic’s path, because Anthropic’s path raises a different question that no court and no S-1 has yet tested.
FIG. 02 — THE LONG-TERM BENEFIT TRUST
The mechanism that is both the protection and the discount
The same design choice makes Anthropic immune to the conversion challenge and exposed to the control challenge
Anatomy
Trustees
5
Equity held by trustees
$0
Voting instrument
Class T
Mandate
Mission
Investor override
None
Board control escalates over time
2023
2024
2026
~2027
Control concentrates toward a board majority over roughly the period the company would be going and being public — the opposite of the usual dilution-of-insider-control trajectory public markets count on.
“Financially disinterested” means the trustees hold no equity and cannot profit from a higher share price. Roster skews national-security, policy, and AI-safety — Richard Fontaine (CNAS, 2025), Mariano-Florentino Cuéllar (Carnegie, Jan 2026); earlier Matheny and Christiano stepped down. The same Trust that makes the charitable-trust theory inapplicable to Anthropic is the feature public-market investors will scrutinize hardest. The protection and the discount are the same object viewed from two directions.
FIG. 03 — TWO S-1s, TWO DIFFERENT HARDEST SECTIONS
The risk-factors section is where the structural difference becomes legible
OpenAI must convince investors its structure is durable · Anthropic must convince them its structure is profitable
OpenAI · hardest disclosures
Existential-structure questions · is the corporate existence durable and lawful
  • Conversion history · nonprofit → capped-profit → PBC · $130B Foundation equity + control
  • The litigation · Musk case dismissed on timing, on appeal · underlying theory unreached
  • Regulatory overhang · AG settlement + oversight · IRS conversion review · future plaintiffs
  • Microsoft entanglement · AGI clause · $38B revenue-share cap · 27% equity · access through 2032
Anthropic · hardest disclosures
Control-and-incentive questions · will the mission governance subordinate returns
  • The Long-Term Benefit Trust · Class T voting · escalating board control · mission-balancing mandate
  • Hyperscaler concentration · Google ~14% / $40B · Amazon $25B · much in credits · antitrust at IPO
  • Compute dependency · AWS / GCP reliance · SpaceX 300MW / 220,000 GPUs · unit-economics proof
  • Mission-vs-margin tension · ad-free pledge · Pentagon dispute cost a contract OpenAI won
The cruel symmetry: Anthropic’s governance is most concerning to investors precisely to the extent that it is most effective at its stated purpose. An investor who believes mission-governance is theater discounts Anthropic less (the Trust is toothless) and OpenAI more (the conversion might unwind). An investor who believes it is real discounts Anthropic more (the Trust will subordinate returns) and OpenAI less (the conversion is done and defended). The two discounts are inversely correlated with the same belief.
FIG. 04 — THE FINANCIAL BACKBONE · THE CLEANER-BURN CANDIDATE
On financial grounds, the cleanest IPO candidate of the AI labs
Narrower burn, earlier breakeven, enterprise-weighted revenue that renews — the load-bearing valuation argument
METRIC
ANTHROPIC
OPENAI
Revenue run-rate · early 2026
~$30B
~$25B
Revenue mix
80% enterprise
Consumer-heavy
2026 operating burn
$8-12B
~$17B
Operating breakeven
2027-28
~2030s
Confirmed valuation
$380B (Series G)
$852B-$1T (target)
Structure on charitable-trust
Clean
Contested
Series G: $30B at $380B post-money (Feb 2026, GIC + Coatue, second-largest private tech round on record). ARR ramp $9B (end-2025) → $14B (mid-Feb) → ~$30B (early April). Eight of Fortune 10 are Claude customers; 1,000+ business customers spend $1M+ annually. The narrower burn and earlier breakeven are the single biggest reasons Anthropic is treated as the cleanest IPO candidate on financial grounds. The financial strength is what would let Anthropic command a premium — if the governance discount does not eat the premium.
FIG. 05 — THE GOVERNANCE DISCOUNT · A DIFFERENT DISCOUNT, NOT NO DISCOUNT
What public markets do to mission-controlled companies
Anthropic trades the conversion-durability discount for a mission-subordination discount with less precedent to calibrate against
OpenAI’s discount
Conversion-durability risk
The risk that the structure gets unwound — that the conversion is found unlawful, the AG reopens, the IRS examines, or a future plaintiff with standing prevails. Litigation-and-regulatory in nature.
The Musk verdict cleared the most-visible challenge on procedural grounds — but the underlying charitable-trust law was never reached on the merits.
Mission-subordination risk
Anthropic’s discount
The risk that the structure works as designed — that the mission trust actually subordinates returns when mission and margin conflict. The trustees are financially disinterested; they cannot be assumed to want the stock to go up. Control-and-incentive in nature.
Snap / Lyft / dual-class precedent — but those founders held equity and stayed aligned with shareholders. A financially-disinterested mission trust is categorically different, and escalates over time.
Most founder-control structures dilute as the company matures and insiders sell. Anthropic’s mission control escalates toward a board majority over exactly the period public-shareholder economic pressure intensifies. A public investor buying at the IPO is buying into a structure where the mission trust’s control is increasing, not decreasing. The countervailing case: in an era of rising regulatory scrutiny, the safety-first governance reads as risk-mitigation, and the 80% enterprise base may value the reliability the mission underwrites. The valuation lands between those two readings.
The cleaner cap table is not the cleaner valuation. Anthropic dodged the exact problem that consumed three weeks of OpenAI’s litigation — by adopting a structure that introduces a governance question public markets have never priced at this scale. It is a different discount, not no discount.
Thorsten Meyer · The Cleaner Cap Table · AI Governance 02

Implications of Anthropic’s Trust-Based Governance for Public Market Valuations

Anthropic’s design demonstrates a deliberate effort to create a cleaner, legally simpler structure that avoids the controversy and legal risks associated with nonprofit-to-for-profit conversions like OpenAI’s. However, this structure introduces a different governance discount, as investors may view the mission trust as subordinating shareholder value. This raises broader questions about how mission-oriented governance frameworks are valued in public markets, especially for AI companies at scale.

Both Anthropic and OpenAI are entering the public markets with governance structures that challenge traditional investor expectations. The market’s valuation of these firms will depend heavily on how it perceives the trade-offs between mission protection and shareholder returns, influencing the future of AI company governance and investor appetite.

Corporate Governance Matters

Corporate Governance Matters

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Structural Differences Between Anthropic and OpenAI

OpenAI, founded in 2015, transitioned from a nonprofit to a for-profit entity in 2019, with the conversion scrutinized in legal and regulatory contexts, including a recent federal jury case that dismissed Elon Musk’s challenge on procedural grounds. Its structure involves a capped-profit model, but the conversion raised questions about the lawful extraction of charitable value.

In contrast, Anthropic was built from the ground up as a Public Benefit Corporation with a Long-Term Benefit Trust, designed explicitly to prevent the legal and governance issues associated with conversion. This Trust, with its independent trustees, holds significant control over the company’s governance and prioritizes safety and public benefit, regardless of shareholder pressure.

While Anthropic avoided the conversion controversy, its governance structure introduces a new kind of market discount, as investors may perceive the mission trust as limiting profit maximization or subordinating shareholder interests, thus affecting valuation prospects.

“Anthropic’s structure was designed explicitly to avoid the legal issues faced by OpenAI during its conversion, offering a potentially cleaner path to public markets.”

— Thorsten Meyer

Intermediate Accounting 1: a QuickStudy Laminated Reference Guide (Quickstudy Reference Guide)

Intermediate Accounting 1: a QuickStudy Laminated Reference Guide (Quickstudy Reference Guide)

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Unresolved Questions About Market Reception and Valuation

It remains unclear how public investors will ultimately value Anthropic’s mission trust relative to traditional profit-driven structures. The degree to which the trust’s subordinated governance will be viewed as a discount, and how it compares to OpenAI’s conversion-related overhang, is still uncertain. Market reactions to Anthropic’s S-1 and subsequent IPO are yet to be seen.

Amazon

trustee voting stock model

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Next Steps for Anthropic’s Public Market Entry and Market Assessment

Anthropic is expected to file its S-1 in mid-2026, after which investor scrutiny of its governance structure and valuation will intensify. The company’s ability to communicate how its mission trust aligns with shareholder value will be critical. Market reactions to the IPO will set a precedent for mission-driven AI companies seeking public listings.

Your Company Mandated AI: Instructions Not Provided

Your Company Mandated AI: Instructions Not Provided

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

How does Anthropic’s mission trust differ from a traditional corporate structure?

The mission trust is an independent body of trustees with the authority to control the company’s governance and prioritize safety and public benefit over shareholder returns, unlike traditional corporate structures where shareholders have overriding control.

Will Anthropic’s structure give it a valuation advantage over OpenAI?

It is uncertain. While the structure avoids legal risks associated with conversion, it introduces a governance discount that investors may scrutinize, impacting valuation prospects compared to OpenAI’s more conventional, though legally complex, structure.

What are the risks associated with Anthropic’s governance model?

The primary risk is that the mission trust’s control could limit shareholder returns, leading to a governance discount. Additionally, regulatory and investor acceptance of such structures at scale remains untested.

When is Anthropic expected to go public?

Anthropic is expected to file its S-1 in 2026, with a potential IPO following shortly after, depending on market conditions and investor reception.

Could Anthropic’s structure influence future AI company governance?

Yes. If successful, Anthropic’s approach could serve as a model for integrating mission focus with public market access, potentially reshaping governance standards for AI and other mission-driven tech firms.

Source: ThorstenMeyerAI.com

You May Also Like

The Memento Constraint: Why Continual Learning Is the Trillion-Dollar Bottleneck Nobody Is Pricing

Exploring how the inability of current models to learn continuously shapes the AI industry and why solving this could reshape the trillion-dollar enterprise AI economy.

The Compute Concentration Audit: When Sovereign Wealth Funds Notice Three Companies Own the Frontier

Global regulators are investigating the dominance of AWS, Microsoft Azure, and Google Cloud in AI infrastructure, impacting frontier AI labs and sovereign investments.

Anchor. The Schwarz Group model.

Analysis of Schwarz Group’s €11B data center and AI infrastructure investment as a scalable industrial-anchor model for Europe.

Zeroserve: A zero-config web server you can script with eBPF

Zeroserve introduces a fast, zero-config web server that uses eBPF for scripting, serving static sites over HTTPS with hot reload and sandboxed middleware.