This article was first published in March 2026 and was substantially rewritten on June 21, 2026, to reflect Anthropic's 65 billion dollar Series H round, its confidential IPO filing, and the Pentagon dispute that unfolded in between.
Anthropic says it is worth $965 billion. No outside auditor has confirmed a single number yet.
Image Credit: Leonardo AI
Five years ago, Anthropic did not exist. In June 2026, the company quietly submitted paperwork to American regulators that could turn it into one of the largest public listings in stock market history, and it did so without releasing a single financial number to the public.
The filing came four days after Anthropic closed a $65 billion funding round that valued the company at $965 billion, a figure that, for the first time, placed it ahead of its longtime rival OpenAI in private-market value. Wall Street has seen this before with hyped private companies that never delivered once their paperwork went public. This time, three of the most closely watched technology companies in the world, Anthropic, OpenAI, and Elon Musk's SpaceX, are all heading toward public markets within months of each other, and the numbers behind each filing matter more than the headlines they generate.
This piece walks through what Anthropic actually filed, what a confidential filing legally commits the company to, where the revenue numbers came from and what they leave out, who actually owns the company, the unresolved legal fight with the Pentagon that could shape its government revenue, and the practical ways an everyday investor can get exposure before the stock trades on a public exchange.
In this article
- What Anthropic is and why it suddenly matters
- Anthropic has already filed to go public
- The revenue numbers behind the valuation
- What the big numbers leave out
- Who owns Anthropic, and who gets paid first
- Anthropic, OpenAI, and now SpaceX are all racing to Wall Street
- The Pentagon dispute explained
- The risks an honest investor should sit with
- A framework for valuing a company with no real comparable
- How to get exposure before the listing
- Frequently asked questions
- Desidaily take
What Anthropic Is and Why It Suddenly Matters
Most people have used a chatbot. Far fewer could say who actually builds Claude, the AI model that has become one of the most trusted assistants for enterprise software, legal work, and software development.
Anthropic was founded in 2021 by seven former OpenAI employees, led by Dario Amodei as chief executive and Daniela Amodei as president. They left OpenAI over disagreements about how quickly and how safely AI should be developed. That decision turned into one of the more consequential exits in the history of the technology industry.
The company operates as a public benefit corporation, a legal structure that requires it to formally balance commercial success with its stated mission. That structure became more than symbolic in 2026, when Anthropic chose to fight the United States government in court rather than remove safety restrictions on how Claude could be used by the military, a dispute explained in full later in this article.
Claude is already replacing entire categories of enterprise software inside many of the world's largest companies, a shift covered in more depth in Claude AI Is Quietly Replacing Major Software Tools in 2026.
Anthropic Has Already Filed to Go Public
On June 1, 2026, Anthropic confidentially submitted a draft registration statement to the United States Securities and Exchange Commission, the formal first step toward a stock market listing. The filing came four days after the company closed its $65 billion Series H funding round, led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, which valued Anthropic at $965 billion.
Anthropic has not announced a date, a ticker, or a share price. Reports from bankers and analysts point to an October 2026 listing on the Nasdaq, with Goldman Sachs, JPMorgan, and Morgan Stanley expected to lead the offering and proceeds potentially exceeding $60 billion, which would rank among the largest technology listings on record. None of that is confirmed, and the next section explains exactly why an Anthropic IPO date should still be treated as unconfirmed.
What a Confidential Filing Actually Commits the Company To
A surprising number of articles about this filing treat it as equivalent to a confirmed listing date. It is not. Under rules that trace back to the 2012 JOBS Act, companies can submit a draft registration privately, which lets regulators review the financials without immediately exposing them to competitors, customers, or the public. The company only has to make that filing public on the SEC's EDGAR system at least fifteen days before it begins its investor roadshow.
That single fact matters more than most coverage suggests. It means that even if Anthropic's October target holds, the public will only see audited revenue, profit, customer concentration, and risk disclosures a few weeks before the shares actually price, not the months that the headlines have implied.
Confidential filings can also be quietly withdrawn. The most cited case in finance is WeWork, which confidentially filed for its IPO in December 2018, made that filing public in August 2019, and withdrew it entirely a month later after investors balked at its governance and valuation, according to Wikipedia's account of the company's history. WeWork's private valuation had peaked at 47 billion dollars earlier that year. By the time the listing was abandoned, the figure being discussed publicly had fallen to roughly 10 billion dollars. The company eventually went public two years later through a SPAC merger at a fraction of its original ambitions and filed for bankruptcy protection in 2023. None of this means Anthropic will follow a similar path. It means a confidential filing is the start of a process with several exit ramps, not a guarantee of where it ends.
For now, the more reliable way to judge whether Anthropic will actually list in October is to watch for the public version of its S-1 appearing on EDGAR, and to watch how SpaceX's own listing, which began trading earlier this month, performs in its first weeks, since that performance is widely viewed as a test of whether public investors will keep paying private market prices for AI adjacent companies. The mechanics of that listing, and who actually profits from it, are covered in If SpaceX Goes Public, Who Actually Gets Rich.
The Revenue Numbers Behind the Valuation
Anthropic's revenue growth is the central argument for its valuation, and the trajectory is hard to overstate. The company generated roughly 100 million dollars in annual revenue in 2023. That grew to about 1 billion dollars by the end of 2024. By Anthropic's own account, its annualized run-rate revenue crossed 47 billion dollars in May 2026, up from roughly $9 billion at the end of 2025.
Enterprise customers drive most of that growth. Eight of the Fortune 10 companies are now Claude customers, and the number of clients spending more than 1 million dollars a year with Anthropic has continued to climb. Claude Code, the company's AI coding assistant, has become a sizable business on its own, reportedly crossing $1 billion in annualized revenue within roughly six months of launch.
Dario Amodei has said publicly that the company's revenue grew from $1 billion to a $7 billion annualized run rate within nine months, a pace he has described as unmatched by any software company in history, in a statement posted on Anthropic's own site. Anthropic has also signaled it is close to its first profitable quarter, a notable shift from earlier guidance that pointed to full-year profitability only by 2028. The two claims are not contradictory. A single profitable quarter, especially one landing while the company absorbs tens of billions of dollars in new compute commitments, is not the same thing as sustained annual profitability. That gap is exactly the kind of detail that tends to get flattened in headline coverage of this IPO.
What the Big Numbers Leave Out
Several of the figures attached to this filing get repeated as plain facts when they are actually specific financial constructs with real boundaries. Reading them correctly changes how much weight each one should carry.
| Headline figure | What it actually measures | Why it matters |
|---|---|---|
| $965 billion valuation | The price new investors agreed to pay for new shares in the most recent round is not a market-clearing price for every share outstanding | Liquidation preferences attached to preferred shares mean common shareholders and employees would not necessarily receive a proportional share of that value in a downside scenario |
| $47 billion run-rate revenue | The most recent month's revenue multiplied by twelve, not audited trailing revenue | Run rate figures have a long history of overstating what eventually appears in audited financial statements, and the figure says nothing about margin after compute costs |
| Filed confidentially for an IPO | A draft registration submitted privately to the SEC that can be amended, delayed, or withdrawn without public notice | The only document that legally commits Anthropic to a listing on any date is a public S-1 that has been declared effective, and that document does not exist yet |
| Anthropic overtook OpenAI in valuation | A comparison of two private funding rounds priced weeks apart, not a measure of market share or active users | OpenAI's ChatGPT still reaches a far larger base of everyday consumers than Claude does, even though Anthropic leads in enterprise and coding revenue |
| Funds like AGIX or DXYZ give you Anthropic exposure | A small slice of a diversified basket of dozens of private companies, often bought at a price well above the fund's own reported net asset value | Explained in full later in this article, including how that gap has behaved around other recent listings |
Who Owns Anthropic and Who Gets Paid First
Anthropic has raised approximately $125 billion across its funding history, according to private market researcher Sacra. The capital structure has been built in distinct stages. A $13 billion Series F in September 2025 valued the company at $183 billion. A $30 billion Series G in February 2026, led by GIC and Coatue and co-led by D.E. Shaw Ventures, Dragoneer, Founders Fund, ICONIQ, and Abu Dhabi's MGX, pushed the valuation to $380 billion. The $65 billion Series H that closed in late May 2026, led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, brought it to $965 billion.
Other major backers across these rounds include Amazon and Google, each having invested roughly 8 billion dollars, Microsoft and Nvidia, which have committed up to $5 billion and $10 billion, respectively, alongside large compute supply deals, and Salesforce, which holds a stake reported to be worth around 5 billion dollars. BlackRock, Blackstone, Fidelity, Lightspeed Venture Partners, Temasek, TPG, and the Qatar Investment Authority have also participated in earlier rounds.
In February 2026, Anthropic ran an employee tender offer worth $5 to $6 billion, letting current and former staff sell shares at roughly the $350 billion valuation set earlier that month. Liquidity events like this are common at companies delaying a listing, and they are worth watching for again before any actual IPO, since they often signal how insiders themselves are pricing the stock ahead of public scrutiny.
Two of Anthropic's biggest investors, Amazon and Google, are also two of its most direct competitors in cloud infrastructure and AI models. That arrangement rarely gets questioned in coverage of the company, but it is a structural feature of how frontier AI gets funded right now. The companies with the deepest pockets and the most compute to sell are often the same companies a startup eventually has to compete against.
Anthropic, OpenAI, and Now SpaceX Are All Racing to Wall Street
Anthropic is not the only company rushing toward a listing this year. OpenAI confidentially filed its own draft registration around May 22, 2026, and confirmed the filing publicly on June 8, eight days after Anthropic's announcement. OpenAI was last valued at 852 billion dollars in a March 2026 funding round, and reports point to a possible listing window between September and November 2026, with Goldman Sachs, Morgan Stanley, and JPMorgan involved as underwriters.
Image Credit: Leonardo AI
SpaceX moved first. It confidentially filed in April 2026 and began trading on Nasdaq under the ticker SPCX in June, pricing around $135 a share in what is being described as one of the largest IPOs ever completed. How that debut performs in its first weeks of trading is being watched closely as a signal for whether public investors will pay private market prices for AI-adjacent companies. The mechanics of that listing, and who actually profits from it, are covered in If SpaceX Goes Public, Who Actually Gets Rich.
Anthropic's argument for deserving a premium over OpenAI rests mostly on capital efficiency. Multiple outlets covering both filings have pointed to Anthropic spending considerably less per dollar of revenue than OpenAI does, while OpenAI's own disclosures suggest it continues to lose more than it earns on every dollar of revenue, a gap it has been trying to close partly by renegotiating its compute contracts. Whether that efficiency gap survives the scrutiny of a public S-1 is one of the more interesting open questions heading into the back half of 2026, and it connects directly to the AI infrastructure spending race covered in Elon Musk Just Unveiled a 20 Billion Dollar Chip. Here Is Why It Matters.
The Pentagon Dispute Explained
One risk that gets a single line in most coverage of this filing deserves a full explanation, because it is genuinely unusual for a frontier AI company and could move Anthropic's government revenue by billions of dollars depending on how it resolves.
The dispute began in February 2026, after Dario Amodei met with Defense Secretary Pete Hegseth and the two sides failed to agree on how the Pentagon could use Claude. Anthropic had drawn two firm lines in its contract negotiations. It would not allow its models to be used in fully autonomous weapons without human oversight of targeting decisions, and it would not allow them to be used for domestic mass surveillance. On February 27, President Trump posted on social media, ordering federal agencies to stop using Anthropic's technology, and Hegseth announced the same day that the company would be designated a supply chain risk, a label historically reserved for firms tied to foreign adversaries. OpenAI struck its own deal with the Pentagon within hours of that announcement.
On March 4, the designation was formalized under two separate statutes, one covering the federal government broadly and one specific to the Department of Defense. Anthropic sued five days later in two courts at once, arguing the designation violated its First Amendment rights, denied it due process, and was issued without following the risk assessment and notification procedures Congress requires, according to legal filings summarized by Pearl Cohen. Researchers at OpenAI and Google DeepMind filed a personal capacity amicus brief supporting Anthropic, warning the designation could chill open discussion of AI safety industry-wide.
Image Credit: Leonardo AI
On March 26, federal judge Rita Lin issued a sweeping preliminary injunction in Anthropic's favor, writing in a 43-page ruling, as reported by CNN, that the record suggested the designation was pretextual and that the government's real motive was unlawful retaliation against the company's protected speech. The win was only partial. On April 8, a federal appeals court in Washington denied Anthropic's request to block the narrower, Department of Defense-specific version of the designation while the case continues, ruling, per CNBC's reporting, that military operational control outweighed the financial harm to a single private company.
The practical result, as of this writing, is a split decision. Anthropic remains excluded from direct Department of Defense contracts, and contractors working with the DOD cannot use Claude in that work, but Anthropic can continue operating across every other part of the federal government while the underlying legal fight plays out, likely well past any October listing.
This matters for an IPO prospectus in a specific way. Litigation unresolved at the time of listing has to be disclosed as an active risk factor, and the size of that risk depends entirely on which way the two separate court tracks eventually go. If Anthropic wins outright, the episode could become a genuine selling point, evidence that the company will hold safety commitments even under direct government pressure, which is the same trust argument that has won it customers in other regulated industries. If the designation survives in its current form, the company permanently loses a category of government revenue that competitors who accepted fewer restrictions do not have to give up.
The Risks an Honest Investor Should Sit With
Beyond the Pentagon dispute, four other risks are worth holding onto before treating this filing as a straightforward growth story.
- Cost risk. Anthropic has committed to tens of billions of dollars in compute and data center spending, including a long-term GPU supply agreement with SpaceX worth roughly 1.25 billion dollars a month through May 2029, and reported plans to raise up to $36 billion in chip-linked debt on top of its equity raises. Spending of this size makes the margin far more sensitive to chip pricing and utilization than a typical software business.
- Competition risk. Google, Meta, and Elon Musk's xAI are all shipping competing models on a similar cadence, and enterprise buyers can and do switch providers when pricing or performance shifts.
- Pricing power. Anthropic's net revenue retention has been reported well above typical enterprise software benchmarks, but that figure can compress quickly if cheaper open-weight or in-house models start covering more of the same workloads.
- Margin timeline. Anthropic's gross margin reportedly moved from deeply negative in 2025 toward roughly 40 percent in 2026, with an internal target near 77 percent by 2028. Missing that 40 percent milestone in its first quarter as a public company would reopen the entire valuation debate immediately.
A Framework for Valuing a Company With No Real Comparable
A revenue multiple is the fastest way to argue about whether $965 billion is reasonable, and it is also close to meaningless on its own. At roughly 20 times run-rate revenue, Anthropic trades cheaper than some peak-era software companies and richer than almost every mature enterprise software business on the public market today. That single number compresses three variables that are all moving at once: how fast revenue grows, what share of it survives as gross margin once compute is paid for, and how sticky existing customers turn out to be.
A more useful way to think about it is to build three explicit scenarios rather than rely on one multiple. In a bull case, gross margin keeps climbing toward management's 77 percent target by 2028, net revenue retention holds near its reported 140 percent level, and the Pentagon litigation resolves in Anthropic's favor, reinforcing the trust advantage that wins it regulated industry customers. In a base case, margin improvement slows as compute costs keep pace with revenue growth, net revenue retention drifts toward the 110 to 120 percent range considered normal for mature enterprise software, and the Department of Defense designation stands while other government work continues unaffected. In a bear case, a wave of cheaper competing models compresses pricing across the industry at the same time Anthropic's growth rate decelerates the way most hypergrowth software companies eventually do, roughly halving every twelve to eighteen months after an inflection point, pushing the 2028 revenue base that the trillion-dollar narrative depends on meaningfully lower.
The single variable that dominates all three scenarios is not the valuation multiple itself. It is whether Anthropic's growth rate decelerates gently or sharply over the next two years, because that one assumption changes the 2028 revenue base by tens of billions of dollars, and everything else in the model follows from it.
How to Get Exposure Before the Listing
Retail investors cannot buy Anthropic shares directly until the company actually lists. Two publicly traded vehicles currently offer indirect exposure: the KraneShares AGIX ETF and the closed-end fund Destiny Tech100, traded under the ticker DXYZ.
Both carry a mechanic that gets glossed over in most coverage. Closed-end vehicles like DXYZ often trade well above their own reported net asset value, simply because shares offering exposure to companies like SpaceX, OpenAI, and Anthropic are scarce. DXYZ reported a net asset value of $19.97 per share at the end of 2025, while its market price has at various points traded at a premium ranging from roughly 50 percent to well over double that figure, according to market data tracked by Benzinga. The fund added roughly $127 million of exposure to Anthropic alongside positions in CHAOS Industries and Hermeus in early 2026, on top of an existing stake in SpaceX that made up its single largest holding.
That premium has nothing to do with how well Anthropic, or any of the fund's other holdings, actually perform. It reflects scarcity, and scarcity premiums tend to compress the moment the underlying company everyone wants exposure to actually starts trading on its own. Analysts covering DXYZ flagged exactly this question once SpaceX's own listing arrived, namely whether a fund built around exclusive access to private companies stays just as relevant once its biggest holding is available to anyone with a brokerage account.
| Route | What you actually get | Key thing to know |
|---|---|---|
| KraneShares AGIX | A diversified AI-themed ETF with a stake in Anthropic among many holdings | Exposure is spread across the fund's full portfolio, not concentrated in Anthropic alone |
| Destiny Tech100, DXYZ | A closed-end fund holding direct stakes in Anthropic, SpaceX, OpenAI, and dozens of other private companies. | Has historically traded at a steep premium to its own reported net asset value |
| Waiting for the public listing | Direct ownership of Anthropic stock once it actually trades | Comes with no scarcity premium, but only arrives after the company prices and lists |
Three further mechanics are worth knowing before buying into either vehicle. The stated net asset value itself is an estimate based on the latest funding round and limited secondary trades, not a live market price, so it can lag what shares would actually fetch if sold today. Both funds hold dozens of positions, which means a dollar invested in the ticker is not a dollar of pure Anthropic exposure. Once Anthropic does list, its own early shareholders, including employees, will typically face lockup periods of ninety to one hundred eighty days before they can sell, and share prices for newly public companies have historically tended to soften around the point those lockups expire, regardless of how strong the opening days of trading looked.
Frequently Asked Questions
Is there an actual Anthropic IPO date yet? No fixed date exists. Anthropic confidentially filed a draft registration statement on June 1, 2026. Reports point to an October 2026 target on the Nasdaq, but the company has not confirmed a date, and the public version of its filing, the document that would make a date credible, has not yet appeared on the SEC's EDGAR system.
What is Anthropic's valuation today? Anthropic's most recent private valuation is $965 billion, set by a $65 billion Series H funding round that closed in late May 2026, led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital.
Is Anthropic profitable in 2026? Anthropic has said it is close to reporting its first profitable quarter, a faster timeline than its earlier 2028 full-year profitability target. A single profitable quarter does not yet confirm sustained annual profitability, particularly given the scale of new compute and data center commitments the company has taken on.
How can I invest in Anthropic before the IPO? Retail investors cannot buy Anthropic shares directly before the listing. The KraneShares AGIX ETF and the closed-end fund Destiny Tech100, traded as DXYZ, both currently hold positions in Anthropic, though both come with the tradeoffs explained earlier in this article.
Will Anthropic or OpenAI list first? Anthropic filed confidentially first, on June 1, 2026, with OpenAI following on June 8. Anthropic is reportedly targeting an October 2026 listing, while OpenAI's reported window runs from September through November 2026, so the order is not yet settled.
What is the Pentagon dispute about? The US Department of Defense designated Anthropic a supply chain risk in early March 2026 after the company refused to remove restrictions on using Claude for autonomous weapons and domestic surveillance. A federal court has partially sided with Anthropic, but the company remains excluded from direct Department of Defense contracts while the case continues.
Desidaily Take
Strip away the headline numbers, and Anthropic's case rests on something fairly simple. Revenue has grown faster than almost any enterprise software company on record, and it has done so while spending less per dollar of revenue than its closest rival. Eight of the Fortune 10 are paying customers, and the company is reportedly close to its first profitable quarter, years ahead of its original target.
The case for caution is just as concrete. The valuation jumped from $380 billion to $965 billion in about fifteen weeks, a repricing speed that has reportedly made even some of the venture investors involved uneasy. The Pentagon dispute remains legally unresolved and could permanently remove a category of government revenue. The company is simultaneously taking on tens of billions of dollars in compute and chip-linked debt, and it is racing two other AI giants, OpenAI and SpaceX, to the same pool of public market capital within the same few months.
Neither story cancels the other out. Both are true at the same time, and which one ends up mattering more will only become clear once Anthropic's actual audited numbers, not its run rate estimates, are sitting in a public S-1 on EDGAR. Until then, every valuation argument about this company, including the ones in this article, is built on private market numbers that have not yet been tested by public disclosure. This is not investment advice, and anyone considering exposure to Anthropic through AGIX, DXYZ, or a future listing should weigh these facts against their own financial circumstances.
Related reading on DesiDaily
- Claude AI Is Quietly Replacing Major Software Tools in 2026
- If SpaceX Goes Public, Who Actually Gets Rich
- AI Is Suddenly Everywhere. Here Is What Is Really Happening
- Imagine Opening Your Monthly AI Bill and Seeing This Number
- Elon Musk Just Unveiled a 20-Billion-Dollar Chip. Here Is Why It Matters
- 24,000 Accounts, One AI Target: What Claude Faced
- The Quiet Reset Inside Meta and Why It Matters for AI
- AI in Warfare: What It Can Do and What It Cannot
- The First Major AI Defense Breakup: What It Means