Anthropic IPO Explained: ANT Stock, Valuation and the Risks Behind Claude

Anthropic’s proposed IPO puts a remarkable business in front of public investors: the developer of Claude generated $41 million in annual revenue in 2023 and $11.555 billion in the second quarter of 2026 alone. The harder question is what ownership in that business will cost—and what rights and obligations will come with it.

The October 26 S-1 filing opens the books on the company’s growth, losses, computing commitments and governance. It does not yet supply the final IPO price, offering size or first trading date. As of October 28, the latest public SEC filing remains that initial registration statement.

This is the distinction at the center of the Anthropic IPO. Demand for Claude is already measurable. The attractiveness of ANT shares will depend on the price, dilution, capital requirements and shareholder protections attached to that demand. A fast-growing company can still be an expensive investment.

The following guide starts with the offering itself, then examines the revenue engine, financing and risks investors would inherit. Claude’s products and research matter throughout, because they must ultimately support the cash flows on which any public-market valuation rests.

Where the Anthropic IPO Stands on October 28

Anthropic intends to list Class A common stock on the Nasdaq Global Select Market under the proposed symbol ANT. Its IPO announcement confirms the public filing and says the number of shares and price range have not yet been determined. Goldman Sachs and Morgan Stanley are the lead representatives named for the offering.

The SEC’s current filing record, checked for this October 28 publication, lists the October 26 S-1 as the latest submission. It does not show a subsequent public pricing amendment or notice of effectiveness. There is therefore no public ANT trading price to quote at the research cutoff.

Anthropic IPO: Confirmed Details and Open Terms

IPO Item	Status at the October 28 Check
Public registration	Form S-1 filed October 26; latest filing checked October 28
Proposed exchange and ticker	Nasdaq Global Select Market · ANT
Offered security	Class A common stock
Lead representatives	Goldman Sachs and Morgan Stanley
IPO price range and share count	Not yet set in the public filing
First trading date	Not announced by the company
SEC effectiveness	No effectiveness notice in the checked filing record
Public ANT stock price	No public trading price at the research cutoff

Reuters reported before the public filing that Anthropic was pressing ahead with a November offering and expected to begin marketing shares during the week of November 9. Its sources cautioned that the plans and dates could change. That timetable remains reporting about preparations, rather than a company-confirmed first trading day.

An S-1 is a disclosure document in a process. Amendments can add financial information, identify the selling shareholders more fully and set an expected price range. Pricing and the final prospectus then establish the actual transaction. SEC effectiveness allows an offering to proceed; it is not an endorsement of the investment’s merits.

For readers tracking the sequence, our filing coverage records the public registration, while earlier timetable reporting explains the preparations. Keeping those stages distinct prevents a reported target from turning into an invented listing date.

What a Trillion-Dollar Valuation Would Require

Anthropic’s last announced major private funding round provides one reference point. Its May 28 Series H raised $65 billion at a $965 billion post-money valuation. Private preferred securities and a future public Class A share are not automatically comparable investments, even when the headline company values can be placed side by side.

Reuters reported that some investors considered a potential $1.5 trillion valuation. Other offering targets have circulated at still higher levels, as covered in our offering analysis. None of those discussions supplies a final IPO valuation. A large expected fundraise also should not be confused with the total value of the company.

The sensitivity to price is substantial. The table below uses hypothetical equity values to illustrate how the denominator changes the story. It compares each value with completed 2025 revenue and with twice the revenue recorded in the first half of 2026. The latter is a mechanical annualization, not a forecast.

How the Revenue Denominator Changes Valuation Multiples

Hypothetical Equity Value	Value / 2025 Sales	Value / Annualized H1 2026 Sales
$1.0 trillion	218.1×	30.7×
$1.5 trillion	327.2×	46.1×
$2.0 trillion	436.2×	61.4×

These are equity-value-to-sales illustrations, not enterprise-value multiples or estimates of fair value. They do not adjust for cash, debt, preferred claims or the conversion terms that will shape the post-IPO company. They also do not imply that Anthropic will be offered at any of the illustrated values.

The exercise shows why “revenue grew 112 times” cannot by itself answer whether shares are attractive. A valuation can already anticipate years of expansion. The return earned by a new shareholder depends on the gap between the results ultimately delivered and the results embedded in the price paid.

A high-growth company also presents a denominator problem: last year’s sales can understate its current scale, while annualizing a recent period can overstate the durability of that pace. The eventual IPO case needs a credible operating forecast, a realistic cost structure and an explicit account of uncertainty. Neither an old revenue multiple nor an optimistic run rate is enough on its own.

What Buyers of ANT Would Actually Own

The proposed shares are an interest in Anthropic, PBC, the company behind Claude. They are not a claim on one chatbot, one model release or a separate pool of subscription revenue. Shareholders would own exposure to the whole business, including its research, applications, platform relationships and financial obligations.

Anthropic was founded in 2021 by Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah. The founding team emerged from OpenAI. Dario is chief executive; Daniela is president and board chair. The organization combines a frontier-model laboratory with a rapidly expanding commercial operation.

There are three connected layers to that operation. The models provide the underlying intelligence. Claude Platform lets developers build applications using it. First-party products, including Claude Code and the evolving Claude.ai and Cowork experience, supply ready-made ways to put it to work.

This combination offers multiple routes to revenue. An individual can subscribe, an organization can deploy Claude internally and a software developer can build a product that repeatedly calls the models. The final user may encounter Anthropic through a cloud provider or another application without maintaining a direct relationship with the company.

The economic appeal is breadth: one research improvement can support several products and customer workflows. The accompanying risk is breadth of responsibility. Investors would be financing continuing model development, infrastructure access, distribution and safe deployment, rather than buying a mature product whose development cost is already behind it.

Class A ownership also needs to be separated from control. Anthropic’s planned voting structure gives other share classes and the Long-Term Benefit Trust powers that public investors will not obtain simply by buying more Class A stock.

Who Would Control ANT After the IPO?

Anthropic’s public benefit status means its stated purpose extends beyond maximizing shareholder returns alone. Its governance also separates economic ownership, general voting power and the authority to choose directors. Buying a public share will not necessarily provide proportionate influence over every important decision.

The planned structure gives a single Class F share, held through an entity initially owned by the seven cofounders, 50.1% of voting power on matters submitted to shareholders generally. That share has no economic rights. Class A shares carry one vote each, while Class N shares, held primarily by strategic partners, generally have no voting rights. Employees other than the cofounders would also receive non-economic Class E shares, with voting rights on limited matters.

A separate Class T share belongs to the Long-Term Benefit Trust, which generally has the power to appoint and remove a majority of authorized directors. The company’s April board announcement confirmed a trust-appointed majority after Vas Narasimhan joined. The trust is not the same thing as the founders’ voting entity.

The structure is intended to protect long-term decision-making and the public-benefit mission. The tradeoff for minority investors is that a large economic stake need not translate into control over corporate direction. The terms of the shares and the appointment rights matter more than a simple ownership percentage.

Another separate mechanism is the proposed Public Benefit Fund. In its October 18 fund announcement, Anthropic said an initial 5% share commitment was subject to stockholder approval, with possible additional allocations linked to valuation milestones and board discretion. An equity commitment is not the same as cash already distributed.

For readers following that initiative, our fund coverage explains the distinction between the new vehicle and earlier founder and employee pledges. Governance, philanthropy and commercial performance are related parts of Anthropic’s identity, but they should be evaluated on their own terms.

The Revenue Growth Is Real, but the Periods Matter

The annual figures show revenue rising from $41 million in 2023 to $386 million in 2024 and $4.585 billion in 2025. The 2025 total was approximately 112 times the 2023 figure. That is a comparison between two full financial years, rather than a claim that every intervening month grew at the same rate.

The acceleration continued in the first half of 2026. Revenue reached $16.286 billion, compared with $1.163 billion a year earlier. The June quarter alone produced $11.555 billion, up from $787 million in the June 2025 quarter. That quarterly comparison supports the filing’s approximately 1,368% growth figure.

Revenue by Reporting Period

Period	Revenue USD billions	Period Length
2023	$0.041	Full year
2024	$0.386	Full year
2025	$4.585	Full year
H1 2025	$1.163	Six months
H1 2026	$16.286	Six months
Q2 2025	$0.787	Three months
Q2 2026	$11.555	Three months

Annual revenue, quarterly revenue and a revenue run rate answer different questions. Anthropic’s defined run rate annualizes recent consumption fees and current subscription fees. It is useful for describing the latest pace, but it is neither revenue already earned over a completed year nor a forecast guaranteed to materialize.

The distinction is especially important when growth is this fast. Multiplying a recent month or quarter can produce an impressive annual number, while saying little about the sequence of demand, capacity and spending that will follow. Seasonal usage, new releases and unusually large customers can all affect the observation.

There is also a difference between public financial statements and later management updates. MarketSpeaker’s September-quarter report covered figures communicated to investors. The October 26 filing’s detailed historical statements run through June, alongside selected September operating disclosures. Those sources should not be treated as identical sets of financial statements.

The company itself warns that recent growth rates are unlikely to persist. A larger business can add more dollars while growing more slowly in percentage terms. The meaningful test is whether customer expansion and improving economics remain strong as the comparison base becomes much larger.

What 900% Net Revenue Retention Really Means

The prospectus’s approximately 900% net revenue retention is one of its most unusual disclosures. It measures expansion, contraction and churn within a defined existing customer cohort. It does not mean Anthropic retained nine customers for every one it previously had.

The definition compares revenue from qualified customers in the three months ending on the measurement date with revenue from those same customers in the corresponding quarter a year earlier. Qualified customers generated at least $1 in the base period. Fixed-price Claude Pro and Max subscribers are excluded.

For illustration, if that cohort generated $1 million in the earlier quarter and $9 million in the later quarter, net revenue retention would be 900%. Revenue from the cohort would be nine times as large—an 800% increase. New customers outside the original cohort are not what produces the expansion in this calculation.

It is tempting to translate this into “a $10,000 customer becomes a $90,000 customer.” That can illustrate the ratio, but it is not a statement about the average or typical individual account. A few very large expansions can affect the aggregate substantially, and customers within the cohort can shrink or leave.

At 100%, the cohort produces the same revenue after expansion, contraction and churn. Above 100%, expansion more than offsets the revenue lost elsewhere in the cohort. That makes 100% a mathematical reference point, rather than a universal gold standard for every software business.

The Anthropic figure is powerful evidence that existing usage expanded from a much smaller base. It does not establish permanent pricing power, identical retention across products or another ninefold expansion next year. Future disclosures will show how the measure behaves as early deployments mature.

Who Is Paying Millions for Claude?

As of September 30, Anthropic counted more than 11,900 organizational customers generating at least $100,000 in trailing-twelve-month revenue. More than 1,900 generated at least $1 million, and approximately 250 generated at least $10 million.

These are nested thresholds. The $10 million customers are included in the million-dollar group, which is included in the $100,000 group. Adding the three counts would double-count customers. They describe revenue recognized over the preceding twelve months, rather than three separate piles of annual subscription contracts.

Customers by Trailing-Twelve-Month Revenue

Revenue per Customer	Customer Count
$100,000 or more	More than 11,900
$1 million or more	More than 1,900
$10 million or more	Approximately 250

The filing also clarifies that Anthropic does not separately define and internally track an “enterprise” customer category. It uses the $100,000 threshold as a measure of enterprise adoption because spending at that level generally reflects organizational use. The shorthand should not be mistaken for a rigid commercial segmentation rule.

What can justify such spending? A company might embed Claude in a product used by thousands of its own customers, run a large software-development operation or automate a high-volume internal process. Millions of dollars can represent many small units of useful work, rather than an exceptionally expensive chat subscription.

One published customer example is Novo Nordisk’s NovoScribe documentation platform, built with Claude Code, Amazon Bedrock and MongoDB Atlas. The case study describes drafting clinical documentation and other materials for human review. Its reported time savings come from that particular deployment; they are not proof that regulatory approval or drug development can be automated wholesale.

Large accounts also create concentration risk. The top five customers represented approximately 25% of revenue in the first nine months of 2026, although no single customer exceeded 10% over that period. Wide adoption and dependence on several very large buyers can exist at the same time.

The $80.7 Billion Loss and the Operating Business

The first-half accounts contain three numbers that can appear incompatible: a $693 million operating loss, an $80.748 billion net loss and $3.929 billion of cash provided by operating activities. They describe different parts of the financial picture.

Operating results compare revenue with operating expenses. Net income includes additional items, including financing-related changes in value. The cash-flow statement then reconciles accounting earnings with actual cash movements and changes in operating balances. None can be substituted for the others.

Three Views of First-Half 2026 Performance

Measure	USD billions	What It Measures
Revenue	$16.286	Services delivered during the period
Operating expenses	$16.979	Operating resources and costs recognized
Operating loss	($0.693)	Revenue less operating expenses
Financing fair-value charge	$72.388	Changes in notes, derivatives and SAFEs
Net loss	($80.748)	Result after financing and other items
Operating cash flow	$3.929	Cash provided by operating activities
Adjusted operating income	$1.618	Company-defined non-GAAP measure

The first-half net loss included a $72.388 billion charge for changes in the fair value of convertible notes, derivative liabilities and SAFEs. A financing instrument can become more valuable to its holder as the company’s valuation rises, creating an accounting expense for the issuer without an equivalent cash payment in that period.

That is why the net-loss headline should not be read as $80.7 billion spent running Claude. Equally, removing the largest charge does not establish a particular level of sustainable cash generation. Other costs and financing effects remain, and future investment can require substantial cash.

The June quarter did produce $974 million of operating income, after a $1.667 billion operating loss in the March quarter. Anthropic also reported non-GAAP adjusted operating income of $1.618 billion for the first half. That adjusted measure excludes specified items, including stock compensation; it should be assessed alongside the reconciliation, not used to erase the GAAP result.

The filing places compute and infrastructure within operating expenses, covering research, model development, customer serving and internal use. Its presentation should not be casually converted into a conventional software gross-margin comparison. Our accounting analysis examines the separate questions of note valuations, operating costs and dilution.

Where the IPO Money Would Go

The prospectus describes several purposes for the offering: increasing capitalization and financial flexibility, creating a public market and facilitating an orderly distribution of shares for selling stockholders. Those are related objectives, but they send money to different destinations.

New shares sold by Anthropic would raise capital for the company, after underwriting discounts and offering expenses. Shares sold by existing holders would provide liquidity to those holders. Anthropic states that it will receive no proceeds from those secondary sales. A headline deal size therefore cannot automatically be treated as fresh cash available to build Claude.

The intended uses of company proceeds include working capital, operating expenses, compute and infrastructure investment, and hiring and retention. The filing also permits acquisitions or investments, while saying there were no agreements or commitments for material acquisitions or investments at the time described.

Another intended use is tax withholding and remittance associated with settling certain restricted stock units. That matters because a portion of the cash raised can support employee-equity obligations rather than new computing capacity. The amount depends in part on the eventual offering price and award settlement.

Management would retain broad discretion over proceeds. Investors should therefore connect the financing plan with the capacity timetable and expected operating cash generation, rather than assume every IPO dollar corresponds to a particular new asset.

A simple illustration separates the arithmetic. If a hypothetical offering sold $10 billion of newly issued shares and $2 billion of existing shares, gross deal size would be $12 billion, but the company would receive only the primary portion before its applicable fees and costs. These are explanatory numbers, not proposed Anthropic terms.

The eventual prospectus should make the primary and secondary mix, net proceeds and planned capital structure more concrete. Those details will reveal how much additional financial flexibility the IPO actually creates.

The Computing Bill Behind the Software

Anthropic’s most important physical dependency is computing capacity. Training advances the models; inference runs them for users. Both require infrastructure that has to be financed, supplied and powered. A software interface can scale quickly only if the systems underneath it can support the demand.

Compute and infrastructure expense reached $8.945 billion in the first half of 2026. In the June quarter, it was $5.558 billion, or approximately 48% of revenue. A year earlier, it had exceeded revenue. Rapid growth therefore improved the relationship between usage revenue and the compute expense recorded in those periods.

The improvement is encouraging, but the timing is important. Anthropic says it typically procures compute 12 to 18 months before availability. Costs can rise when committed capacity comes online, while customer demand may arrive earlier or later than expected. Short-term capacity can fill a gap, generally at a higher cost per unit.

The prospectus describes SpaceXAI arrangements worth up to $120.5 billion through terms extending into 2029. Once all capacity is delivered, the aggregate monthly amount would be approximately $3.1 billion, with different fees during ramp-up. Cancellation provisions and initial non-cancelable periods matter; the maximum is not an immediate cash bill.

Separate Broadcom equipment leases involve approximately $161.2 billion of rent over their lives, excluding interest. These obligations differ from purchases of cloud services, current-period expense and notes already outstanding. Adding every disclosed number without checking timing, overlap and contract type would produce a misleading total.

The financial challenge is a mismatch between relatively flexible customer consumption and commitments extending years into the future. Strong demand can make early access to capacity valuable. Weak demand can leave costly capacity underused. The company’s infrastructure commitments are central to evaluating the IPO’s eventual use of capital.

Its Partners Are Also Investors and Competitors

Anthropic’s May 28 funding announcement described a $65 billion Series H round at a $965 billion post-money valuation. That was a private financing benchmark with its own securities and terms, rather than an exchange price available to public investors.

The announcement also described the company’s relationships with Amazon, Google, Broadcom and SpaceX, and identified AWS as its primary cloud provider and training partner. These arrangements help explain how a company with a comparatively small workforce can obtain infrastructure at extraordinary scale.

But the relationships are not simple supplier contracts. Strategic partners may supply compute, distribute Claude, invest capital and sell competing AI products. Anthropic benefits from their reach while remaining exposed to their priorities, pricing, capacity allocation and ability to develop alternatives.

The right question is what each transaction does. Investment supplies financing. A compute contract buys access to resources. A distribution agreement helps reach customers. Revenue still depends on providing a service under the applicable accounting rules. The existence of linked relationships warrants scrutiny, but does not by itself establish artificial demand or improper accounting.

At June 30, Anthropic held $131.651 billion in cash, cash equivalents and short-term investments. Its balance sheet also carried $108.566 billion of convertible notes. That carrying value is not the same as the original cash principal borrowed. Conversion terms affect ownership and dilution, rather than simply representing a bill that must all be paid back in cash at the IPO.

For prospective shareholders, the eventual capitalization table matters as much as the financing headlines. Preferred stock, convertible instruments, employee awards and newly sold shares determine the economic interest remaining for each group. A large fundraise can strengthen the company while changing the fraction owned by everyone else.

Selling Shareholders and Lockups Could Shape Trading

A public company can have a very large total valuation while only a limited portion of its shares is available for immediate trading. The offer size, existing holders’ sales, employee liquidity and restrictions on other shares determine that initial supply.

Anthropic’s preliminary prospectus contemplates sales by both the company and selling stockholders, but the transaction quantities remain blank. Until those terms are filled in, investors cannot calculate the exact initial public float or how much of the offering represents existing investors cashing out.

The lockup provisions are more detailed than a single six-month expiration. The company’s own restrictions generally run until the earlier of a specified point after its first-quarter 2027 results release or the 180th day after the prospectus date, subject to exceptions and possible consent from the underwriters’ representatives.

Different holder groups have different early-release schedules. For example, certain non-founder directors, officers and current or former employees or consultants may transfer up to 10% of their specified holdings when trading begins, with up to another 35% after the full-year 2026 results release under the stated conditions.

Some other holders have staged releases tied to elapsed days and earnings announcements. Certain holders above 0.5% of pre-offering fully diluted shares have a longer baseline tied to third-quarter 2027 results or the 365th day, again with early-release provisions. These are descriptions of the preliminary filing, not a calendar of confirmed sale dates.

The practical implication is a sequence of potential changes in tradable supply. A release allows eligible stock to be sold; it does not require holders to sell or establish that the price will fall. Demand, performance and the broader market can outweigh the effect of additional supply.

Founders’ and employees’ charitable commitments add another reason to read the detailed terms. A pledge, a permitted transfer and an open-market sale are different events. The final offering documents and subsequent disclosures will be needed to connect any of them with actual public trading.

Can Claude Build the Economics the IPO Needs?

The most durable argument for the IPO is that Claude becomes valuable enough in repeatable work for customers to keep expanding their use. That claim has to survive competition and the cost of delivering the service, not merely a succession of impressive demonstrations.

Claude’s model families occupy different combinations of capability, speed and cost, including Haiku, Sonnet and Opus, with more advanced systems described in the Mythos class. A customer may use cheaper models for routine work and more capable ones for difficult tasks. The relevant outcome is reliable work completed at an acceptable total cost.

The September Opus release illustrates the dynamic. Anthropic said Opus 5.5 reduced costs on typical workloads by approximately 40% compared with Opus 5. That is a company-reported comparison, rather than a universal saving. Efficiency can improve serving economics and encourage more usage, while competition can push some of the benefit into lower customer prices.

Claude Code, introduced as a research preview in February 2025 alongside Claude 3.7, brought agentic coding into the terminal. Cowork extended the approach to broader knowledge work in January 2026. The filing says Anthropic began merging Cowork and Claude.ai in September.

Agents can make repeated model calls while inspecting documents, changing software and checking results. That offers a route to much greater consumption than occasional questions. It also raises the reliability standard: a system permitted to act needs controls, review and recovery procedures.

Connections to company data help turn general models into useful services. Anthropic’s open MCP standard was introduced in November 2024 to connect AI applications with information and tools. Existing integrations and tested workflows can make a platform more useful and more expensive to replace, while open standards can also help customers move between suppliers.

OpenAI, Google, other developers and open-weight models limit how much of that customer value Anthropic can keep. Our competition coverage follows one distribution channel, not the whole market. For the IPO, the key question is whether differentiation can support attractive margins as alternatives improve.

The People and Costs Behind the Revenue per Employee

Anthropic reported approximately 5,300 employees worldwide as of September 30, with a majority working in research, engineering and product development. It also disclosed trailing-twelve-month revenue per employee above $6.6 million, up from $1.4 million a year earlier.

That is an exceptional sales-to-headcount ratio. It is not profit per employee, compensation per employee or a direct measurement of any individual’s productivity. A rapidly growing revenue numerator and a point-in-time headcount denominator can produce a ratio that moves dramatically.

The organizational boundary matters too. Many people manufacturing chips, supplying electricity and operating the data centers used by Anthropic are employed elsewhere. Their work and the associated costs do not disappear because they are outside Anthropic’s own payroll.

The company is also an intensive user of its own technology. It says Claude wrote more than 90% of contributions to its codebase in September. That is a company-reported measure of code contributions, not proof that 90% of engineering judgment, product design or successful delivery has been automated.

Hiring can remain expensive even in an efficient organization. First-half personnel expense was $3.152 billion, including substantial stock-based compensation. Equity awards can reduce immediate cash requirements while diluting ownership and creating future accounting and tax consequences.

Talent retention involves more than competing on salary. Researchers and engineers also choose projects, colleagues and a mission they want to pursue. As the organization grows, preserving that attraction can become harder. The durable question is whether Anthropic can retain the people needed to keep improving Claude at a cost the business can support.

Safety Is Both the Proposition and a Business Risk

Anthropic sells the usefulness of capable models while warning about their limitations and potential harms. That tension is part of the company’s strategy, rather than an issue that can be separated neatly from the product. A system trusted with important work must be evaluated for what it can get wrong.

Its system cards document capabilities, evaluations and deployment decisions for particular models. The transparency disclosures describe measures such as misuse detection, account enforcement and defenses against attempts to bypass safeguards. These are descriptions of the company’s approach, not guarantees that failure is impossible.

For an ordinary business deployment, the immediate risks include incorrect outputs, confidential information being exposed and an agent taking an unwanted action. More capable systems can also create concerns involving cybersecurity, biological misuse and behavior that is difficult to evaluate. Capability and deployment safeguards need to be considered together.

Safety measures have commercial consequences. Restricting access, delaying a release or terminating abusive usage may protect users and reduce liability, but can also affect sales. Anthropic estimates that customers later judged to violate its usage policy or commercial terms generated approximately 5% to 7% of revenue in the first nine months of 2026.

That estimate does not establish a certain future loss of the same percentage. Detection, enforcement and customer behavior can change. It does show why revenue quality matters alongside growth: usage that cannot be responsibly retained may be less durable than usage from a well-governed production deployment.

Copyright and government policy add further uncertainty. The filing records a $1.5 billion Bartz litigation settlement charge in 2025 and describes continuing legal and regulatory risks. A recorded settlement expense addresses a particular matter; it does not settle every future question about training data, generated outputs or where advanced models may be supplied.

Offer Price, Opening Price and Pre-IPO Exposure

The SEC’s IPO guide distinguishes buying in an offering from buying in the public market. The IPO offer price is the price set for shares sold in the offering. The opening market price is established when public trading begins. They can differ, and a later buyer’s entry price can differ from both. A headline first-day gain does not mean every interested investor could have bought at the offer price.

Access to an IPO allocation depends on the broker, distribution arrangements, eligibility and demand. Filing an S-1 does not make the shares available through every brokerage account. Until the final terms and distribution are established, a promise of easy access should not be confused with an official allocation.

Pre-IPO investments need a separate analysis of the instrument. A private share, an interest in a fund or special-purpose vehicle, and a token or contractual claim marketed as exposure can carry different ownership, transfer, fee and redemption terms. The same company name on a screen does not make those rights identical.

MarketSpeaker’s token coverage examines one such route to marketed exposure. Those arrangements should not be displayed as a live Nasdaq quote for ANT or assumed to convert automatically into the Class A shares described in the prospectus.

Once trading begins, the relevant comparison remains the ownership acquired for the price paid. An order placed after a large opening move is an investment at that new price, not at the private-round valuation or IPO allocation price used in earlier headlines.

That distinction also explains why a successful listing and an attractive purchase can be different outcomes. Anthropic could raise substantial capital and gain a liquid market while buyers at different prices experience very different returns.

The IPO Disclosures That Matter Next

The first checkpoint is a public amendment setting the price range, share count and primary-secondary split. It should make the expected proceeds, dilution and post-offering ownership easier to calculate. Any change in the timetable should be attributed to an official announcement or clearly identified reporting.

The second checkpoint is the financial information available by pricing. Detailed first-half accounts and selected September metrics do not replace a full set of later results. Updated revenue, operating income, cash flow and capacity commitments would help test whether the rapid expansion is translating into durable economics.

The third is the capital plan after the IPO. Track capacity coming online, compute expense, utilization and liquidity together. A larger cash balance can buy time, but it does not remove the need to earn an adequate return on the infrastructure secured.

Customer quality is another test. Retention, large-account counts, concentration and the effect of enforcing usage policies should be read together. A remarkable expansion ratio can coexist with reliance on a handful of customers or with growth that becomes harder to sustain from a much larger base.

Governance and tradable supply will matter after the opening bell. The rights attached to ANT, the founders’ and trust’s powers, convertible instruments and phased lockup releases shape the investment beyond the first day’s performance.

Anthropic has already demonstrated that its technology can generate substantial commercial demand. The IPO will test what price public investors are willing to pay for that demand, how much capital the company retains and whether future results justify the expectations. The answer will come from the final terms and subsequent execution, not from the size of the growth figures alone.

---

Reporting note: Prepared and verified October 28, 2026, against Anthropic’s October 26 S-1, the SEC company filing record and dated company disclosures. The SEC record showed no subsequent public pricing amendment or effectiveness notice at the research cutoff. Interim financial figures are unaudited. Valuation multiples and the proceeds example are MarketSpeaker illustrations, not offering terms or forecasts. No public ANT trading price was available.