The Anti-Sell — Anthropic’s S-1, read in two columns

One piece of advice, which we heard in slightly different forms from multiple people, stuck with us: never let your company be worth a dollar more than it is actually worth.

Document Stories

Form S-1 · filed with the SEC on October 26, 2026 · proposed Nasdaq symbol ANT

Revenue · FY2023
$41M
Year ended Dec 31, 2023 · $ millions · GAAPS-1 · p.33 · L.913
Revenue · FY2024
$386M
Year ended Dec 31, 2024 · $ millions · GAAPS-1 · p.33 · L.914
Revenue · FY2025
$4,585M
Year ended Dec 31, 2025 · $ millions · GAAPS-1 · p.33 · L.915
Revenue · six months to Jun 30, 2026
$16,286M
Six months ended Jun 30, 2026 · $ millions · vs $1,163M a year earlierS-1 · p.33 · L.917

Anthropic’s S-1, read in two columns

01 — Four Things

$200M
Committed to economic-response studies · Committed “to partner with governments to run studies on different ways of responding to changes in the economy” · $ millions · founders’ letter · S-1 · p.6 · L.204

There are four things that we would like any investor in Anthropic to understand: the incredible potential of AI, the risks we must address, Anthropic's unique approach to building a company and operating in the world, and the ways in which our behavior may differ from that of a purely profit-maximizing company.

founders’ letter · S-1 · p.3 · L.154

1 of 4 — AI's potential

4.6 billion

People without essential health services · Estimated people lacking access to essential health services, “more than half the world's population”, per WHO and the World Bank (letter) · founders’ letter · S-1 · p.4 · L.178

Parser-described artwork: red bacterial colonies in a petri dish; a cross-section of a biological structure with cellular compartments; a vintage celestial map showing constellations and astronomical coordinates.
p.3 · founders’ letter · render
WHO daily situation reportfounders’ letter · S-1 · p.5 · L.182

A full day → under an hour

Prep time for the WHO Regional Office for Africa's daily report, 2026 Ebola outbreak (DRC)

Free clinical search for doctorsfounders’ letter · S-1 · p.5 · L.182

~100 countries

Free AI-powered clinical search for doctors in “roughly 100” low- and middle-income countries

Utah state agenciesfounders’ letter · S-1 · p.5 · L.186

10 agencies

Developers across ten state agencies using Claude to modernize public-service software

Clinical trial reportsfounders’ letter · S-1 · p.4 · L.172

Months → minutes

Time it takes to write clinical trial reports

It is Anthropic's thesis that the exponential in cognition is likely to continue, eventually reaching the level of what we have called "a country of geniuses in a data center"—millions of virtual agents with the ability to revolutionize every area of human endeavor: software, physics, math, biology, medicine, finance, economics, commerce, robotics, military and intelligence capabilities, and much more.

founders’ letter · S-1 · p.4 · L.168

2 of 4 — AI's risks

Parser-described artwork: a profile of a human head with colorful circles overlaid on it, suggesting cognitive processes or data points.
p.5 · founders’ letter · render

There are four broad categories of risk: misuse, loss of control, authoritarian abuse, and economic disruption.

founders’ letter · S-1 · p.6 · L.196

Broad categories of AI riskfounders’ letter · S-1 · p.6 · L.196

4

Misuse, loss of control, authoritarian abuse, economic disruption

Powerful AI widely availableS-1 · p.17 · L.481

Within a few years

Company belief (forward-looking), prospectus summary

a country of geniuses in a data center

the filing’s words, in its own quotation marks · founders’ letter · S-1 · p.4 · L.168

3 of 4 — Anthropic's unique approach

Over $6.6M

TTM revenue per employee · Trailing 12-month revenue per employee as of Sep 30, 2026 · $ millions · founders’ letter · S-1 · p.7 · L.218

Parser-described artwork: a landscape photo of a forest and hills overlaid with a grid and green circular data points.
p.7 · founders’ letter · render
Regretted leavers who returnedfounders’ letter · S-1 · p.7 · L.222

19%

Of employees who left whom the company would have preferred to retain

Directors elected by the LTBTfounders’ letter · S-1 · p.8 · L.234

4 of 7

Co-founders elect the other three

Equity set aside for the public goodfounders’ letter · S-1 · p.8 · L.236

5%

Of the company's equity, before the IPO; +2% per sustained 50% share-price increase, at the board's discretion

Employees who pledged equityfounders’ letter · S-1 · p.7 · L.220

Just over half

Of all employees have pledged at least some equity (donation match)

But we also believe that, for a technology as consequential as AI, entrepreneurial capitalism must be balanced and augmented by mechanisms that protect the common good and share the upside as widely as possible.

founders’ letter · S-1 · p.6 · L.212

4 of 4 — On investor expectations

~$100M

Claude credits given to cyber defenders · “around $100 million”, in the letter’s account of the Mythos Preview delay · $ millions · founders’ letter · S-1 · p.9 · L.264

Parser-described artwork: three circular cutouts of varying sizes revealing a cloudy sky against a dark, textured background.
p.10 · founders’ letter · render

These fall into two categories: business bets that derive from AI's wild exponential trajectory but seem strange or unorthodox in isolation, and decisions that depart from profit maximization in order to appropriately balance the public good.

founders’ letter · S-1 · p.8 · L.248

Co-founder pledgesfounders’ letter · S-1 · p.8 · L.252

9%

Of fully diluted shares as of Jun 30, 2026 (80% of each co-founder's personal equity)

Employee donation-match pledgesfounders’ letter · S-1 · p.8 · L.252

8%

Of fully diluted shares as of Jun 30, 2026

Additional public-benefit set-asidefounders’ letter · S-1 · p.8 · L.236

+2%

For every sustained 50% increase in share price from the prior contribution, at the board's discretion

AI's potential1
AI's risks2
Anthropic's unique approach3
On investor expectations4

Our theory of change has always been what we call a race to the top: by showing that it is possible to succeed commercially while acting in the public interest, we hope to persuade other companies to do the same.

founders’ letter · S-1 · p.8 · L.242

02 — The Exponential

$11.6B
Revenue · Q2 2026 · $11,555M as reported, rounded here · S-1 p.168

We expect our revenue growth rate to decline over time. In addition, because the AI industry is very nascent, we have struggled to accurately forecast our revenue in the past and may struggle to do so in the future.

Risk factors · S-1 p.73

Scroll to draw the quarters, in order.

Q1 2025$376M01 / 06
Fig. 01 · Revenue by quarter · $ millions · Quarterly results · S-1 p.168Every quarter the filing tabulates. No later quarter is tabulated.

The next quarter, as the filing states it

Claude has experienced the fastest enterprise adoption of any technology in history, as demonstrated by revenue growing from $0 at commercial launch in the first quarter of 2023 to $ for the three months ended September 30, 2026.

left blank in the filing · Prospectus summary · S-1 p.16

Customers, by trailing-twelve-month revenue

Customers ≥ $100K TTM revenue
>11,900
Distinct organizations with $100,000 or more of revenue on a TTM basis · as of Sep 30, 2026S-1 p.19
Customers ≥ $1M TTM revenue
>1,900
Distinct organizations with $1 million or more of revenue on a TTM basis · as of Sep 30, 2026S-1 p.19
Customers ≥ $10M TTM revenue
~250
Distinct organizations with $10 million or more of revenue on a TTM basis · as of Sep 30, 2026S-1 p.20
Fig. 02 · Distinct organizations with $100,000 or more of revenue on a TTM basis · MD&A · S-1 p.145Sep 30, 2025 is from the filing’s prose, not its chart. Qualifiers (~, >) as printed.

Reach, and revenue per employee

Fortune 10
Every company
Customers with ≥ $100K TTM revenue included every company in the Fortune 10 · as of Sep 30, 2026S-1 p.205
Fortune 500
>50%
Share of the Fortune 500 that were ≥ $100K TTM customers · as of Sep 30, 2026S-1 p.205
Global 2000
>30%
Share of the Global 2000 that were ≥ $100K TTM customers · as of Sep 30, 2026S-1 p.205
TTM revenue per employee · Sep 30, 2025
$1.4M
$ millions, trailing twelve monthsS-1 p.21
TTM revenue per employee · Sep 30, 2026
>$6.6M
$ millions, trailing twelve months; ~5,300 employees (p206)S-1 p.21

As the filing defines it

Net Revenue Retention
~900%
“approximately 900%” as of Sep 30, 2026. Definition excludes Claude Pro and Max subscribers and compares a three-month current period with the same three months a year earlier (p13–14)S-1 p.21

03 — The Loop

Over90%
Code contributions written by Claude · Share of contributions to Anthropic's codebase in September 2026 · S-1 p.194

As a result, we may not identify a significant acceleration in our rate of development until after it has occurred.

Risk factors · S-1 p.39

Graphic · Labels as printed in the filing’s graphic · Business · S-1 p.194

Inside the company

Code contributions written by Claude
>90%
Share of contributions to Anthropic's codebase in September 2026Business · S-1 p.194
Daily code volume per engineer
>45x
Average daily code contributed per engineer, Q3 2026 vs the FY2024 averageBusiness · S-1 p.194
Success on hard, open-ended engineering problems
~90%
September 2026, up approximately 50 percentage points in nine months; internal benchmarksBusiness · S-1 p.194
Value of internal Claude usage
>250x
Growth at list prices, January 2026 → September 2026Business · S-1 p.195
Internal usage value > personnel expense
April 2026
Month the monthly value of internal usage first exceeded monthly personnel expense, excluding stock-based compensationBusiness · S-1 p.195
Headcount growth
>5x
Dec 31, 2024 → Sep 30, 2026Business · S-1 p.194
Quarterly revenue multiple
86x
Revenue for Q3 2026 vs Q4 2024 (a ratio the filing gives although Q3 2026 revenue itself is blank)Business · S-1 p.194

Each version of a Claude model assists in building the next, helping to compress development timelines and produce compounding improvements in the speed at which we innovate. We refer to this recursive dynamic as "closing the loop," and we believe it is one of the most significant structural advantages in frontier AI and can greatly accelerate the rate at which capabilities advance.

Business · S-1 p.191

The filing’s term

Virtual Collaborator

an AI system that works alongside humans to perform complex, multi-step work at professional levels.

Business · S-1 p.192

Ultimately, millions of Virtual Collaborators can run in parallel, each working independently on unrelated problems or collaborating seamlessly to achieve a shared objective. The result is something without historical precedent: what we refer to as a country of geniuses in a data center, available to every organization and sector of the economy.

Prospectus summary · S-1 p.19

21 case studies · S-1 pp.219–238

p.219 · Accenture
p.220 · AIG
p.221 · Bristol Myers Squibb
p.222 · DoorDash
p.223 · International Rescue Committee
p.224 · Jane Street
p.225 · Legora
p.226 · Lilly
p.227 · Mercado Libre
p.228 · Mozilla
p.229 · PayPal
p.230 · Rakuten
p.231 · Salesforce
p.232 · SAP
p.233 · Shopify
p.234 · Travelers
p.235 · UST
p.236 · Verizon
p.237 · Walmart
p.238 · Shy Bird
p.238 · KBF Design Gallery
select a page to open it

Results, as the case studies state them

Mozilla · Finding the vulnerabilities attackers haven't
271
Two months later, Mozilla shipped fixes for an additional 271 vulnerabilities identified by Claude Mythos Preview.Case study · S-1 p.228
AIG · Reinventing commercial underwriting
>90%
Where deployed, Underwriting by AIG Assist cut submission review time by more than 90% and enabled review of 100% of applicable submissions.Case study · S-1 p.220
Verizon · Rebuilding and defending America's largest network
>33,000
Today, more than 33,000 engineers build with Claude Code.Case study · S-1 p.236

The same filing, on the loop’s risks.

As we increasingly leverage Claude to accelerate progress within Anthropic, our organization may lose visibility across certain processes in a way that may become difficult for any individual or group of humans to understand and control. Because our models write a large and growing share of our code and increasingly perform our research, including our safety research, misaligned behavior in internal use could undermine that work, introduce vulnerabilities into our systems or into the training of future models, or result in models operating outside of our intended controls.

Risk factors · S-1 p.59

Using AI to help build more capable AI could also pose a considerable safety risk, particularly if it sharply accelerates the pace of progress.

Business · S-1 p.195

04 — The Bet

2.08×
Revenue per dollar of compute and infrastructure expense · Q2 2026 · up from 0.40x in Q1 2025 · S-1 p.147

In the first category, Anthropic will often make forward investments (especially in compute) that may seem irresponsibly large at first glance.

founders’ letter · S-1 p.9

Fig. 03 · Revenue per dollar of compute · by quarter · S-1 p.147in-period revenue ÷ compute and infrastructure expense, as printed
Fig. 03b · Compute and infrastructure expense as % of revenue · by quarter · S-1 p.169

We manage compute as a single, fungible pool from multiple strategic partners, and we dynamically direct available compute between workloads based on where the marginal unit is expected to be most productive at any given time.

business · S-1 p.191

our multi-chip compute architecture across AWS Trainium, Google TPUs, NVIDIA GPUs, and AMD GPUs allows us to orchestrate workloads to optimize for price and performance.

business · S-1 p.199

  1. Microsoftat least $44.9 billionat least · Nov 2026 – May 2033 · p.181
  2. Googleat least approximately $114.5 billionat least approximately · Apr 2026 – Jul 2033 · p.181
  3. Amazonat least $110.0 billionat least · May 2026 – Apr 2036 · p.182
  4. SpaceXAIup to $120.5 billionup to · Terms extending into 2029 · p.182
  5. Nscaleup to approximately $44.6 billionup to approximately · Terms extending into 2033 · p.182
  6. Lambdaapproximately $41.3 billionapproximately · Terms extending into 2034 · p.182
  7. Broadcomapproximately $161.2 billion excluding interestapproximately · excluding interest · Five-year equipment leases · p.182
  8. AMDexceed $20.0 billionexceed … if the full initial capacity is procured (not in the contractual table) · Over the term of the agreement (initial capacity in calendar 2027; no fixed dollar commitment) · p.182
  9. Aggregate committed spend, including the portions due before December 31, 2030 and in the next 12 months[BLANK] in the filingno figure printed · p.148
Fig. 04 · Named compute and infrastructure agreements, as stated · unit tick = six months of revenue ($16,286M) · S-1 pp.181–182amounts as stated in the filing; periods differ; no sum is shown
Unit tick · revenue, six months to Jun 30, 2026
$16,286M
Six months ended Jun 30, 2026 · $ millions · vs $1,163M a year earlierS-1 p.33

Compute is procured and committed in advance and the expense is incurred as it is brought online and added to our fungible pool, often in large, discrete steps. The fungible pool can support all workloads dynamically, including research, model development, serving customers, and internal usage of Claude. As a result, there is a lag between when investment is committed and when the full revenue associated with such investment is realized.

management’s discussion and analysis · S-1 p.144

Additionally, we have in the past sought quantities of compute from CSPs in excess of what those partners were willing to provide.

risk factors · S-1 p.41

In the past, we have significantly underestimated our compute requirements, which left us capacity constrained and impacted both product development and deployment.

risk factors · S-1 p.42

Our relationships with our strategic partners—whether with these CSPs, who simultaneously serve as our investors, customers, distribution partners, and competitors, or with Broadcom, which simultaneously serves as a supplier and lessor—create potential conflicts of interest that are difficult to fully anticipate or mitigate.

risk factors · S-1 p.85

But some level of risk is inherent to the extreme upside possibilities offered by the AI exponential. In general, we will aim to be bold but sane, and to win trust with a track record of successful bets.

founders’ letter · S-1 p.9

05 — Red Ink

$80.7B

Net loss · six months to Jun 30, 2026 · $(80,748)M · S-1 p.33

These changes were primarily driven by an increase in the fair value of our outstanding convertible promissory notes resulting from an increase in our enterprise valuation during each of the respective periods.

summary consolidated financial data, footnote 2 · S-1 p.33

Net loss · six months to Jun 30, 2026 · $(80,748)M

Q2 2026 alone · $(59,597)M of the six-month $(72,388)M · Three months ended Jun 30, 2026 · $ millions; the filing ties it to the Series H financing (p170 L4740) · S-1 p.168

Operating loss$(693)MS-1 p.33 · line 976
Change in fair value of convertible notes, derivative liabilities, and SAFEs$(72,388)MS-1 p.33 · line 992
Other income (expense), net$(8,757)MS-1 p.33 · line 1000
Interest income+$1,111MS-1 p.33 · line 984
Provision for income taxes$21MS-1 p.33 · line 1024
Net loss · six months to Jun 30, 2026$(80,748)MS-1 p.33 · line 1032
Fig. 05 · Anatomy of the six-month net loss · $ millions · S-1 p.33non-GAAP terms avoided; labels as printed
Cash, cash equivalents, and short-term investments
$131,651M
As of Jun 30, 2026 · $ millionsS-1 p.35
Cash and cash equivalents
$112.4B
As of Jun 30, 2026 · $ billionsS-1 p.21
Short-term investments
$19.3B
As of Jun 30, 2026 · $ billionsS-1 p.21
Operating cash flow · six months to Jun 30, 2026
$3,929M
Net cash provided by operating activities · $ millionsS-1 p.179
Operating cash flow · six months to Jun 30, 2025
$(1,513)M
Net cash used in operating activities · $ millionsS-1 p.179
Capital raised to date
$141.1B
Aggregate capital raised as of Jun 30, 2026 · $ billions (not cash, not a valuation)S-1 p.21
Accumulated deficit
$132.9B
As of Jun 30, 2026 · $ billions ($(132,902)M on the balance sheet, p35)S-1 p.172
…of which fair-value changes since 2023
$113.2B
Changes in fair value of convertible notes and derivative liabilities driven by increases in enterprise valuation, including $72.4B in the six months to Jun 30, 2026 · $ billionsS-1 p.172
Total stockholders’ deficit
$(127,619)M
As of Jun 30, 2026 · $ millions. Redeemable convertible preferred stock is recorded “outside of permanent equity” (notes, p.369) and converts into common stock at the IPO (p.30)S-1 p.35
Redeemable convertible preferred stock
$157,510M
Carrying value as of Jun 30, 2026 · $ millions; converts in the Preferred Stock Conversion (p30)S-1 p.35

05 — Red Ink

  1. Series A · $2.57 · S-1 p.392
  2. Series B · $11.23 · S-1 p.392
  3. Series C-1 · $11.23 · Mar–Jun 2023 · S-1 p.294
  4. Series D-1 · $30.00 · Jan–May 2024 · S-1 p.415
  5. Series E-1 · $56.09 · Jan–Mar 2025 · S-1 p.415
  6. Series F-1 · $140.97 · Aug–Sep 2025 · S-1 p.415
  7. Series G-1 · $259.14 · Jan–Feb 2026 · S-1 p.415
  8. Series H-1 · $589.01 · May–Jun 2026 · S-1 p.416
Fig. 06 · Price per preferred share, by series, as printed · $ per share · linear scale · S-1 pp.294, 392, 415, 416preferred share counts are not split-adjusted; dates as printed where stated
Series H gross proceeds
$65.2B
Series H-1 and H-2 sold at $589.0095 per share in Q2 2026 (May–June) · $ billions. Final figure from Note 10; Note 16 (as of June 1, 2026) gave $60.8B with additional sales expected in June (p409 L13492)S-1 p.392
Bartz copyright settlement
$1.5B
Settled September 2025; approved July 2026; $300M paid in 2025, $1.2B owed at June 30, 2026 (p405 L13368). Settlement of the Bartz authors' class action · $ billions; recorded as a FY2025 operating expense ($1,500M, p33)S-1 p.252

06 — What They Won’t Do

5%
Initial Public Benefit IssuanceOf fully diluted shares as of Jun 30, 2026, assuming full conversion of the Amazon NotesS-1 · p.23 · L.634

These fall into two categories: business bets that derive from AI's wild exponential trajectory but seem strange or unorthodox in isolation, and decisions that depart from profit maximization in order to appropriately balance the public good.

founders’ letter · S-1 · p.8 · L.248

01 / 10

No advertising on Claude

We have also committed to keeping Claude free of advertising. Advertising funds much of the modern internet, and it would likely be a large source of revenue for us. But people bring some of their most important work and most personal questions to Claude, and we want Claude's only incentive to be giving them a helpful answer.

founders’ letter · S-1 · p.9 · L.266

02 / 10

Mythos Preview not released widely

Claude credits to cyber defenders

~$100M

“around $100 million” · $ millionsfounders’ letter · S-1 · p.9 · L.264

We therefore chose not to make it generally available and offered its capabilities to the public (as Claude Fable 5) only after we had built additional safeguards. The delay meant forgoing substantial revenue and ceding ground to competitors (we also gave away around $100 million in Claude credits to cyber defenders), but we believe it was the right decision, and we expect to face similar decisions again as our models grow more capable.

founders’ letter · S-1 · p.9 · L.264

03 / 10

May advance more slowly

Building AI safely may require Anthropic (and ideally the rest of the industry) to modulate the extremely fast pace at which AI capabilities are compounding, especially when this is driven by AI systems building themselves. While progress will still be very fast in absolute terms—potentially much faster than today—there may be times when we choose to advance more slowly than we could in order to make our systems safer, even if that puts us at a competitive disadvantage.

founders’ letter · S-1 · p.9 · L.268

04 / 10

No image or video generation

Similarly, we have chosen not to develop certain commercially attractive offerings, such as image and video generation models, in order to direct our compute toward our research and safety priorities.

S-1 · p.121 · L.2858

05 / 10

No mass surveillance, no fully autonomous weapons

For example, our usage policy prohibits the use of our models, platform, and applications to develop or operate fully autonomous weapons or to conduct mass surveillance, including mass domestic surveillance, and we have declined to permit such uses based on our Company Purpose, notwithstanding the revenue those or similar uses could generate. Maintaining these limitations has had, and may continue to have, adverse commercial and financial consequences.

S-1 · p.121 · L.2858

06 / 10

Pre-training delayed over a month

For example, we recently chose to delay pre-training a new frontier base model for more than a month to allow time to further enhance our security measures related to preventing model weight exfiltration and protecting our intellectual property.

S-1 · p.62 · L.1605

07 / 10

Public Benefit Fund

Class N shares contributed to the Fund

846,840,960

Initial Public Benefit Issuance = 5% of fully diluted shares as of Jun 30, 2026 (assuming full conversion of the Amazon Notes)S-1 · p.23 · L.634

Prior to our initial public offering, we are setting aside 5% of the company's equity for a fund devoted to the public good. We will set aside an additional 2% for every sustained 50% increase in share price from the prior contribution (starting with our initial public offering price), at our board's discretion.

founders’ letter · S-1 · p.8 · L.236

08 / 10

Founders’ 80% pledges

Co-founder pledges

9%

Of fully diluted shares as of Jun 30, 2026 (employee donation-match pledges: 8%)founders’ letter · S-1 · p.8 · L.252

In addition, our co-founders have each pledged to dedicate 80% of their personal equity in Anthropic to philanthropic causes. These commitments are voluntary, revocable expressions of each co-founder's philanthropic intent and do not constitute legally binding obligations.

S-1 · p.129 · L.3024

09 / 10

Economic-disruption studies

Committed to studies with governments

$200M

$ millionsfounders’ letter · S-1 · p.6 · L.204

We also suspect more radical actions will be needed and have already committed $200 million to partner with governments to run studies on different ways of responding to changes in the economy.

founders’ letter · S-1 · p.6 · L.204

10 / 10

Electricity price pledge

We have publicly committed to covering the cost of electricity price increases to consumers caused by our data centers, including both the cost of electricity itself and the infrastructure necessary to deliver it.

S-1 · p.45 · L.1292

The numbers of giving.

Anthropic Public Benefit Fund

Initial Public Benefit Issuance
5%
Of fully diluted shares as of Jun 30, 2026, assuming full conversion of the Amazon NotesS-1 · p.23 · L.634
Class N shares to the Anthropic Public Benefit Fund
846,840,960
Shares (count)S-1 · p.23 · L.634
Each further issuance
Up to +2%
Of then-outstanding shares each time the share price appreciates 50% above the prior milestone, starting from the IPO price; at the board's discretionS-1 · p.23 · L.634

Pledges and the donation match

Each co-founder's personal equity pledged
80%
To charitable causes; voluntary and revocable (p129)founders’ letter · S-1 · p.7 · L.220
Co-founder pledges as share of the company
9%
Of fully diluted shares as of Jun 30, 2026founders’ letter · S-1 · p.8 · L.252
Employee donation-match pledges
8%
Of fully diluted shares as of Jun 30, 2026founders’ letter · S-1 · p.8 · L.252
Donation-match expense · Q3 2026
$2.0B
Recognized in other operating and administrative expense for matching employee donations made in August 2026 · $ billions (unaudited subsequent event)S-1 · p.411 · L.13555

03 / 10

May advance more slowly

Building AI safely may require Anthropic (and ideally the rest of the industry) to modulate the extremely fast pace at which AI capabilities are compounding, especially when this is driven by AI systems building themselves. While progress will still be very fast in absolute terms—potentially much faster than today—there may be times when we choose to advance more slowly than we could in order to make our systems safer, even if that puts us at a competitive disadvantage.

founders’ letter · S-1 · p.9 · L.268

07 — Who Holds the Pen

50.1%
Class F common stockA fixed 50.1% of total voting power on matters submitted to stockholders generallyS-1 · p.300 · L.8523

Holders of our Class A common stock generally will not be able to affect the outcome of any matter submitted to a vote of our stockholders or the election of our directors, regardless of how they vote, unless a majority of the Continuing Founders cannot agree on how our Class F common stock will be voted (in which case our Class F common stock will generally be voted proportionately with the votes of our Class A common stock cast on the matter).

S-1 · p.313 · L.8823

5 classes or series of common stock, 7 board seats.

4 Class T directors3 other directorsTTTTClass AClass NClass TClass EClass F

Select a class to see what it votes on.

The Long-Term Benefit Trust (one share)

Class T common stock

No general vote; elects and removes the Class T directors, a majority of the board

S-1 · p.26 · L.698

The Founder LLC (one share; seven co-founders, one vote each)

Class F common stock

A fixed 50.1% of total voting power on matters submitted to stockholders generally

S-1 · p.304 · L.8634

Fig. 07 · Share classes and the board they elect · founders’ letter · S-1 · p.8 · L.234; class terms S-1 · p.304

When Class F ends

  1. 01

    Seven Continuing Founders

    S-1 · p.316 · L.8893

  2. 02

    Fewer than three for 90 consecutive days

    S-1 · p.316 · L.8895

  3. 03

    450-day Class F Transition Period

    S-1 · p.316 · L.8895

  4. 04

    Class F and Class E redeemed for no consideration

    S-1 · p.316 · L.8895

The Long-Term Benefit Trust

Trustee and Chair (since June 2023)
Dr. Neil “Buddy” Shah
CEO, Clinton Health Access Initiative; an initial manager of the Anthropic Public Benefit FundS-1 · p.256 · L.7149
Trustee (since April 2025)
Richard Fontaine
CEO, Center for a New American SecurityS-1 · p.257 · L.7160
Trustee (since June 2026)
Dr. Ben S. Bernanke
Distinguished Fellow, Brookings; Federal Reserve Chairman 2006–2014S-1 · p.257 · L.7164
Trustee (since October 2026)
Dr. Christopher Stubbs
Professor of Physics and of Astronomy, Harvard UniversityS-1 · p.257 · L.7168
Trustee seats
5 (one vacant)
“Of the five authorized seats for Trustees, one is currently vacant.”S-1 · p.256 · L.7147
Annual cash compensation per Trustee
$400,000
US dollars per year; the Chair receives an additional $25,000; not tied to company profitsS-1 · p.257 · L.7176
Trustee shareholdings
Prohibited
LTBT policies prohibit a Trustee from directly purchasing or holding any Anthropic securitiesS-1 · p.258 · L.7185

We have a novel capital and corporate governance structure that has no direct public company precedent.

S-1 · p.118 · L.2791

Thus, our directors are not merely permitted, but obligated, to consider our Company Purpose and the interests of other stakeholders.

S-1 · p.122 · L.2868

The Founder LLC (one share; seven co-founders, one vote each)

Class F common stock

The Founder LLC Agreement will provide that the share of Class F common stock will be voted, consented or not consented, in the manner determined by a vote of the Continuing Founders as prescribed by the Founder LLC Agreement, with each Continuing Founder entitled to one vote.

S-1 · p.301 · L.8531

08 — In Their Own Words

88
risk-factor headings in 6 categories · S-1 pp.38–130

For example, we have observed models behaving differently when they detect or infer that they are being evaluated, which can limit our ability to accurately assess real-world risks or failure modes prior to deployment. Such model awareness of our evaluation efforts creates a significant limitation on our ability to assess model safety.

S-1 · Risk Factors · p.57 · L.1520

In their own words

88 risk-factor headings · S-1 pp.38–130 · select a marked heading to read the passage

Page 58 of the S-1, from the risk factors
p.58 · render

Risks Related to our Business and Industry · 1–46 of 88

  • Our success depends in large part on our ability to continue to develop the most advanced and capable AI models, and any failure to do so could adversely affect our business.
  • If we are unable to secure adequate compute or if we fail to direct our compute effectively across our workloads, we may be unable to train models that remain competitive with those of our competitors or adequately serve our customers.
  • If our actual compute requirements differ materially from our projections, we may fail to secure sufficient compute to meet our needs or we may be committed to excess compute that remains underutilized.
  • Constraints on the critical inputs supporting compute could limit our ability to secure adequate compute.
  • We are highly dependent on the continued service of our co-founders, senior management team, and key technical employees, and the loss of any of these individuals may adversely affect our business, financial condition, and results of operations.
  • We may require additional capital to meet our substantial ongoing obligations and support our business growth, and our ability to generate sufficient revenue and secure additional capital on acceptable terms is critical to our continued operations and competitive position.
  • Our company culture has contributed to our success and, if we cannot maintain this culture as we grow, our business could be adversely affected.
  • We face intense competition, and if we are unable to compete effectively, our business, financial condition, and results of operations could be adversely affected.
  • If we fail to attract new customers, retain existing customers, or expand usage of our models, platform, and applications, our business, financial condition, and results of operations could be adversely affected.
  • If the market for our models, platform, and applications develops more slowly than we expect or declines, our growth may slow or stall, and our business, financial condition, and results of operations could be adversely affected.
  • The political environment and our engagement with policymakers and public positions on AI regulation and safety could subject us to heightened scrutiny, political backlash, or disproportionate regulatory action.
  • If we are unable to maintain and enhance our brand and reputation, our business, financial condition, and results of operations may be adversely affected.
  • We derive substantially all of our revenue from a limited number of products and applications, and our failure to maintain and expand our product portfolio could adversely affect our business, financial condition, and results of operations.
  • A substantial portion of our revenue and accounts receivable is driven by a limited number of customers and distribution channels, and the loss of, or a significant reduction in, spend from one or more of our top customers or disruption through those channels could adversely affect our prospects, business, and financial condition.
  • We have incurred significant indebtedness in order to operate and grow our business, including to fund our substantial compute commitments, and our ability to service this indebtedness may be constrained by our cash flows and liquidity.
  • We may not have the ability to raise the funds necessary to repay the principal and accrued interest of our outstanding convertible promissory notes or the A&R Credit Agreement in cash when required by the terms of the notes or the A&R Credit Agreement, as applicable, or at their maturity.
  • Our pricing structures for our models, platform, and applications change from time to time, which could adversely affect our business, financial condition, and results of operations.
  • If we are unable to attract, retain, or motivate highly skilled personnel, we may not be able to grow or operate effectively.
  • We have a limited operating history in an evolving and nascent industry, which makes it difficult to evaluate our future prospects and may increase the risk that we will not be successful.
  • We are at an early stage of our revenue growth, we have experienced rapid growth in recent periods, and our historical growth rates are not indicative of future performance or our ability to sustain or forecast future revenue growth.
  • If we fail to manage our growth effectively, our brand, business, financial condition, and results of operations could be adversely affected.
  • We have a history of net losses and may not generate sufficient revenue to achieve or sustain profitability in the future.
  • Certain of our operating metrics are subject to assumptions and limitations. As a result, these metrics may not provide an accurate indication of our actual performance or future results.
  • Our business is subject to an evolving and uncertain regulatory and political environment, which could increase our costs, restrict our operations, or adversely affect demand for our models, platform, and applications.
  • We are pursuing initiatives to develop our own data center infrastructure, which involves significant capital commitments, operational complexities, and execution risks that could adversely affect our business, financial condition, and results of operations.
  • We may not successfully build, expand, and deploy our sales and marketing organization in a timely manner, or we may not successfully hire, retain, train, and motivate our sales personnel.
  • We have a limited history under our current go-to-market strategy, and our efforts to scale and evolve our customer acquisition, engagement, and retention model may not be successful.
  • Sales to enterprise customers involve complexities and risks that could adversely affect our business, financial condition, and results of operations.
  • If our marketing strategies, such as our free trials and free tiers, fail to generate the benefits we expect, or if their costs exceed those benefits, our ability to grow our revenue could be adversely affected.
  • If we fail to maintain high-quality customer support, our ability to retain and expand customers could suffer, which may adversely affect our prospects, business, financial condition, and results of operations.
  • Our product and investment decisions may not prioritize short-term financial results and may not align with investor expectations or preferences, which could adversely affect the market price of our Class A common stock.
  • Our investments in new products, technologies, and business initiatives may not produce the expected benefits and could adversely affect our business.
  • If we fail to manage our technical operations infrastructure, or experience service outages, interruptions, or delays in the deployment of our models, platform, and applications, our results of operations may be adversely affected.
  • If we are unable to make acquisitions and investments, or successfully integrate them into our business, our business, financial condition, and results of operations could be adversely affected.
  • Unfavorable media coverage regarding our business, our models, platform, and applications, the impacts of AI, and the AI industry more broadly could adversely affect our business, financial condition, and results of operations.
  • We will incur increased costs and demands upon management as a result of complying with the laws and regulations affecting public companies, which may adversely affect our business, financial condition, and results of operations.
  • Our business, financial condition, and results of operations may be adversely affected by war, terrorism, natural disasters, public health crises, political crises, or other unexpected events.
  • If we are unable to ensure that our models, platform, and applications interoperate with a variety of software applications that are developed by others, including our integration partners, we may become less competitive and our business, financial condition, and results of operations may be adversely affected.
  • Many of the risks we face are interconnected and may compound or trigger cascading adverse effects across our business, which could amplify their overall impact.

Risks Related to Privacy, Cybersecurity, and Intellectual Property · 47–58 of 88

  • The application of certain copyright law doctrines and contractual interpretations of open-source licenses to modern AI model training are uncertain.
  • We are subject to complex and evolving laws, regulations, and other legal obligations related to privacy and data protection, and any actual or perceived failure to comply with such obligations could adversely affect our business, financial condition, and results of operations.
  • We process personal and sensitive data for AI safety, and any actual or perceived failure to comply with the legal and regulatory obligations applicable to such processing could adversely affect our business, financial condition, and results of operations.
  • Any security breaches or other incidents, including those that affect our networks, systems, or data, or the networks, systems, or data of our service providers, suppliers, or customers, could adversely affect our business, financial condition, and results of operations.
  • We employ automated data collection across the web and use other methods to collect data for training our AI models, and such activities may subject us to liability or limit our ability to use the collected data, which may adversely affect our business, financial condition, or results of operations.
  • We collect, integrate, and analyze data via agreements from multiple sources, including data provided by our customers and users and data sourced from third parties under contractual arrangements, and such activities may subject us to liability or limit our ability to use the collected data, which may adversely affect our business, financial condition, or results of operations.
  • We have been, are currently defending against, and may in the future be, subject to intellectual property rights claims by third parties, which are extremely costly to defend, may require us to pay significant damages and may limit our ability to use certain technologies, which may adversely affect our business, financial condition, and results of operations.
  • If we fail to adequately protect our intellectual property rights, our competitive position may be impaired, and we may lose valuable assets, generate reduced revenue, and incur costly litigation to protect our rights, which may adversely affect our business, financial condition, and results of operations.
  • We incorporate open-source software in our models, platform, and applications, which may subject us to litigation or other requirements, each of which may adversely affect our business, financial condition, and results of operations.
  • We currently, and may in the future, license technology from third parties for use in our models, platform, and applications, and our inability to maintain those licenses could adversely affect our business, financial condition, and results of operations, and our use of such third-party technology could subject us to litigation.
  • Our development of solutions that replace or reduce reliance on third-party vendors may expose us to litigation and other disputes.
  • We may be subject to claims that we or our employees have misappropriated the intellectual property of other companies, particularly as we hire employees from other AI and technology companies.

Risks Related to Legal and Regulatory Matters · 59–64 of 88

  • We are subject to anti-corruption and anti-bribery laws, anti-money laundering laws, and similar laws.
  • To continue to grow our business, we will need to continue to expand internationally, and our non-U.S. operations and sales to customers outside of the United States expose us to risks inherent in international expansion, sales, and operations.
  • Increasing scrutiny and changing expectations from global regulations, our investors, customers, and employees with respect to ESG may result in additional compliance risk and costs and may impact our reputation and business.
  • If we or our payment processing partners fail to manage the risks related to our payment processing activities, we could be subject to fines, regulatory scrutiny, disruptions to our business, or reputational harm.

Risks Related to Tax and Accounting Matters · 65–71 of 88

  • Changes in tax laws or regulations could be enacted, or existing tax laws or regulations could be applied to us or our customers in a manner that could increase the costs of our models, platform, and applications and adversely affect our business.
  • Our business, financial condition, and results of operations may be adversely affected if we are required to collect sales or other related taxes for subscriptions to our models, platform, and applications in jurisdictions where we have not historically done so.
  • Our ability to use our net operating loss carryforwards and certain other tax attributes may be limited.
  • Our corporate structure and intercompany arrangements are subject to the tax laws of multiple jurisdictions, and we could be required to pay additional taxes, which could adversely affect our business, financial condition, and results of operations.
  • If our judgments or estimates relating to our critical accounting policies prove to be incorrect, our results of operations may be adversely affected.
  • If we fail to maintain an effective system of disclosure controls and internal control over financial reporting, our ability to produce timely and accurate financial statements or comply with applicable regulations may be impaired, which may adversely affect our business, financial condition, and results of operations.
  • We are exposed to fluctuations in interest rates and currency exchange rates, which could negatively affect our business, financial condition, and results of operations.

Risks Related to our Governance Structure · 72–80 of 88

  • Our capital stock structure features certain provisions that are novel or otherwise not common among other corporations with multiple-class structures, and may effectively eliminate the ability of holders of our Class A common stock to influence important decisions, including a change in control.
  • Our capital structure will concentrate voting power with the LTBT and the Founder LLC for the foreseeable future and will result in the LTBT and the Founder LLC having voting power that is not derived from or proportionate to economic interests in our company.
  • Our governance structure and the adoption of our amended and restated certificate of incorporation may negatively affect the decision by certain institutional investors to purchase or hold shares of our Class A common stock.
  • We are considered a “controlled company” within the meaning of the Nasdaq Global Select Market corporate governance requirements, and, as a result, we qualify for, and currently rely on, exemptions from certain corporate governance requirements.
  • We operate as a Delaware public benefit corporation. As a public benefit corporation, we cannot provide any assurance that we will achieve our Company Purpose.
  • Our focus on the long-term best interests of our company as a PBC, and the duty of our directors to balance the pecuniary interests of our stockholders, the best interests of those stakeholders materially affected by our conduct, and our Company Purpose, may conflict with short-, medium-, or long-term financial interests and business performance, which may negatively impact the value of our Class A common stock.
  • As a Delaware public benefit corporation, we may be subject to increased litigation concerning our board of directors’ obligation to balance our stockholders’ pecuniary interests, the best interests of those materially affected by our conduct and our Company Purpose, the occurrence of which may have an adverse impact on our business, financial condition, and results of operations.
  • Our amended and restated bylaws will designate a state or federal court located within the State of Delaware as the exclusive forum for substantially all disputes between us and our stockholders, and also provide that the federal district courts will be the exclusive forum for resolving any complaint asserting a cause of action arising under the Securities Act of 1933, as amended, each of which may limit our stockholders’ ability to choose the judicial forum for disputes with us or our directors, officers, stockholders, or employees.
  • Provisions in our amended and restated certificate of incorporation and amended and restated bylaws could make a merger, tender offer, proxy contest, or other significant corporate transaction difficult, potentially depressing the market price of our Class A common stock.

Risks Related to Ownership of our Class A Common Stock · 81–88 of 88

  • There has been no prior public trading market for our Class A common stock, and an active trading market may not develop or be sustained following this offering.
  • The market price of our Class A common stock may be volatile, and you may lose all or part of your investment.
  • The issuance of additional stock in connection with financings, acquisitions, investments, our equity incentive plans, the conversion of our outstanding convertible promissory notes, the conversion of our outstanding Class N common stock, our public benefit initiatives, or otherwise will dilute the interests of the holders of our Class A common stock.
  • A substantial number of shares of our Class N common stock will be outstanding upon the completion of this offering, which will dilute the voting power of the holders of our Class A common stock upon conversion.
  • Future issuances of our Class A common stock and Class N common stock will not dilute the voting power of our Class F common stock.
  • Future sales of our Class A common stock in the public market could cause the market price of our Class A common stock to decline.
  • Because we do not anticipate paying any cash dividends on our capital stock in the foreseeable future, capital appreciation, if any, will be your sole source of gain.
  • If securities or industry analysts do not publish research or publish inaccurate or unfavorable research about us, our business or our market, or if they change their recommendation regarding our Class A common stock adversely, the market price and trading volume of our Class A common stock could decline.

09 — The Seven

All seven of us still work at the company, which makes Anthropic the only major AI company to have retained its entire founding team.

founders’ letter · S-1 p.7 · L.222

Talent density matters more than talent mass.

S-1 p.198 · L.5589

Regretted leavers who came back
19%
Of employees who left whom the company would have preferred to retainfounders’ letter · S-1 p.7 · L.222
First 100 employees still there
Over three quarters
Of the first 100 employees still work at the companyfounders’ letter · S-1 p.7 · L.222
TTM revenue per employee
Over $6.6M
As of Sep 30, 2026 · $ millionsfounders’ letter · S-1 p.7 · L.218
Employees
~5,300
Approximately 5,300 employees as of Sep 30, 2026S-1 p.206 · L.5814
Regrettable attrition
~1%
Of the employee base, TTM ended Sep 30, 2026S-1 p.198 · L.5591
Pledged under the 50% match
Over 70%
Of employees who joined before May 2024 made a pledge; over 25% pledged the full 50%founders’ letter · S-1 p.7 · L.230
Participated under the 25% match
Nearly 50%
Of post-May 2024 employees have participated; nearly 30% pledged the full 25%founders’ letter · S-1 p.7 · L.230
Co-founder pay
Equal
“Each co-founder is paid equal compensation.”S-1 p.275 · L.7688
Pay target
90th percentile
Of the AI development market (compensation philosophy)S-1 p.275 · L.7682
  1. Dario Amodei
  2. Daniela Amodei
  3. Tom Brown
  4. Jack Clark
  5. Jared Kaplan
  6. Sam McCandlish
  7. Chris Olah

signature block · founders’ letter · S-1 p.10

Our aim is to help humankind chart a safe course through this technological evolution, open new frontiers of discovery, and bring forth an age of miracles.

founders’ letter · S-1 p.10 · L.278

10 — The Parse

421
pages, read once by a machine
parse manifest · 288 seconds

Fable is generally available and runs with our production safeguards, which block or limit performance on certain sensitive topics, including by routing requests to a less capable Claude model.

S-1 p.210

Pages
421
pages in the PDF, each read onceLlamaParse manifest
Parse time
288 s
wall-clock seconds, start to finishLlamaParse manifest
Lines of markdown
13,847
lines of text in the outputdocument.md
Parse tier
Agentic
page-by-page agentic parsingLlamaParse manifest

Every figure and quote from the filing on this page was read from one PDF by LlamaParse, then audited against the parse. Nothing was hand-transcribed. The parse statistics above come from the parser’s manifest and its output file.

Rendered page 33 of the S-1
p.33 · render · 612 × 792 ptpage_33.json ↓

A header that spans columns, parenthetical negatives, em-dash nulls and superscript footnotes, parsed into one table: the table the Red Ink chapter rests on.

  • heading
  • text
  • table

Exhibit · five pages, five kinds of fidelity

Output · parsed items5 items · 5 boxes

Extracted text, as parsed. Boxes are the parser’s own coordinates, none drawn by hand; a table’s box is its layout region and can run wider than the ruled table.

The wall · all 421 pages · the 20 case-study sheets hang above

p.219 · render
p.220 · render
p.221 · render
p.222 · render
p.223 · render
p.224 · render
p.225 · render
p.226 · render
p.227 · render
p.228 · render
p.229 · render
p.230 · render
p.231 · render
p.232 · render
p.233 · render
p.234 · render
p.235 · render
p.236 · render
p.237 · render
p.238 · render

421 pages · select any page to open its render

Built with Claude Fable 5.1 on LlamaParse output. Not affiliated with Anthropic.