Anthropic
AI safety and research company building reliable, interpretable, and steerable AI systems.
Website: https://www.anthropic.com/
Cover Block
PUBLIC
| Field | Value |
|---|---|
| Name | Anthropic |
| Tagline | AI safety and research company building reliable, interpretable, and steerable AI systems |
| Headquarters | San Francisco, United States |
| Founded | 2021 |
| Stage | Series D+ (most recently a Series G in February 2026) |
| Business Model | API and developer platform, with consumer and enterprise Claude subscriptions |
| Industry | Deeptech, frontier AI |
| Technology Type | Large language models, AI safety research |
| Geography | North America, with global API distribution |
| Growth Profile | Venture scale |
| Founding Team | Co-founders Dario Amodei and Daniela Amodei |
| Funding Label | $100M+ (cumulative disclosed rounds exceed $43B based on Series C, F, and G alone) |
Links
PUBLIC
- Website: https://www.anthropic.com/
- LinkedIn: https://www.linkedin.com/company/anthropicresearch
- Product (Claude): https://anthropic.com/product
- Pricing: https://www.anthropic.com/pricing
- Research: https://www.anthropic.com/research
- Careers (Lever): https://jobs.lever.co/Anthropic
Executive Summary
PUBLIC
Anthropic is a San Francisco frontier AI lab whose Claude family of models has emerged as the principal commercial alternative to OpenAI in both the developer API market and large enterprise deployments [Wikipedia][Crunchbase]. The company was founded in 2021 by Dario and Daniela Amodei, who departed OpenAI in late 2020 alongside several senior researchers over disagreements about commercialization pace and safety prioritization [Observer, 2024]. Its core differentiation is a published constitutional approach to model alignment paired with a Public Benefit Corporation structure, which the company positions as a credible safety story for regulated buyers and government customers [Anthropic]. The product surface centers on Claude (most recently Claude Opus 4.5, released November 24, 2025, which the company describes as its best model for coding, agents, and computer use) delivered through a free-to-enterprise tiered subscription and a developer API [Anthropic, 2025]. Capital formation has accelerated sharply: a $13B Series F at a $183B valuation in September 2025 was followed by a $30B Series G in February 2026 led by GIC and Coatue at a $380B post-money valuation, the second-largest venture round on record [TechCrunch, 2025][Crunchbase, 2026][Orrick, 2026]. CEO Dario Amodei has publicly confirmed annualized revenue exceeding $19B [Yahoo Finance, 2026], a figure that, if sustained, would make Anthropic one of the fastest revenue ramps in software history. The next 12 to 18 months will test three things: whether enterprise gross margins hold as the company competes for compute against OpenAI and xAI, whether the Pentagon and national security pipeline materializes into a durable government franchise [Bloomberg, 2026], and whether the Amodei-led safety posture survives the commercial pressure that comes with a $380B price tag.
Data Accuracy: GREEN -- Confirmed by Crunchbase, TechCrunch, Orrick, Yahoo Finance, and the company's own press releases.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Series G (February 2026) |
| Business Model | API and developer platform, plus tiered Claude subscriptions (Free, Pro, Max, Team, Enterprise) |
| Industry / Vertical | Frontier AI, deeptech |
| Technology Type | Large language models, alignment and interpretability research |
| Geography | North America headquartered, global distribution |
| Growth Profile | Venture scale, hyper-growth |
| Founding Team | Two co-founders, sibling pair, ex-OpenAI |
| Funding | $30B Series G at $380B post-money [Crunchbase, 2026] |
Company Overview
PUBLIC
Anthropic was incorporated in 2021 as a Delaware Public Benefit Corporation with a stated purpose of "the responsible development and maintenance of advanced AI for the long-term benefit of humanity" [Anthropic]. The founding cohort included Dario Amodei (formerly VP of Research at OpenAI) and his sister Daniela Amodei (formerly VP of Safety and Policy at OpenAI), along with several colleagues who left OpenAI in late 2020 over what has been reported as a disagreement about the pace of commercialization and the prioritization of safety research [Observer, 2024][Wikipedia]. The company is headquartered in San Francisco.
The milestone arc is unusually steep. After an early seed and Series A and B largely outside the public record, Anthropic raised a $450M Series C led by Spark Capital in May 2023 [TechCrunch, 2023], followed by strategic investments and commercial commitments from Google and Amazon, the latter of which has been widely reported as Anthropic's primary cloud partner. In September 2025 the company closed a $13B Series F led by Iconiq at a $183B valuation, with Fidelity Management & Research Co. participating [TechCrunch, 2025][Crunchbase]. Five months later, in February 2026, GIC and Coatue led a $30B Series G at a $380B post-money valuation, which Crunchbase identified as the second-largest venture funding deal of all time [Crunchbase, 2026][Orrick, 2026].
Product milestones have moved at a similar tempo: Claude 2.1 introduced a 200K token context window and beta tool use [Anthropic]; Claude Sonnet 4 and Claude Opus 4 launched on May 22, 2025 [Wikipedia]; and Claude Opus 4.5 launched on November 24, 2025 with the company positioning it as a leading model for coding, agentic workflows, and computer use [Anthropic, 2025]. In parallel, Anthropic has publicly disclosed deployments inside U.S. classified networks, the National Laboratories, and the Department of War, framing itself as the first frontier lab to operate at that level of government integration [Anthropic].
Data Accuracy: GREEN -- Confirmed by Anthropic press releases, Wikipedia, Crunchbase, TechCrunch, and Orrick.
Product and Technology
MIXED
The commercial surface area is narrower than the research footprint suggests: substantially all revenue flows from Claude, delivered either through the Anthropic API or through tiered subscriptions branded Free, Pro, Max, Team, and Enterprise [Anthropic]. Claude is positioned by the company as "Anthropic's AI, built for problem solvers," oriented toward complex analysis, code generation, and structured reasoning [Anthropic]. The most recent model, Claude Opus 4.5 (released November 24, 2025), is described by Anthropic as its best model for coding, agents, and computer use [Anthropic, 2025]. Claude Code, an agentic coding product, is available on Team and Enterprise plans with premium seats and admin controls [Anthropic].
The technical posture is distinct from peers in two visible ways. First, Anthropic has published a "Constitution" describing the principles used to train Claude's behavior, which functions both as an alignment artifact and as a marketing wedge for risk-sensitive buyers [Anthropic]. Second, the company invests heavily in interpretability research and an Economic Index that quantifies how Claude is being used across occupations and geographies, with reports published as recently as March 2026 covering Australia and learning curves [Anthropic]. A Claude for Life Sciences program and participation in the multi-billion-dollar federal Genesis Mission extend the surface into vertical scientific computing [Anthropic].
Infrastructure dependencies are partially public. Anthropic's primary compute relationships are widely reported to include Amazon Web Services and Google Cloud, both of which are also strategic investors, and the company has publicly engaged with Nvidia as a hardware supplier even as CEO Dario Amodei voiced criticism of Nvidia at Davos in January 2026 [TechCrunch, 2026]. Open roles for an Engineering Manager, LLM Scaling [Lever.co, 2026] suggest continued internal investment in training infrastructure rather than full reliance on third-party stacks (inferred from job postings).
Data Accuracy: GREEN -- Confirmed by Anthropic.com, Wikipedia, and TechCrunch.
Market Research and Opportunity
PUBLIC
The frontier model market has consolidated faster than any prior software category, and Anthropic now occupies the clear number-two slot behind OpenAI by reported revenue and valuation. Independent third-party TAM estimates for generative AI vary widely and none in the structured record here is from a named institutional report, so the more reliable framing is to anchor on disclosed revenue and valuation comparables among the small set of frontier labs.
Demand drivers visible in the cited record are concrete. On the consumer side, DemandSage reports roughly 18.9 million monthly active users for Claude as of 2026 [DemandSage, 2026]. On the enterprise side, the same source counts more than 300,000 business customers and agencies using the Claude Enterprise API [DemandSage, 2026]. CEO Dario Amodei has confirmed annualized revenue exceeding $19B [Yahoo Finance, 2026], with The Information also citing the $19B annualized figure [The Information]. Government demand is a distinct tailwind: Anthropic states it was the first frontier lab to deploy inside U.S. classified networks and at the National Laboratories, with extensive Department of War usage for intelligence analysis, modeling, and cyber operations [Anthropic].
| Metric | Value | Source |
|---|---|---|
| Annualized revenue | $19B+ | [Yahoo Finance, 2026] |
| Post-money valuation (Series G) | $380B | [Crunchbase, 2026] |
| Post-money valuation (Series F) | $183B | [Crunchbase] |
| Consumer MAU | 18.9M | [DemandSage, 2026] |
| Enterprise API customers | 300,000+ | [DemandSage, 2026] |
Analyst takeaway: at roughly 20x annualized revenue against the Series G valuation, Anthropic is being priced at a premium to mature enterprise software but at a discount to its closest comparable, OpenAI, on the most recent reported multiples. The implied bet is that the $19B revenue number compounds rather than plateaus.
Adjacent and substitute markets matter. Open-weight models from Meta and Mistral AI compress price-performance at the low end, hyperscaler-native models from Google DeepMind bundle into existing cloud contracts, and xAI competes for the same enterprise mindshare with a different distribution wedge through X. Regulatory forces cut both ways: the EU AI Act and emerging U.S. federal guidance reward documented safety practices (an Anthropic strength), while export controls and energy permitting for data centers constrain the entire frontier cohort symmetrically.
Data Accuracy: YELLOW -- Revenue and valuation figures confirmed across multiple outlets; user and customer counts come from a single secondary source (DemandSage).
Competitive Landscape
MIXED
Anthropic competes as the principal safety-forward alternative to OpenAI in a market where five labs control essentially all of frontier model spend.
| Company | Positioning | Stage / Funding | Notable Differentiator | Source |
|---|---|---|---|---|
| Anthropic | Safety-forward frontier lab; Claude API and enterprise | Series G, $380B post-money | Constitutional AI, PBC structure, deep federal deployments | [Crunchbase, 2026][Anthropic] |
| OpenAI | Frontier consumer and enterprise AI; ChatGPT and API | Late stage, multi-hundred-billion valuation reported | Consumer brand leadership and Microsoft distribution | [TechCrunch, 2026] |
| Google DeepMind | Frontier lab inside Alphabet; Gemini family | Public parent | Native bundling into Google Cloud, Workspace, Search | [PUBLIC] |
| xAI | Frontier lab tied to X distribution; Grok | Late stage venture | Distribution through X and Tesla data assets | [PUBLIC] |
| Mistral AI | Open-weight European frontier lab | Late stage venture | Open weights, EU sovereignty positioning | [PUBLIC] |
| Meta | Open-weight Llama family | Public | Free distribution at scale, ecosystem effects | [PUBLIC] |
The segment map breaks into three layers. At the frontier closed-model layer, the contest is essentially Anthropic, OpenAI, Google DeepMind, and increasingly xAI, all racing on model capability, agentic reliability, and inference economics. At the open-weight layer, Meta's Llama and Mistral apply downward price pressure on commodity inference and serve as the default for cost-sensitive buyers and sovereign deployments. At the application layer, every frontier lab is also a partial competitor to its own customers, a tension that affects how aggressively Anthropic builds first-party products like Claude Code.
Anthropic's defensible edges today are four. First, an enterprise reputation for safety and predictability that is reinforced by published constitutional and interpretability work [Anthropic]. Second, an unusual government franchise: first-mover deployments inside classified networks, National Labs, and the Department of War give Anthropic a credentialed channel that is hard to replicate quickly [Anthropic][Bloomberg, 2026]. Third, capital depth: $30B raised in a single February 2026 round at $380B [Crunchbase, 2026] funds multi-year compute commitments. Fourth, a coding-and-agents reputation around Claude Opus 4.5 and Claude Code that has gained traction with developer audiences [Anthropic, 2025]. Edges one and four are perishable, since competitor models leapfrog on roughly six-month cycles. Edges two and three are stickier.
The exposures are equally specific. OpenAI retains a consumer brand and Microsoft distribution channel that Anthropic does not own. Google DeepMind can bundle Gemini into Workspace and Search at zero marginal acquisition cost, a channel Anthropic cannot match. Meta's Llama and Mistral compress the floor for commodity inference, which over time may force Anthropic's API pricing down even as enterprise gross margin assumptions hold up. Public friction with OpenAI over Pentagon messaging in March 2026 [TechCrunch, 2026] and CEO criticism of Nvidia at Davos [TechCrunch, 2026] suggest a willingness to pick visible fights, which can sharpen positioning but also concentrates execution risk on the CEO.
The most plausible 18-month scenario: Anthropic wins if Claude becomes the default model for regulated enterprise and federal workloads, locking in multi-year contracts that compound into a high-margin franchise; Anthropic loses ground if OpenAI's distribution through Microsoft and Google's bundling through Workspace turn the consumer and SMB layer into a two-horse race that excludes Anthropic from the top of the funnel.
Data Accuracy: GREEN -- Confirmed by Crunchbase, TechCrunch, Bloomberg, and Anthropic press releases.
Opportunity
PUBLIC
The size of the prize, if Anthropic executes, is to become one of the two or three default infrastructure providers for the next generation of enterprise and government software.
The headline opportunity. The single largest plausible outcome is that Anthropic settles into a durable role as the safety-credentialed frontier model provider for regulated industries, federal and allied governments, and developer platforms that prefer not to build on OpenAI. The cited evidence makes that outcome reachable rather than aspirational: $19B+ in annualized revenue confirmed by the CEO [Yahoo Finance, 2026], 300,000+ enterprise API customers reported [DemandSage, 2026], first-mover federal deployments across classified networks and the Department of War [Anthropic], and a $30B war chest from the February 2026 Series G [Crunchbase, 2026]. Few companies in software history have arrived at that combination of revenue, customer breadth, government access, and capital simultaneously.
Growth scenarios.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Federal default | Anthropic becomes the standard frontier model inside U.S. and allied national security agencies | Multi-year DoW contracts and Genesis Mission expansion | First frontier lab deployed in classified networks and National Labs [Anthropic] |
| Enterprise coding standard | Claude Code and Claude Opus 4.5 become the default agentic coding stack inside the Fortune 500 | Continued model leadership on coding benchmarks, premium seat expansion on Team and Enterprise plans | Opus 4.5 positioned as best for coding and agents [Anthropic, 2025]; 300,000+ enterprise API customers reported [DemandSage, 2026] |
| Regulated industry API | Claude becomes the default API for healthcare, financial services, and life sciences | Claude for Life Sciences and Genesis Mission partnerships scale into commercial pharma and biotech [Anthropic] | PBC structure, published constitution, and interpretability research lower procurement friction in regulated verticals [Anthropic] |
What compounding looks like. The flywheel sits at the intersection of capital, compute, and credentialed distribution. Each major round (Series F at $183B, Series G at $380B [Crunchbase, 2026]) funds a larger training run, which lifts model capability, which lifts enterprise and federal contract value, which justifies the next round. Layered on top is a data and feedback loop: 18.9M consumer MAU [DemandSage, 2026] generate usage signal that improves Claude, while 300,000+ enterprise customers [DemandSage, 2026] generate the high-value workflow data that matters most for agentic products like Claude Code. The federal franchise compounds differently: once a model is accredited inside classified networks, switching costs are measured in years, not quarters [Anthropic].
The size of the win. A credible comparable is OpenAI itself, which has been reported at valuations several multiples of Anthropic's $380B post-money [Crunchbase, 2026]. If the federal default and enterprise coding scenarios both play out and annualized revenue compounds from the cited $19B+ base [Yahoo Finance, 2026] at the trajectory implied by the past two funding rounds, Anthropic could plausibly reach a valuation band associated with the largest software platforms of the prior generation (scenario, not a forecast). Even a more conservative path, in which Anthropic holds roughly its current share of frontier API spend and grows with the category, supports a multi-hundred-billion outcome consistent with where the Series G priced the business [Crunchbase, 2026].
Data Accuracy: YELLOW -- Revenue and valuation anchors are GREEN; scenario constructions are analyst inferences from the cited record.
Sources
PUBLIC
[Anthropic] Home | https://www.anthropic.com/
[Anthropic] Company | https://www.anthropic.com/company
[Anthropic] Anthropic Raises $450 Million in Series C Funding | https://www.anthropic.com/news/anthropic-series-c
[Anthropic] The AI for Problem Solvers, Claude product page | https://anthropic.com/product
[Anthropic] Plans and Pricing | https://www.anthropic.com/pricing
[Anthropic] Introducing Claude 2.1 | https://anthropic.com/news/claude-2-1
[Anthropic] Claude's Constitution | https://www.anthropic.com/constitution
[Anthropic] Statement from Dario Amodei on Department of War discussions | https://www.anthropic.com/news/statement-department-of-war
[Anthropic] Research index | https://www.anthropic.com/research
[Anthropic, 2025] Claude Opus 4.5 launch coverage | https://www.anthropic.com/
[Wikipedia] Anthropic | https://en.wikipedia.org/wiki/Anthropic
[Crunchbase] Anthropic profile | https://www.crunchbase.com/organization/anthropic
[Crunchbase] Anthropic Nearly Triples Valuation To $183B With Massive New Funding | https://news.crunchbase.com/venture/generative-ai-anthropic-funding-iconiq/
[Crunchbase, 2026] Anthropic Raises $30B At $380B Valuation In Second-Largest Venture Funding Deal Of All Time | https://news.crunchbase.com/ai/anthropic-raises-30b-second-largest-deal-all-time/
[Crunchbase] Report: Anthropic Raising $5B At A $170B Valuation | https://news.crunchbase.com/ai/unicorn-anthropic-fund-valuation-raise-iconiq/
[Crunchbase] Anthropic Reportedly In Talks To Raise $750M At $18B-Plus Valuation | https://news.crunchbase.com/ai/unicorn-anthropic-funding-menlo-goog-amzn/
[TechCrunch, 2023] Anthropic Series C coverage | https://techcrunch.com/
[TechCrunch, 2025] Anthropic CEO Dario Amodei is trying to duck a deposition in an OpenAI copyright lawsuit | https://techcrunch.com/2025/01/31/anthropic-ceo-dario-amodei-is-trying-to-duck-a-deposition-in-an-openai-copyright-lawsuit/
[TechCrunch, 2026] Anthropic CEO Dario Amodei calls OpenAI's messaging around military deal 'straight up lies' | https://techcrunch.com/2026/03/04/anthropic-ceo-dario-amodei-calls-openais-messaging-around-military-deal-straight-up-lies-report-says/
[TechCrunch, 2026] Anthropic CEO Dario Amodei could still be trying to make a deal with Pentagon | https://techcrunch.com/2026/03/05/anthropic-ceo-dario-amodei-could-still-be-trying-to-make-a-deal-with-pentagon/
[TechCrunch, 2026] Anthropic's CEO stuns Davos with Nvidia criticism | https://techcrunch.com/2026/01/20/anthropics-ceo-stuns-davos-with-nvidia-criticism/
[Bloomberg, 2026] Feud with Pentagon over AI safety and military use restrictions | https://www.bloomberg.com/
[Yahoo Finance, 2026] CEO confirmed annualized revenue exceeding $19B | https://finance.yahoo.com/
[The Information] Annualized Revenue $19B reporting | https://www.theinformation.com/
[DemandSage, 2026] Claude user and customer statistics | https://www.demandsage.com/
[Observer, 2024] Why Dario and Daniela Amodei left OpenAI | https://observer.com/
[Orrick, 2026] Series G round coverage | https://www.orrick.com/
[Contrary Research] Anthropic Business Breakdown | https://research.contrary.com/company/anthropic
[Lever.co, 2026] Anthropic open roles | https://jobs.lever.co/Anthropic
Articles about Anthropic
- Anthropic Is Putting Claude Inside the Pentagon's Classified Networks — The Amodei siblings' safety lab now runs at $19B in annualized revenue while picking fights with OpenAI and Nvidia over how AI should be sold to governments.