What is Anthropic?
AI safety company behind the Claude family of frontier models, sold via API and the Claude apps.
- Category
- Frontier AI lab
- Headquarters
- San Francisco, CA
- Founded
- 2021
- Employees
- ~5,000
- Total funding
- ~$125B+
- Valuation
- $965B (Series H, May 2026)
What is Anthropic?
Anthropic is a San Francisco-based AI safety company and frontier research lab, founded in 2021 by former OpenAI leaders, that builds the Claude family of large language models. It sells Claude to developers and enterprises through an API and to individuals and teams through the Claude apps, positioning itself as the safety-focused alternative among the handful of labs at the AI frontier.
Claude ships as a tiered model family — Haiku (fast and cheap), Sonnet (balanced), and Opus (most capable, currently Opus 4.8) — and powers chat, the Claude Code coding agent, and a desktop app aimed at non-technical users. Most usage runs through the API, where Claude is accessed directly or via Amazon Bedrock and Google Vertex AI, so Anthropic functions as core infrastructure for thousands of other software products.
Growth has been extraordinary. Anthropic's annualized run-rate revenue crossed roughly $47 billion in mid-May 2026 (and the company has told investors it expects to pass $50 billion by the end of June), up from about $9-10 billion at the end of 2025 — one of the fastest revenue ramps ever recorded. It serves more than 300,000 business customers, with enterprise accounting for roughly 80% of revenue, and over 1,000 customers each spending more than $1 million a year (a figure that doubled from 500 in just two months in early 2026).
The company is organized as a public benefit corporation and is widely regarded as OpenAI's closest competitor at the frontier. On May 28, 2026 it raised a $65 billion Series H at a $965 billion post-money valuation, and on June 1, 2026 it confidentially filed an IPO prospectus (Form S-1) with the SEC — the clearest signal yet that it is preparing to go public.
What does Anthropic offer?
Anthropic's products all center on the Claude model family — sold as an API for builders and as apps for individuals, teams, and enterprises.
- Claude API· Developer platform
- Claude (chat apps)· Consumer/Team
- Claude Code· Coding agent
- Claude Opus 4.8· Model
- Claude Sonnet 4.6· Model
- Claude Haiku 4.5· Model
- Claude Enterprise· Enterprise
- Claude on AWS Bedrock / Google Vertex AI· Distribution
- Model Context Protocol (MCP)· Open standard
How does Anthropic make money?
Anthropic makes most of its money selling access to Claude by the token through its API, with the rest coming from per-seat subscriptions to the Claude apps. The large majority of revenue is enterprise — roughly 80% — as companies wire Claude into their own products and internal workflows, while consumer and team subscriptions are a smaller slice.
API pricing is usage-based, charged per million tokens (MTok). As of June 2026, Claude Opus 4.8 runs about $5 per million input tokens and $25 per million output; Sonnet 4.6 about $3 / $15; and Haiku 4.5 about $1 / $5, with batch processing halving costs and prompt caching cutting cached input by up to 90%. That token model means revenue scales directly with how much customers use Claude — heavy agentic and coding workloads burn far more tokens than simple chat.
Subscriptions layer on top. Claude Pro is $20/month (about $17 billed annually), Claude Max is $100 (5x) or $200/month (20x) for power users, and Team seats run roughly $25-30 per user (Premium seats with Claude Code at ~$100), with Enterprise priced custom — a seat fee plus usage at API rates. Coding is a standout: Claude Code alone reached an estimated ~$2.5 billion run rate by early 2026.
Growth is driven by enterprise adoption — more than 1,000 customers now spend over $1 million a year, and 8 of the Fortune 10 are customers — and by coding agents in particular. The flip side is cost: frontier models require enormous compute, so Anthropic's economics hinge on usage outrunning the massive spend on training and inference.
Who leads Anthropic?
Anthropic was founded by a group of former OpenAI researchers led by the sibling pair Dario and Daniela Amodei. Dario Amodei is CEO and Daniela Amodei is President, supported by co-founders and a deep executive bench across research, product, finance, and security.
- Dario AmodeiCo-founder & CEOSince 2021Former VP of Research at OpenAI; sets Anthropic's research direction and is a leading public voice on AI safety and scaling.
- Daniela AmodeiCo-founder & PresidentSince 2021Former VP at OpenAI; runs operations, people, finance, and the company's day-to-day business.
- Jared KaplanCo-founder & Chief Science OfficerSince 2021Theoretical physicist whose work on neural scaling laws underpins Anthropic's model strategy.
- Tom BrownCo-founder & Chief Compute OfficerSince 2021Led GPT-3 engineering at OpenAI; now owns Anthropic's compute and training infrastructure.
- Sam McCandlishCo-founder & Chief ArchitectSince 2021Former research lead; moved from CTO to chief architect as Anthropic scaled its technical org.
- Jack ClarkCo-founder & Head of PolicySince 2021Leads policy and the Anthropic-affiliated AI think tank; key public voice on AI governance.
- Krishna RaoChief Financial OfficerSince 2025Ex-Airbnb finance leader; Anthropic's first CFO, steering finances toward a potential IPO.
- Mike KriegerCo-lead, Anthropic Labs (former Chief Product Officer)Since 2024Instagram co-founder; was CPO, then moved in early 2026 to co-lead the new Anthropic Labs experimental-products group.
- Vitaly GudanetsChief Information Security OfficerSince 2025Former Netflix CISO; secures Anthropic's model IP and sensitive enterprise customer data.
How do you contact Anthropic's leadership?
Anthropic uses a first-name email format on anthropic.com (e.g. first@anthropic.com); a first-initial-plus-last variant (jsmith@anthropic.com) is also reported. The individual addresses below follow that verified pattern and are inferred, not personally published — for verified routes, use department inboxes like press@anthropic.com, sales@anthropic.com, or support@anthropic.com.
first@anthropic.comHow much funding has Anthropic raised?
Anthropic has raised well over $125 billion in equity across eight named rounds (Series A through H) plus large strategic investments from Amazon and Google. Its most recent round, the Series H in May 2026, valued the company at $965 billion post-money — making it one of the most valuable private companies ever — and on June 1, 2026 it confidentially filed for an IPO.
The early arc: a $124M Series A in 2021 (led by Skype co-founder Jaan Tallinn), a $580M Series B in April 2022 (led by FTX/Alameda, whose stake was later sold off after FTX collapsed), and a ~$450M Series C in May 2023 led by Spark Capital at a ~$4.1B valuation. Alongside these, Google committed about $2 billion and Amazon committed up to $4 billion as strategic cloud-and-capital partners through 2023-2024.
The rounds then grew enormous. A Series D (February 2024, led by Menlo Ventures) raised over $750M near an ~$18.1B valuation; a $3.5B Series E in March 2025, led by Lightspeed, lifted Anthropic to $61.5B; and a $13B Series F in September 2025, led by ICONIQ with Fidelity and Lightspeed, took it to $183B.
Then valuation roughly doubled twice in months. A $30B Series G in February 2026 (co-led by GIC and Coatue, with Microsoft and Nvidia participating, plus a ~$5-6B employee tender at ~$350B) set a $380B post-money valuation — at the time the second-largest private tech raise ever. A $65B Series H announced May 28, 2026 then reached $965B, led by Altimeter, Dragoneer, Greenoaks, and Sequoia, and including ~$15B in previously announced hyperscaler commitments (Amazon +$5B, Google +$10B). On June 1, 2026, Anthropic confidentially filed its S-1 with the SEC.
How did Anthropic get here?
Anthropic's path: a 2021 spinout of OpenAI researchers, the 2023 launch of Claude, a string of ever-larger megarounds, and a 2026 confidential IPO filing at a near-trillion-dollar valuation.
- 2021FoundedDario and Daniela Amodei and a group of former OpenAI researchers found Anthropic in San Francisco as an AI safety company and public benefit corporation.
- Mar 2023Claude launchesAnthropic releases its first public Claude models and API; Google soon commits ~$2B as a strategic partner.
- Sep 2023Amazon partnershipAmazon agrees to invest up to $4B and becomes a primary cloud partner, with Claude offered on AWS Bedrock.
- Mar 2025$61.5B valuationA $3.5B Series E led by Lightspeed values Anthropic at $61.5B as enterprise demand accelerates.
- Feb 2026$380B Series GA $30B round — then the second-largest private tech raise ever — values Anthropic at $380B, with Microsoft and Nvidia joining.
- May–Jun 2026$965B & confidential IPO filingA $65B Series H values Anthropic at $965B; on June 1 it confidentially files an S-1 with the SEC.
Who are Anthropic's competitors?
Anthropic competes with the other frontier AI labs and model providers — a small, capital-intensive field where the rivalry is over model capability, developer adoption, and enterprise trust.
- OpenAIThe closest rival and consumer leader with ChatGPT; Anthropic counters with a safety-first lean and outsized strength in coding and agentic workloads.
- Google DeepMindGemini models backed by Google's compute and distribution; competes across consumer, cloud (Vertex AI), and research.
- Meta AIOpen-weight Llama models that undercut closed APIs on cost and self-hosting, pressuring per-token pricing.
- xAIElon Musk's lab building Grok, integrated with X and pursuing the frontier with large compute buildouts.
- Mistral AIEuropean challenger with efficient open and commercial models, strong in EU/sovereign-AI deals.
- CohereEnterprise-focused model provider emphasizing private deployment, RAG, and regulated industries.
Anthropic — frequently asked questions
