What Anthropic PBC does

Anthropic PBC is a San Francisco-based AI safety and research company founded in 2021 that develops the Claude family of large language models. The company serves enterprises, developers, government agencies, and consumers through a portfolio spanning Claude (the flagship AI assistant), Claude Code (an AI coding tool with live iOS testing), Claude Cowork (autonomous multi-step task execution across web and mobile), Claude Design, Claude Science (biopharma research workbench), and Claude Security. The underlying technology rests on Anthropic's proprietary Constitutional AI and RLHF training methodologies, supported by a Responsible Scaling Policy with formal AI Safety Level Standards and a co-developed Cyber Jailbreak Severity Framework. The model family is tiered across Mythos (most advanced/restricted), Fable (flagship), Opus (high-performance, cost-efficient), Sonnet (mid-tier, agentic), and Haiku (lightweight).

Anthropic monetizes through a hybrid revenue model combining usage-based API token pricing ($5-25 per million tokens for Opus 5, $10-50 per million tokens for Fable 5), tiered consumer and team subscriptions (Pro at $20/month, Max at $100-200/month, Team at $25/seat/month with regional pricing in India), enterprise licensing for large organizations, and compute infrastructure partnerships. Go-to-market is multi-motion: API-first developer distribution, product-led growth via Claude.ai and Claude Code, enterprise field sales supported by system integrators (McKinsey, Accenture, LTM, TCS, Infosys, Globant, Fujitsu), and cloud marketplace channels on AWS, Google Cloud Vertex AI, Microsoft Azure AI Foundry, and Hugging Face. Customer segments include enterprise organizations (40% agent orchestration platform adoption), software developers, financial institutions (JPMorgan Chase, FIS), insurance (Brown & Brown, Allianz), government agencies (CISA, Government of Alberta), and life sciences research labs.

The company is privately held with Alphabet/Google as the largest backer at a $124.3 billion stake carrying value as of June 2026, AMD as a strategic $5 billion investor, and Apollo/Blackstone as arrangers of a $35 billion debt facility for AI infrastructure. Anthropic confidentially filed an S-1 with the SEC on June 1, 2026, targeting a fall IPO at a valuation approaching $1.2 trillion based on secondary-market transactions.

Name

Anthropic PBC

Legal name

Anthropic PBC

Website

https://anthropic.com

Company type

Private

Founded year

2021

Operating status

Operating

Short description

Anthropic PBC is a San Francisco-based AI safety company that develops the Claude family of large language models and agentic tools, serving enterprises, developers, governments, and consumers through APIs, subscriptions, and multi-cloud distribution.

Ownership category

akta.pro rank

Product category

Foundation Models and AI Platform

NAICS

Software Publishers (5132)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ)

akta.pro secondary industries

Responsible AI, Security & Privacy Platforms (Safety, Guardrails, PII) (HDAEANAG), Regulatory Readiness & Audit Automation (e.g., EU AI Act, NIST AI RMF, ISO/IEC 42001) (HDAAAMAE), AI Governance, Risk & Compliance (GRC) Platforms (HDAAAMAA), AI & Automation in Customer Service (Chatbots, Agent Assist) (EDABACAO)

Keywords

• Large language models
• AI safety research
• Foundation models
• AI coding assistant
• Enterprise AI platform

Headquarters

HQ city

San Francisco

HQ country

United States

HQ region

North America

Offices5 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Infrastructure, Personnel, Marketing or Sales, Operations

Revenue model

1. API/Token-based pricing: Revenue from API usage based on token consumption. Opus 5 priced at $5 per million input tokens and $25 per million output tokens. Higher tiers like Fable 5 priced at $10 per million input tokens and $50 per million output tokens under usage-based model.
2. Subscription Plans: Recurring subscription revenue from Claude Pro ($20/month), Team ($25/seat/month), Max ($100-200/month), and other tiered plans. Run-rate revenue crossed $47 billion.
3. Enterprise Licensing: Enterprise agreements with large organizations for dedicated capacity, custom deployments, and premium support. Includes partnerships with financial institutions, government agencies, and global enterprises.
4. Custom Chip Development: Revenue from custom AI chip development partnerships including $35 billion financing arrangement with Apollo and Blackstone for Broadcom and Anthropic AI infrastructure expansion.

Pricing tiers

Go-to-market motion3 records

Distribution channels7 records

Marketing channels7 records

Core offering

Anthropic builds and sells the Claude family of large language models and AI products, accessed primarily through a public developer API and consumer/enterprise applications including the Claude.ai assistant, Claude Code for software development, Claude Cowork for autonomous task execution, Claude Design, Claude Science for biopharma research, and Claude Security for cybersecurity. Underlying training uses Constitutional AI and RLHF methodologies, and models are distributed across Anthropic-owned channels and major cloud marketplaces (AWS, Google Cloud Vertex AI, Microsoft Foundry).

Product overview

Anthropic PBC is an AI safety and research company offering a comprehensive product portfolio centered on Claude, its family of large language models. The core offering is the Claude AI assistant available across web, desktop (Mac/Windows/Linux), and mobile (iOS/Android) platforms. The product architecture follows a platform-plus-modules design: the Claude Platform provides API access for developers, while specialized products include Claude Code (AI coding tool), Claude Cowork (autonomous task execution), Claude Design (design work), Claude Science (biopharma research), and Claude Security (cybersecurity). The model family includes five tiers: Mythos (most advanced/restricted), Fable (flagship), Opus (high-performance), Sonnet (mid-tier), and Haiku (lightweight). Subscription tiers range from Claude Pro (individual) through Claude Max and Claude Team to enterprise offerings including Claude Code Enterprise. Claude is available natively on AWS, Google Cloud, and Microsoft Foundry. Claude Opus 5 launched July 2026 as a cost-effective alternative approaching frontier-level intelligence.

Differentiator

Problem solved

Functional benefit

Products and services

• Claude

Quantifiable outcome

• 2x to 8x productivity gains using Claude Code in enterprise settings, with troubleshooting time cut by 80-90%
• +4 more outcomes

Named customers16 records

Segments7 records

Ideal customer profiles5 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration9 records

AI capability10 records

Feature9 records

Partnerships

17 partnerships are on record, tiered core, major, strategic and minor.

• AMDcoreStrategic or Co-development PartnerAMD announced strategic investment of up to $5 billion in Anthropic in exchange for commitment to purchase tens of billions of dollars worth of AMD servers. Anthropic will deploy up to 2 gigawatts of MI450 series GPUs in Helios rack systems starting H1 2027. AMD will also adopt Claude to accelerate ROCm software platform development. Deal positions AMD as direct challenger to Nvidia while diversifying Anthropic's compute infrastructure.
• TeraWulfcoreStrategic or Co-development Partner$19 billion, 20-year lease agreement for Anthropic to house AI workloads at Justified Data Campus in Hawesville, Kentucky. Initial capacity expected H2 2027, full ramp-up by early 2028 providing 401 megawatts of computing power. TeraWulf plans $3-4 billion facility investment.
• Samsung ElectronicsmajorStrategic or Co-development PartnerSamsung Chairman Lee Jae-yong met with Anthropic CEO during San Francisco AI Summit. Discussions focused on AI semiconductor cooperation and Korean investment plans. Anthropic indicated intent to sign MOU with Korea's science ministry on AI security cooperation.
• Meta PlatformsmajorStrategic or Co-development PartnerEarly-stage discussions for potential $10 billion computing lease over two years. Anthropic proposed arrangement in June 2026 with monthly installments and early exit provisions. Would help Anthropic boost compute capacity while providing Meta new revenue stream from infrastructure investments.
• JPMorgan ChasestrategicTechnology or IntegrationTesting Anthropic's Mythos AI model since April 2026 to assess cyber defenses. CEO Jamie Dimon warned Mythos could be dangerous if widely released, comparing broad access to 'giving ballistic missiles to individuals.' Model identified tens of thousands of previously unknown vulnerabilities.
• Brown & BrownstrategicImplementation/ SI/ Consulting PartnerInsurance brokerage selected Anthropic along with McKinsey and Accenture for enterprise AI transformation. Early pilots using Claude Code show productivity gains of 2x to 8x. Plans to deploy across 23,000-person workforce.
• LTMstrategicImplementation/ SI/ Consulting PartnerStrategic partnership to accelerate enterprise adoption of Claude, Claude Code, and Claude Cowork. LTM's BlueVerse AI Delivery Fabric serves as implementation layer. Targeting BFSI, Hi-Tech, Consumer sectors. AI1000 initiative training thousands of Claude-certified architects.
• FujitsustrategicImplementation/ SI/ Consulting PartnerFujitsu's AI-driven Modernization Service combines Claude with Kozuchi platform and Takane LLM. Claims 40% reduction in migration timelines. Targets finance, healthcare, manufacturing sectors in Japan.
• FISstrategicImplementation/ SI/ Consulting PartnerPartnership to develop AI agents that automate financial crime investigations. First agents expected in H2 2026. Part of FIS's broader AI strategy including Project Keystone tokenized deposit network.
• GlobantstrategicImplementation/ SI/ Consulting PartnerMulti-year AI alliance to scale agentic AI services and modern web deployments for global enterprises. AI Pod initiative signed 18 clients and drives higher-margin recurring revenue.
• 1PasswordstrategicTechnology or IntegrationIntegration enabling Claude to sign into websites without accessing actual passwords. Zero-exposure architecture with Agentic Mode for AI agent security. First event-focused connector natively in Claude.
• CISAstrategicTechnology or IntegrationUS Cybersecurity and Infrastructure Security Agency using Anthropic's Mythos AI model to scan federal government code repositories for security vulnerabilities. White House mandated agencies strengthen cyber defenses within 30 days.
• Amazon Web ServicesmajorChannel Partner/ Reseller/ DistributorClaude available on AWS through Claude on AWS partnership. AWS carries $364 billion commercial backlog plus $100 billion-plus Anthropic deal. Andy Jassy secured $225 billion in Trainium capacity commitments from Anthropic and OpenAI.
• Google CloudmajorChannel Partner/ Reseller/ DistributorClaude available through Google Cloud Vertex AI. Google Cloud reported record $22.4 billion revenue (64% YoY growth) with $514 billion order backlog. Alphabet raised capex guidance to $195-205 billion for AI infrastructure.
• Microsoft FoundrymajorChannel Partner/ Reseller/ DistributorClaude available through Microsoft Azure AI Foundry. Microsoft AI CEO Mustafa Suleyman stated Anthropic is 'extremely expensive' as company shifts to internal MAI models.
• SK Group / SK HynixmajorStrategic or Co-development PartnerSK Group Chairman revealed Anthropic requested supplies from SK Hynix to manufacture its own semiconductors - remarkable development for AI company to pursue chip-making ambitions. SK Group deal includes $500 billion with NVIDIA; Anthropic seeking to reduce dependence on external chip suppliers.
• OpenAIminorOthersNot partners but major competitor. Both companies have accused Chinese AI companies of distillation. OpenAI reached 10 million AI agent users versus Anthropic's Claude agent platform.

Scale indicators12 records

Recent moves9 records

Expansion highlights9 records

Direct peers

• Cursor: AI-first IDE and coding assistant that competes directly with Claude Code for developer mindshare and seat-based revenue. Operates in the same AI coding workflow segment where Anthropic reports 2x–8x enterprise productivity gains.
• Perplexity AI: AI-native search and answer engine built on foundation models, competing for consumer attention, subscription dollars, and enterprise knowledge workflows. Uses comparable frontier LLMs (including Claude itself) and pressures Anthropic on consumer monetization.
• OpenAI: Directly comparable frontier AI lab offering competing foundation models (GPT series), an API platform, and consumer/enterprise products (ChatGPT, Codex). Both compete for the same developer, enterprise, and consumer mindshare, and Anthropic is explicitly described as OpenAI's most significant competitor.
• Google DeepMind: Alphabet's AI lab behind Gemini and related foundation models, distributed through Google Cloud Vertex AI—the same channel through which Anthropic also reaches enterprise customers. Competes head-to-head on frontier benchmarks, enterprise deals, and developer APIs.
• xAI: Elon Musk's AI company building the Grok family of frontier models with comparable capital intensity, custom infrastructure ambitions, and a mix of consumer and developer/API distribution. Competes for the same top-tier model customers and hyperscaler partnerships.

Emerging players

• Mistral AI: European frontier/open-weight model provider offering competitive LLMs with strong enterprise and developer adoption in EMEA. Comparable in API-first GTM and platform strategy, though with a smaller model portfolio and a stronger open-weight focus than Anthropic.
• Cohere: Enterprise-focused LLM provider offering models via API and through cloud partners, with emphasis on retrieval, agents, and private deployments. Comparable to Anthropic's enterprise go-to-market and partnership-driven distribution model.

Broad incumbents

• Microsoft Copilot / Azure AI: Microsoft's portfolio of Copilot products and Azure AI Foundry offers enterprise AI assistants and a model marketplace that includes Claude. Overlaps with Anthropic on enterprise productivity, coding assistance, and cloud-distributed foundation models.
• AWS (Bedrock / Amazon Q): Hyperscaler-owned AI platform (Bedrock) that hosts Anthropic's models alongside its own Titan/Q families. Comparable as both a distribution channel for Claude and a competitor with proprietary models targeting the same enterprise AI workloads.

Others

• Hugging Face: Open model hub and inference platform that lists Claude alongside other LLMs and offers tooling for agents and fine-tuning. Not a direct model competitor but a shared ecosystem venue for developers and enterprises evaluating Anthropic's APIs.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks6 records

Key highlights7 records

Customer concentration

• GitHubgithub.com/anthropics
• LinkedInlinkedin.com/company/anthropicresearch
• Xx.com/AnthropicAI

Compliance6 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles5 records

Subsidiaries3 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Anthropic PBC