Key Partners
- Amazon Bedrock — Claude model IDs for enterprise AWS buyers.
- Google Cloud — Claude available with Google-published IDs.
- Microsoft Foundry — Claude deployments for Azure/Foundry customers.
- Workplace embeds — Claude for Microsoft 365, Chrome, Slack Tag, connectors/MCP.
- Long-Term Benefit Trust + PBC board (governance partners for mission).
- Cloud/GPU suppliers for training and inference capacity.
Key Activities
- Train and evaluate frontier models (Fable 5.1, Opus 5, Sonnet 5, Haiku 4.5).
- Operate Claude apps (web/desktop/mobile) and the Claude API / Console.
- Ship Claude Code, Design/Slides/Docs, Managed Agents, connectors.
- Safety science — interpretability, misuse detection, policy engagement.
- Price and package Free → Pro → Max and Team/Enterprise seats.
- Enterprise sales, AWS Marketplace, and partner enablement.
Key Resources
- Frontier model IP (Fable / Opus / Sonnet / Haiku families).
- Claude consumer brand across web, desktop, iOS, Android.
- API platform, Console, Claude Code, Managed Agents, docs.
- People — research, engineering, policy, safety, and GTM teams.
- Governance — Public Benefit Corporation + Long-Term Benefit Trust.
- Trust assets — SSO/SCIM, audit logs, HIPAA-ready Enterprise path.
Value Propositions
- Claude — chat, files, Artifacts, research, voice, scheduled tasks.
- Claude Code + Design/Slides/Docs for end-to-end work products.
- Frontier models with long context (up to 1M tokens on flagship lines).
- API + Managed Agents — ship products on Claude with tools and runtime.
- Enterprise trust — SSO/SCIM, audit logs, HIPAA-ready, spend controls.
- Safety brand — reliable, interpretable, steerable systems (Company).
Customer Relationships
- Free forever on-ramp — no seat required to try Claude.
- Self-serve billing for Pro / Max (card); annual discount on Pro.
- Team — central billing, seat types, SSO, spend controls.
- Enterprise — sales-assisted path, CS at thresholds, ACH/invoicing.
- Help Center + docs; usage credits when session limits hit.
- Team/Enterprise — no model training on content by default.
Channels
- Primary doors — claude.ai / apps and anthropic.com product pages.
- PLG — Free → invite/share → paid Pro/Max or Team seats.
- Developer channel — API docs, Console, prepaid/usage billing.
- Cloud marketplaces — Bedrock, Google Cloud, Microsoft Foundry.
- Content — research blog, newsroom, Help Center, customer stories.
- Contact sales / AWS Marketplace for sales-assisted Enterprise.
Customer Segments
- Individuals — Free, Pro ($17–20), Max (from $100).
- Power users / developers — Max + Claude Code all day.
- SMBs & teams — Team 2–150, Standard or Premium seats.
- Enterprises — $20/seat + usage; SCIM, HIPAA-ready, CS.
- Developers & ISVs — Claude API token customers and agents.
- Cloud-platform buyers — Bedrock / GCP / Foundry consumers.
Cost Structure
- Training compute — multi-month frontier model runs.
- Inference COGS — Claude traffic + API token serving at scale.
- Research, engineering, and safety/policy talent.
- Product surfaces — Claude apps, API, Claude Code, agents.
- Sales, support, and solutions for Team/Enterprise.
- Compliance, trust programs, and multi-cloud distribution.
Revenue Streams
- Free — $0 Claude on web/desktop/mobile.
- Pro — $17/mo annual equiv. or $20 monthly; Claude Code included.
- Max — from $100/mo for 5× or 20× Pro usage.
- Team — Std $20 / Prem $100 per seat/mo annual ($25 / $125 monthly).
- Enterprise — $20/seat/mo + usage at API rates; annual billing.
- API — pay-as-you-go tokens (Fable $10/$50; Opus $5/$25; Sonnet $2/$10; Haiku $1/$5).
- Add-ons — Managed Agents $0.08/session-hour; code execution overage $0.05/hr.