Sells banks Claude seats and API access directly and through cloud marketplaces, and in 2026 added embedded engineering with Goldman Sachs and FIS, restricted cyber-defence access to its Mythos model for a few large banks, and a services firm co-founded with Goldman Sachs.
What is Claude, and which banks use it?
Claude is the model family and workplace assistant from Anthropic, a San Francisco public benefit corporation founded in 2021 and valued at $965 billion in May 2026; it filed confidentially for a US IPO in June 2026. Among our 100 banks, Goldman Sachs and JPMorganChase offer Claude to staff, BankUnited and Flagstar run Claude in internal AI platforms, and several large banks test the restricted Mythos model. Claude Enterprise lists at $20 per seat per month plus usage at API rates.
Category
Founded
2021
Headquarters
San Francisco, California, US
Ownership
Product of Anthropic PBC, a private public benefit corporation; valued at $965 billion post-money after a $65 billion Series H in May 2026, with Amazon, Google, Microsoft and Nvidia among strategic backers; confidentially filed for a US IPO on June 1, 2026.[6][4][10][11]
Deployment
SaaS workspace (Claude Team and Enterprise) and API (cloud); Claude models are also sold through Amazon Bedrock and run inside Microsoft 365 Copilot
Pricing
Per-seat subscriptions for Claude Team; Claude Enterprise at a per-seat fee plus usage billed at API rates; API billed per million input and output tokens by model
Banks among the 100
7
- Anthropic's pricing page lists Claude Team at $20 per standard seat and $100 per premium seat per month billed annually, and Claude Enterprise at US$20 per seat per month billed annually plus usage at API rates; API prices are published per million tokens by model.
- Anthropic raised $30 billion at a $380 billion valuation in February 2026 and $65 billion at $965 billion in May 2026, reporting run-rate revenue above $47 billion; it confidentially filed for a US IPO on June 1, 2026.
- In March 2026 the Pentagon designated Anthropic a supply-chain risk after a dispute over the company's usage safeguards, barring contractors from using Claude in Department of War work; on August 27, 2026 a federal judge in San Francisco ruled the designation unlawful, though a separate case over a second designation was still pending in Washington, D.C., so CNBC reported Anthropic technically remained a supply-chain risk.
- Anthropic restricted its Claude Mythos Preview model, which it says has found thousands of high-severity software vulnerabilities, to Project Glasswing launch partners such as JPMorganChase and more than 40 additional organisations; Reuters reported in May 2026 that US banks were rushing to fix weaknesses the tool flagged.
- JPMorganChase removed Claude models from the approved list for its Hong Kong staff in June 2026, a decision based on language in Anthropic's usage terms, according to the Financial Times (citing three people familiar with the matter) as reported by Investing.com; the FT had earlier reported a similar move by Goldman Sachs.
- Microsoft runs Claude inside Microsoft 365 Copilot as a subprocessor and switches it on by default for most commercial customers outside the EU/EFTA and UK, but excludes Anthropic models from its EU Data Boundary commitments.
- Outside our 100 US banks, Anthropic's financial-services materials cite Commonwealth Bank of Australia, AIG, IG Group and Moody's as customers.
What does Claude sell to banks?
| Product | What it does | Status |
|---|---|---|
| Claude Enterprise[1][2] | Seat-based workplace plan with usage billed at API rates, SCIM, audit logs, compliance API, custom data retention, network-level access control and a HIPAA-ready option. | Generally available |
| Claude Team[1] | Workplace plan for 2 to 150 users with standard and premium seats, single sign-on, admin controls and no model training on content by default. | Generally available |
| Claude API[1] | Per-token access to Claude models, with batch processing at 50% off, prompt caching and a US-only inference option. | Generally available |
| Claude for Financial Services[3][9] | A financial-analysis solution launched July 2025 that connects Claude to market data and internal data platforms such as FactSet, S&P Global, Databricks and Snowflake. | Generally available |
| Claude Mythos Preview (Project Glasswing)[7] | A restricted model used for defensive security work by Glasswing launch partners and more than 40 additional organisations; Anthropic says it does not plan to make it generally available. | Limited availability |
| Claude on Amazon Bedrock[26] | Claude models sold through AWS's Bedrock platform at per-token rates set on AWS's pricing page. | Generally available |
How much does Claude cost?
How it is charged: Per-seat subscriptions for Claude Team; Claude Enterprise at a per-seat fee plus usage billed at API rates; API billed per million input and output tokens by model.
| Product | Plan | List price | Unit | Checked |
|---|---|---|---|---|
| Claude Enterprise | Claude Enterprise[1] | US$20 | per seat per month, billed annually, plus usage at API rates | Sep 24, 2026 |
| Claude Team | Standard seat[1] | $20 | per seat per month billed annually ($25 if billed monthly) | Sep 24, 2026 |
| Claude Team | Premium seat[1] | $100 | per seat per month billed annually ($125 if billed monthly); 5x the usage of standard seats | Sep 24, 2026 |
| Claude API | Fable 5.1[1] | $10 input / $50 output | per 1M tokens (MTok) | Sep 24, 2026 |
| Claude API | Sonnet 5[1] | $2 input / $10 output | per 1M tokens (MTok) | Sep 24, 2026 |
List prices are published on Anthropic's pricing page; enterprise usage cost scales with the models and tasks used, so total cost depends on consumption. Negotiated terms for banks are not published, and access to Claude Mythos Preview is not sold on a public price list.
Which banks use Claude?
| Bank | What | Status |
|---|---|---|
| Goldman Sachs | Claude offered to staff in the GS AI assistant; Claude-based agents for trade accounting and client onboarding in early development with embedded Anthropic engineers (February 2026); has access to Mythos.[18][17][14] | In production |
| JPMorgan Chase | Claude models on the bank's internal list of approved LLMs under a licensing agreement (withdrawn for Hong Kong staff in June 2026); Project Glasswing launch partner testing Mythos.[19][7][16] | In production |
| BankUnited | SAVI internal policy assistant built on Anthropic's Claude 2 in Amazon Bedrock (95% accuracy reported by AWS).[23] | In production |
| Flagstar Bank | StarIQ multi-model gen-AI platform, deployed early 2026, integrating Anthropic Claude with other models on AWS.[22] | In production |
| Bank of America | Part of Project Glasswing and testing Mythos internally, according to a Reuters source.[14] | Pilot |
| Citigroup | Has access to Mythos for internal tests (Reuters source); Citi says it is involved in Project Glasswing.[14][20] | Pilot |
| Morgan Stanley | Permissioned on Claude Mythos Preview, per the CEO on the Q1 2026 earnings call.[14] | Pilot |
Among the 100 largest US banks on this site, from sources that name the bank: the bank, Claude or the press. Status is the furthest stage a source confirms.
What is Claude Enterprise?
Anthropic's enterprise plan, launched September 2024, priced per seat plus usage at API rates, with SCIM, audit logs, a compliance API, custom data retention, IP allowlisting and no model training on customer content by default, per Anthropic's pricing page.[1][2]
What has Claude done, and when?
- Sep 14, 2026
Schwab joins Claude for Financial Advisors[24]
Schwab Advisor Services became the first RIA custodian in Anthropic's Claude for Financial Advisors, connecting RIAs to Schwab Advisor Center from Claude.
- Aug 27, 2026
Court rules the Pentagon designation unlawful[13]
U.S. District Judge Rita Lin in San Francisco found the Pentagon violated the First Amendment by designating Anthropic a supply-chain risk in retaliation; a separate case over a second designation remained pending in Washington, D.C.
- Jun 18, 2026
JPMorgan withdraws Claude for Hong Kong staff[19]
JPMorgan's Hong Kong staff could no longer select Claude models, a decision based on language in Anthropic's usage terms, following a similar move by Goldman Sachs, according to the Financial Times as reported by Investing.com.
- Jun 1, 2026
Confidential IPO filing[11]
Anthropic said it had confidentially filed for a US initial public offering.
- May 28, 2026
$65 billion Series H at $965 billion[6]
Anthropic raised $65 billion and said run-rate revenue had crossed $47 billion.
- May 12, 2026
Banks rush to fix Mythos findings[15]
Reuters reported US banks were rushing to fix IT weaknesses flagged by Mythos, raising the possibility of customer disruption.
- May 4, 2026
- Apr 20, 2026
Large US banks gain Mythos access[14]
Reuters reported that Bank of America, Morgan Stanley, Goldman Sachs and Citigroup had access to or were testing Mythos, as regulators reviewed its cyber risks.
- Apr 7, 2026
Project Glasswing and Mythos Preview[7]
Anthropic gave launch partners including JPMorganChase restricted access to Claude Mythos Preview for defensive security work and committed up to $100 million in usage credits.
- Mar 5, 2026
Pentagon designates Anthropic a supply-chain risk[12]
The designation barred contractors from using Anthropic's technology in their military work; Anthropic said it applied only to direct Department of War contracts and challenged it in court.
- Feb 12, 2026
$30 billion Series G at $380 billion[5]
Anthropic raised $30 billion led by GIC and Coatue and reported run-rate revenue of $14 billion.
- Feb 6, 2026
Goldman Sachs co-develops Claude agents[17]
Goldman's CIO said embedded Anthropic engineers had spent six months co-developing Claude-based agents for trade accounting and client vetting and onboarding.
- Jul 15, 2025
Claude for Financial Services[3]
Anthropic introduced a financial-analysis solution connecting Claude to market data and internal data platforms.
- Sep 10, 2024
Claude Enterprise plan launched[2]
Anthropic launched the Claude Enterprise plan with an expanded context window, more usage capacity and a native GitHub integration.
What do banks use Claude for?
| Use | What it does | Status | Use case |
|---|---|---|---|
| Employee assistants and back-office agents[18][17][19] | Claude in Goldman's GS AI and JPMorganChase's approved-model list; Goldman's Claude-based agents for trade accounting and client onboarding. | In production | Generative & agentic AI |
| AML investigation agent[21] | FIS's Financial Crimes AI Agent, built with Anthropic, assembles evidence for AML alerts and SAR narratives; BMO is in development with it. | Pilot | AML / KYC |
| Vulnerability discovery with Mythos[14][7] | JPMorganChase, Bank of America, Morgan Stanley, Goldman Sachs and Citigroup use or test the restricted Mythos model to find and fix software vulnerabilities. | Pilot | Cybersecurity |
| Internal policy assistants for branch staff[23] | BankUnited's SAVI answers employee questions from about 400 policy documents using Claude on Amazon Bedrock. | In production | Customer-facing chatbots |
| Usage terms and jurisdiction[19][25] | JPMorganChase withdrew Claude for Hong Kong staff over Anthropic's usage terms, and Microsoft excludes Claude from its EU Data Boundary inside Copilot. | In production | Third-party & vendor AI |
What has Claude disclosed in numbers?
| Metric | Value | As of |
|---|---|---|
| Post-money valuation | $965 billion, after a $65 billion Series H[6] | May 28, 2026 |
| Run-rate revenue (company-reported) | More than $47 billion, up from $14 billion in February 2026[6][5] | May 28, 2026 |
| Customers spending over $1 million a year (company-reported) | More than 500; eight of the Fortune 10 are Claude customers[5] | Feb 12, 2026 |
| Mythos Preview usage credits committed | Up to $100 million, plus $4 million to open-source security groups[7] | Apr 7, 2026 |
| BankUnited SAVI accuracy (as reported by AWS) | 95% accuracy, responses in under 10 seconds[23] | Sep 24, 2026 |
“Whether it is entrepreneurs, startups, or the world's largest enterprises, the message from our customers is the same: Claude is increasingly becoming critical to how businesses work,”[5]
Krishna Rao · Chief Financial Officer, Anthropic · Feb 12, 2026
“We have the model. We're working closely with Anthropic and all of our security vendors to kind of harness frontier capabilities wherever it's possible.”[14]
David Solomon · CEO, Goldman Sachs · Apr 20, 2026
Who runs Claude?
Which rules apply to a bank using Claude?
| Authority | Why it matters here | Documents |
|---|---|---|
| OCC | Licensing Claude directly, through FIS or through a cloud marketplace is a third-party relationship; the 2023 interagency guidance expects banks to review terms such as the usage restrictions that led JPMorganChase to withdraw Claude for Hong Kong staff. | OCC Bulletin 2023-17 FDIC FIL-29-2023 |
| NY DFS | NY DFS's May 2026 guidance frames the cyber risks from frontier AI models as Part 500 obligations, which covers both defending against such models and fixing the vulnerabilities that tools like Mythos surface. | DFS Frontier AI Models Industry Letter (May 2026) 23 NYCRR Part 500 |
| Federal Reserve | SR 26-2 leaves generative and agentic AI outside model-risk scope, so Claude-based agents in accounting, onboarding and AML review are governed through enterprise risk management and the bank's own controls. | SR 26-2 |
| UK (BoE / PRA / FCA) | The May 2026 Bank of England, FCA and HM Treasury statement says frontier models' cyber capabilities exceed a skilled practitioner's and expects firms to identify and remediate vulnerabilities at scale, the work banks with Mythos access describe doing, and to manage third-party AI suppliers. | BoE/FCA/HMT joint statement on frontier AI and cyber resilience (May 2026) |
| ECB | The ECB's July 2026 letter to significant institutions says emerging AI models can find vulnerabilities and generate working exploits at unprecedented speed and requires action plans under DORA, the capability Anthropic describes for Mythos. | ECB 'Dear CEO' letter on AI-enabled cybersecurity threats (SSM-2026-0301) |
Linked authorities have a page in the AI Regulation Tracker; the others are named for completeness.
What should a bank note before buying Claude?
Analysis by BankingNewsAI from the sources cited on this page. Observations, not advice or a recommendation.
01
Seat price plus usage
Claude Enterprise lists a $20 per-seat monthly fee plus usage at API rates, so the total bill depends on which models and how much work staff run.
02
Most large-bank activity is restricted-model testing
Bank of America, Citigroup and Morgan Stanley appear here for access to or testing of Mythos, a model Anthropic does not plan to release generally; that is different from a staff-wide Claude deployment.
03
Usage terms have jurisdictional effects
JPMorganChase withdrew Claude for Hong Kong staff in June 2026 over language in Anthropic's usage terms, after a similar move by Goldman Sachs, according to the Financial Times as reported by Investing.com.
04
Government dispute is in the record
The Pentagon's March 2026 supply-chain-risk designation applied to Department of War contract work; a San Francisco federal judge ruled it unlawful in August 2026 while a parallel case in Washington, D.C. was still pending; banks with defence-related contracts may want the current status.
05
Several deployments reach Claude through another vendor
BankUnited uses Claude through Amazon Bedrock, Flagstar through its AWS-based StarIQ platform, and BMO Financial Group (the Canadian group; US scope not stated) through an FIS agent; the direct contract may be with that intermediary.
What are the alternatives to Claude?
Microsoft 365 Copilot
AI assistant inside Word, Excel, Outlook and Teams, grounded in a firm's Microsoft 365 data, plus Copilot Studio for building agents.
ChatGPT Enterprise (OpenAI)
OpenAI's workplace ChatGPT with admin controls and no training on business data by default, plus API access to GPT models.
Gemini Enterprise (Google Cloud)
Google Cloud's workplace AI app and agent platform on Gemini models, connecting to Google Workspace, Microsoft 365 and business data.
Amazon Bedrock (AWS)
AWS service giving one API to foundation models from Amazon, Anthropic, OpenAI, Meta and others, plus AgentCore for running agents.
GitHub Copilot
AI coding assistant and coding agent in the IDE, CLI and GitHub.com, with Business and Enterprise plans billed per seat plus AI credits.
How much does Claude Enterprise cost a bank?
Anthropic's pricing page lists Claude Enterprise at US$20 per seat per month, billed annually, plus usage billed at API rates, so cost scales with the models and tasks used (checked September 24, 2026). Claude Team is $20 per standard seat or $100 per premium seat per month billed annually, and API models such as Sonnet 5 list at $2 per million input tokens and $10 per million output tokens.
Which banks use Claude?
Among the 100 banks we track, Goldman Sachs and JPMorganChase offer Claude to staff through internal AI platforms, BankUnited runs an internal assistant on Claude via Amazon Bedrock, and Flagstar's StarIQ platform integrates Claude. Bank of America, Citigroup and Morgan Stanley have access to or test the restricted Mythos model. Each has a cited source on this page.
Who owns Anthropic?
Anthropic is a private public benefit corporation launched in 2021 by founders including CEO Dario Amodei and Daniela Amodei. It was valued at $965 billion after a $65 billion Series H in May 2026; Amazon, Google, Microsoft and Nvidia have committed investments, and it confidentially filed for a US IPO on June 1, 2026.
What is Claude Mythos and which banks have it?
Claude Mythos Preview is a restricted Anthropic model that Anthropic says has found thousands of high-severity vulnerabilities. It is available to Project Glasswing partners, including JPMorganChase, and more than 40 additional organisations for defensive security work. Reuters reported that Bank of America, Morgan Stanley, Goldman Sachs and Citigroup also had access or were testing it in April 2026.
Was Anthropic banned by the US government?
In March 2026 the Pentagon designated Anthropic a supply-chain risk, barring contractors from using its technology in their work for the US military; Anthropic said commercial use outside those contracts was unaffected and sued. On August 27, 2026, a federal judge in San Francisco ruled the designation was unlawful First Amendment retaliation; a separate case over a second designation was still pending in Washington, D.C., so Anthropic technically remained designated, CNBC reported.
Claude vs ChatGPT Enterprise: how do they compare for banks?
Both are sold as workplace plans and APIs. Anthropic publishes Claude Enterprise at $20 per seat plus usage, while OpenAI lists ChatGPT Enterprise as custom pricing. Several banks, including Goldman Sachs and JPMorganChase, offer both through internal portals that let staff choose a model, so the choice is often per task rather than exclusive.
- Plans & Pricing | Claude by Anthropic (undated page, retrieved 2026-09-24)
- Claude for Enterprise
- Claude for Financial Services
- Company (undated page, retrieved 2026-09-24)
- Anthropic raises $30 billion in Series G funding at $380 billion post-money valuation
- Anthropic raises $65B in Series H funding at $965B post-money valuation
- Project Glasswing: Securing critical software for the AI era
- Building a new enterprise AI services company with Blackstone, Hellman & Friedman, and Goldman Sachs
- Deploying Claude across financial services
- 1. Anthropic
- Anthropic moves toward IPO, stepping up race with OpenAI
- Pentagon designates Anthropic a supply chain risk
- Judge blocks Pentagon blacklist of Anthropic as supply chain risk
- Banking industry scrambles for Anthropic's Mythos as global regulators review risks
- Anthropic's Mythos sends US banks rushing to plug cyber holes
- Jamie Dimon says Anthropic's Mythos reveals 'a lot more vulnerabilities' for cyberattacks
- Goldman Sachs taps Anthropic's Claude to automate accounting, compliance roles
- Goldman Sachs staff now write a million gen AI prompts a month
- JPMorgan blocks Hong Kong staff from using Anthropic AI tools- FT
- A Competitive Differentiator: David Griffiths on Building the World's Most AI-empowered Financial Institution
- FIS Brings Agentic AI to Banking with Anthropic, Starting with Financial Crimes
- Flagstar Bank Announces Proprietary Technology Transformation
- BankUnited Enhances Customer and Employee Satisfaction for SMB Banking Using AWS (undated case study, retrieved 2026-09-24)
- Charles Schwab and Anthropic to Bring Claude to Independent Registered Investment Advisors
- Anthropic models in Microsoft Online Services (undated page, retrieved 2026-09-24)
- Amazon Bedrock Pricing (undated page, retrieved 2026-09-24)
When a bank signs with Claude, you'll read it here first.
when a bank picks an AI vendor, or a vendor ships for banks, it is in the next morning's brief · six sourced stories · 7 am ET · free
plus every tracker, bank and agent page update, the morning after · leave any morning