Claude Weekly: IPO Filing, $65B Raise, Glasswing Goes Global

This was the week Anthropic went from “fastest-growing private AI company” to “soon-to-be-public AI company.” On June 1, Anthropic filed a confidential draft S-1 with the SEC, hot on the heels of a $65 billion Series H that values the company at $965 billion. Project Glasswing expanded to ~150 critical-infrastructure organizations in 15+ countries, Claude Opus 4.8 continued driving new product capabilities, and a significant Agent SDK billing restructure takes effect June 15. This is the Claude Weekly for June 4, 2026. Read last week’s edition.

Anthropic Files for IPO

The single biggest story this week: Anthropic submitted a confidential draft S-1 to the SEC on June 1, beginning the formal process of going public. The number of shares and price range haven’t been set — those details come with the public S-1 filing after SEC review — but the move signals that Anthropic believes its financials and governance are ready for public market scrutiny. (Anthropic, TechCrunch)

The IPO filing follows last week’s $65B Series H at a $965B valuation — co-led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, with strategic infrastructure investments from Samsung, SK Hynix, and Micron. $15B of the round came from previously committed hyperscaler investments including $5B from Amazon.

What matters for enterprise buyers: A publicly traded Anthropic means financial transparency, audited controls, and long-term stability — all of which make procurement conversations easier. The existential “will they be around in three years” question that hung over earlier enterprise evaluations is effectively retired. Anthropic’s run-rate revenue has crossed $47 billion, and the company is positioned for what could be the largest AI IPO alongside an anticipated OpenAI public offering.

Project Glasswing: From 50 Partners to 150

Anthropic published details on the expansion of Project Glasswing, its critical-infrastructure security initiative using Claude Mythos Preview for vulnerability discovery. The program has grown from 50 initial partners to ~150 organizations across 15+ countries, expanding into new sectors: power, water, healthcare, communications, and hardware. (Anthropic, TechCrunch)

Notable new partners include Okta, Samsung, SK Hynix, SK Telecom, NATO, and ENISA. Since launch, Mythos has identified 10,000+ high- or critical-severity flaws across partner codebases — a volume that no human security team could match at this scale.

Why this matters: Glasswing is the clearest demonstration yet of AI delivering asymmetric returns in security. The program also serves as Anthropic’s real-world safety testing ground for Mythos-class models, which the company expects to bring to general availability “in the coming weeks” once cyber safeguards are complete.

Claude Opus 4.8: The Platform Effects Are Starting

Last week’s Opus 4.8 release is producing a wave of downstream product updates worth tracking:

Dynamic Workflows (Research Preview) — Claude can now plan a large task, spawn hundreds of parallel subagents, verify outputs, and report back. Demonstrated on codebase-scale migrations across hundreds of thousands of lines of code. Available on Max, Team, and eligible Enterprise plans, plus through the API and partner platforms (Bedrock, Vertex AI, Foundry). Teams looking to apply Dynamic Workflows to enterprise code migrations can work with Big Hat Group’s application development team. The feature is generating significant community interest — Reddit threads on Dynamic Workflows have been among the most engaged Claude Code conversations this year. (InfoQ)

Claude Security — A new product that uses Opus 4.8 to scan codebases, validate findings (reducing false positives), and suggest patches. Released alongside Project Glasswing expansion and available now for Enterprise customers.

Claude for Small Business — 15 ready-to-run agentic workflows with connectors for QuickBooks, PayPal, HubSpot, Canva, DocuSign, Google Workspace, and Microsoft 365. Covers payroll planning, month-end close, invoice chasing, lead triage, and campaign generation via Cowork. This is Anthropic’s most aggressive push into the SMB segment and signals a widening competitive front against Microsoft Copilot and Google Gemini for Business. (Anthropic)

Claude Platform on AWS — Claude is now available as a managed platform on AWS in addition to Amazon Bedrock. This makes Claude the first frontier model available as a native platform on all three major clouds (AWS, GCP, Azure).

Cowork GA on all paid plans — file-based workspace with local-folder access, now generally available.

Agent SDK Billing Restructure: June 15 Deadline

The June 15 billing restructure is approaching. Agent SDK usage — including claude -p (headless Claude Code), GitHub Actions, and third-party ACP applications — moves out of subscription pools and into separate monthly dollar credits at standard API rates.

The new credit allocations:

The $0 for standard Enterprise seats is a commonly overlooked detail. Standard Enterprise seat holders get no programmatic usage credit at all. Unused credits do not roll over. (Digital Applied)

What teams should do now: Audit your Agent SDK and claude -p usage patterns. If you’re running production automation through subscription-priced accounts, you’ll need to budget separately or switch to direct API billing. The change effectively ends the 15–30× subsidy that made unlimited agent usage at subscription pricing tenable.

Claude Compliance API + Microsoft Purview Integration

Enterprise security teams can now govern Claude usage through Microsoft Purview, using the Claude Compliance API for programmatic access to conversation content and audit data. The integration enables compliance teams to detect, investigate, and report on Claude activity alongside the rest of their SaaS stack. (CloudWars)

This follows the Compliance API’s launch last week with 28 third-party integrations. For organizations where security is the gating factor on AI deployment, these integrations remove a significant blocker.

Community & Ecosystem

Code with Claude 2026 — The global roadshow continues with the Tokyo edition scheduled for June 10. Dario Amodei reported at the San Francisco event that Q1 2026 revenue grew 80× annualized vs. 10× planned — a number that explains both the compute partnerships and the IPO timeline. (InfoQ)

Databricks + Claude Managed Agents — A session at the Data + AI Summit demonstrated autonomous data agents built with Claude Managed Agents and Databricks tooling. (Databricks)

MCP ecosystem continues scaling. Sanity launched a hosted MCP server. Third-party MCP servers for Playwright, GitHub, and financial services connectors emerged. Claude Desktop supports one-click MCP server installation, and the MCP ecosystem now includes thousands of community-built servers.

Anthropic security research — The company published “What we learned mapping a year’s worth of AI-enabled cyber threats,” analyzing AI-transformed cyberattack methods mapped to the MITRE ATT&CK framework. (Anthropic)

What to Watch

• Agent SDK billing cutover (June 10–15) — Teams still on subscription-priced agent usage need to audit and migrate. The clock is ticking.
• Code with Claude Tokyo (June 10) — The APAC roadshow stop may bring region-specific announcements, particularly around cloud partnerships.
• Mythos-class model GA — Anthropic has stated it expects to bring Mythos-class models to all customers “in the coming weeks.” This would be the most significant model release since Opus 4.8.
• OpenAI IPO filing — Expected imminently, which would set up the most consequential AI public-market competition in history.
• Anthropic public S-1 — Watch for the public S-1 filing, which will reveal share count, price range, and financial details now kept confidential.

A week that started with the largest private fundraising in AI history and a confidential IPO filing doesn’t need embellishment. Anthropic is executing across product, security, and financial dimensions at a pace that fundamentally changes the competitive landscape.

If your team is navigating the June 15 Agent SDK billing changes, evaluating Claude Enterprise for production deployment, or building on Claude Managed Agents, talk to Big Hat Group today. We help enterprise teams move from evaluation to production with AI automation and agent deployment.

Check back next week.