Anthropic today made the Claude Platform generally available on Amazon Web Services, bringing the complete Claude API feature set to AWS customers with native AWS authentication and billing. The launch includes managed agents, code execution, web tools, skills, and prompt caching — all accessible under existing AWS commitment structures without a separate Anthropic contract.
Anthropic is reportedly in active fundraising talks for a round of $30–50 billion that would peg the company's valuation at approximately $950 billion — within striking distance of a trillion-dollar mark. The round would cement Anthropic as one of the most valuable private companies in history and follows a string of major enterprise partnerships including Gates Foundation and PwC.
Anthropic's new Claude for Small Business tier brings ready-to-run AI workflows to popular SMB tools including QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace, and Microsoft 365. Pre-built automation handles payroll processing, invoice generation, sales pipeline management, and month-end close — without requiring any custom integration work from business owners.
Anthropic's unreleased Mythos model — described internally as "far ahead" of other AI systems in cybersecurity capability — has prompted concern from governments, financial institutions, and utility companies. Anthropic is restricting access to a vetted group of approved organizations rather than making it generally available, in a significant departure from its typical rollout approach. The move mirrors OpenAI's own tiered government access program for frontier models.
OpenAI introduced Daybreak, a new cybersecurity product that combines frontier model intelligence with Codex's agentic capabilities to help organizations find and patch software vulnerabilities. Daybreak integrates with existing security toolchains through OpenAI's security partners and can both identify weaknesses and validate that patches actually close them — an end-to-end loop that conventional scanners miss.
The OpenAI–Apple partnership is under strain, with OpenAI reportedly preparing potential legal action over how Apple has embedded ChatGPT within iOS. OpenAI's complaint centers on discoverability: the integration is considered too buried in Apple's interface, limiting the reach and usage that underpins the partnership's economics. Internal OpenAI lawyers are evaluating options with outside counsel.
Google's I/O developer conference opens next Tuesday with a keynote at 10 AM PT that is widely expected to center on a major Gemini model upgrade. "Gemini Omni" — a unified text, image, and video generation model — is the most anticipated reveal, alongside Android 17 with deep Gemini Intelligence integration, new Android XR smart glasses, and possible announcements around Veo 4 for video generation. The new model is expected to rival GPT-5.5's performance on coding and scientific benchmarks.
The National Institute of Standards and Technology announced agreements with Google, Microsoft, and xAI allowing the Center for AI Standards and Innovation to evaluate unreleased AI models before public launch. Testing will focus on cybersecurity implications and national security risks. OpenAI separately confirmed it is making its most advanced models available to vetted government agencies under a parallel access program aimed at getting ahead of AI-enabled threats.
GitHub is transitioning all Copilot plans to a credits-based billing model on June 1, 2026. Each plan will include a monthly allotment of GitHub AI Credits, with the option to purchase additional usage. Billing is calculated on token consumption (input, output, and cached tokens) at listed API rates per model. Code completions and Next Edit suggestions remain included in all plans without consuming credits. Existing price points — Pro at $10/month, Business at $19/user — are unchanged.