Need to know
- Anthropic closes a $30B round at a $900B valuation as OpenAI files its IPO prospectus on the same day.
- OpenAI and Anthropic posted $5.7B and $4.7B in Q1 revenue respectively, confirming a two-horse race at the frontier.
- Okta integrates with Anthropic's Claude Compliance API to surface identity risks inside Claude Enterprise deployments.
- Microsoft Fara1.5 browser agents outperform OpenAI Operator and Gemini 2.5 on web automation benchmarks.
- GitLab 19.0 ships agentic merge request workflows, secrets management, and self-hosted open-source model support.
New Releases
GitLab 19.0 is generally available today, adding agentic merge request workflows, expanded secrets management, and support for self-hosted open-source models.
- Agentic MR workflows allow Duo Developer to operate inside discussion threads and propose code changes without a human initiating each step, extending AI beyond autocomplete into review and remediation.
- Self-hosted model support lets enterprises point GitLab Duo at their own open-source models rather than GitLab's hosted inference, a direct answer to data-sovereignty objections that have stalled enterprise rollouts.
Microsoft Research released Fara1.5, a family of 4B, 9B, and 27B browser computer-use agent models that outperform OpenAI Operator and Gemini 2.5 on the Online-Mind2Web benchmark.
- Three model sizes (4B/9B/27B) let enterprises deploy Fara1.5 on-premise or at the edge rather than routing every browser action through a hosted API, which matters for regulated industries.
- Benchmark leadership on Online-Mind2Web is significant because that dataset tests real-world web task completion, not synthetic prompts, making it a credible proxy for enterprise RPA displacement.
Okta announced its Identity Security Posture Management product now integrates with Anthropic's Claude Compliance API, correlating Claude Enterprise user activity with signals from other SaaS tools and cloud providers.
- Okta ISPM will surface identity misconfigurations and risky behavior within Claude Enterprise and Claude Platform alongside its existing coverage of other SaaS and cloud identity sources.
- CrowdStrike also announced a parallel integration feeding Claude Compliance API data into Falcon Next-Gen SIEM and Charlotte Agentic SOAR, suggesting Anthropic is building a compliance data ecosystem around enterprise Claude deployments.
Palo Alto Networks unveiled Idira, a next-generation identity security platform that adds agentic PAM capabilities covering human, machine, and AI agent identities in a single control plane.
- Agentic identity controls extend dynamic privilege management to AI agents running inside enterprise workflows, a gap that traditional CyberArk-style PAM tools were not designed to address.
- Positioned as an upgrade path for existing CyberArk customers, Idira arrives as AI agent proliferation is forcing enterprises to rethink privileged access architecture from scratch.
Funding
Anthropic closed a $30B fundraise at a $900B-plus valuation led by Sequoia, Dragoneer, Greenoaks, and Altimeter on the same day OpenAI filed its IPO prospectus, marking the moment two private AI labs simultaneously reached public-market scale.
OpenAI filed its S-1 today, the most anticipated tech IPO since Coinbase, putting a public-market price on a company reporting $5.7B in Q1 revenue while still burning cash on infrastructure at scale.
Anthropic's annualized revenue reached $45B versus OpenAI's $25B annualized pace, though The Information reported OpenAI led on Q1 actuals at $5.7B versus Anthropic's $4.7B, signaling aggressive Anthropic bookings acceleration in Q2.
Case Studies
Manulife Wealth and Asset Management announced three AI-powered sales and advisor enablement tools designed to reduce routine work and surface client insights faster across its advisor and sales teams.
- The tools target friction reduction in day-to-day advisor workflows, freeing time for relationship and advice work, consistent with the pattern of AI delivering the most measurable ROI in high-touch sales roles.
- No specific productivity metrics were disclosed, but Manulife's framing around advisor time reallocation rather than headcount reduction signals an adoption-first rather than efficiency-first deployment strategy.
Bloomberg reported that Salesforce's Agentforce promotional case studies, including University of Chicago Medicine, involve capabilities not yet fully deployed in production, exposing a gap between marketed outcomes and live results.
- University of Chicago Medicine was featured in Salesforce Agentforce video marketing for autonomous patient appointment and prescription workflows, but the Bloomberg investigation found those capabilities had not shipped at the scale shown.
- The SaaSapocalypse pressure is forcing Salesforce to lead with AI narrative ahead of documented enterprise outcomes, a risky posture as buyers grow more demanding about proof of production deployment.
Trending on X
- OpenAI vs Anthropic revenue race The Information's report that OpenAI led Q1 at $5.7B versus Anthropic's $4.7B is being dissected alongside Anthropic's higher annualized run rate, with the community debating which metric actually signals future dominance.
- Andrej Karpathy joins Anthropic Karpathy leaving OpenAI's orbit and joining Anthropic to train future Claude models is being treated as the clearest signal yet that talent is flowing toward the challenger lab, not the incumbent.
- The model alone is no longer the product OpenAI's Greg Brockman (@gdb) posted that framing and sparked a debate with Ethan Mollick about whether labs' harnesses and product control actually insulate them from commoditization or whether model quality remains the prime mover.
- White House cancels AI pre-release review order A proposed 90-day government vetting gate for frontier models was killed before signing, and X is debating whether that was a win for innovation velocity or a missed opportunity to establish safety accountability before OpenAI's IPO.
- Grok flopping in US federal government Reuters and multiple federal employees confirmed Grok has near-zero adoption across US government agencies despite xAI's investment push, undermining a core part of SpaceX's IPO AI growth narrative.