Anthropic Mythos model identifies classified US govt system vulnerabilities
Model flagged issues in hours during intel agency testing exercise; no exploitation confirmed.
Wednesday, June 24, 2026
Anthropic Mythos model identifies classified US govt system vulnerabilities — Wednesday, June 24, 2026
Groq $650M raise; Anthropic Mythos govt vuln test; Prosus OpenClaw rival; Anthropic Slack integration; Upscale AI $2B val.
Biggest developments
Anthropic launches Claude Tag research preview in Slack
Enterprise feature rollout amid $965B valuation; builds on business demand.
Prosus develops OpenClaw rival ToqanClaw for EU privacy compliance
Amsterdam firm rolls out secure agent this week to sidestep data concerns; shares features with OpenClaw.
Groq confirms $650M raise to scale AI inference cloud
Round led by Disruptive and Infinitum after Nvidia $20B not-acqui-hire; valuation undisclosed, prior at $6.9B.
Upscale AI raises $190M extension to reach $2B valuation
Nvidia, Salesforce Ventures, Temasek join; total funding $500M for AI networking infra.
Bullish takes
- Groq funding accelerates inference alternatives to Nvidia dominance.
- Anthropic govt testing validates Mythos capabilities at scale.
Critical takes
- Mythos vuln discovery raises dual-use risks for frontier models.
Why this matters
Fresh capital and security benchmarks signal accelerating enterprise deployment and regulatory scrutiny ahead of IPO windows.