← Kilroy’s Daily Briefings
🤖 AI News AM

AI News Briefing — May 7, 2026 at 11:28 AM

🤖 AI News AM5/7/2026🕐 6:00 AM⏱ 6:40AudioMorning

Top stories, ranked by relevance.

Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.

▶ Listen at 0:19

#1Anthropic Secures 220K+ GPUs from SpaceX's Colossus 1, Eyes Orbital Compute

Anthropic announced a blockbuster compute deal with SpaceX, gaining access to the full capacity of the Colossus 1 data center in Memphis — over 300 megawatts and 220,000-plus NVIDIA GPUs, available within the month. The deal will directly boost capacity for Claude Pro and Max subscribers. Perhaps most provocatively, Anthropic has "expressed interest" in working with SpaceX to develop multiple gigawatts of compute capacity in orbit.

🔗 CNBC
No image

#2Musk Dissolves xAI, Rebrands All AI Products Under "SpaceXAI"

Elon Musk announced today that xAI will cease to exist as a standalone company, with Grok and all AI products folding into SpaceX under the new "SpaceXAI" brand. The move follows the $1.25 trillion formal acquisition earlier this year and positions SpaceX's AI roadmap around space-based data centers as the company targets a mid-2026 IPO potentially valued at $1.75-2 trillion.

🔗 Republic World
No image

#3Anthropic Launches 10 Financial Services AI Agents, Tanks FactSet Stock 8%

Anthropic released ten ready-to-deploy agent templates for banking, insurance, and asset management — handling pitchbooks, KYC screening, month-end close, earnings analysis, and compliance review. The agents ship as plugins for Claude Cowork and Claude Code. FactSet shares plunged 8.1% on the news, with S&P Global and Morningstar also selling off sharply, as investors priced in direct competition with legacy financial data platforms.

🔗 Bloomberg
No image

#4Pentagon Awards Classified AI Contracts to 8 Companies, Blacklists Anthropic

The Pentagon finalized AI agreements with OpenAI, Google, Microsoft, Amazon, Oracle, Nvidia, SpaceX, and Reflection AI for classified military networks — while formally excluding Anthropic under a "supply chain risk" designation. The blacklisting stems from Anthropic's refusal to grant unrestricted access to Claude for autonomous weapons and mass surveillance. Anthropic has sued the administration, and a federal judge blocked the designation, but the contracts went out anyway.

🔗 CNN
No image

#5Trump Admin Embraces AI Safety Testing — CAISI Signs Google, Microsoft, xAI

In a notable policy pivot, the Center for AI Standards and Innovation signed pre-deployment evaluation agreements with Google DeepMind, Microsoft, and xAI, building on earlier deals with OpenAI and Anthropic. CAISI has now completed over 40 frontier model evaluations, including unreleased models. The White House is reportedly consulting experts on a formal government review process for new AI models — a departure from its prior light-touch approach.

🔗 CNBC

#6Google Kills Project Mariner, Folds Browser Agent Tech into Gemini Agent

Google quietly shut down Project Mariner on May 4, its standalone browser-controlling AI agent that launched in December 2024. The technology — web navigation, form-filling, multi-task automation — has been gradually migrated into Gemini Agent over the past year. It signals Google's consolidation strategy: fewer experimental standalone products, more capability packed into the Gemini umbrella.

🔗 Digital Trends
No image

#7Meta Building "Hatch" Consumer AI Agent and Instagram Shopping Tool

The Information reports Meta is developing "Hatch," a consumer-facing AI agent comparable to OpenClaw but designed for mainstream users within Meta's app ecosystem. Powered internally by Meta's Muse Spark model, Hatch is targeting end-of-June internal testing with mock environments simulating Reddit, Etsy, and DoorDash. Separately, Meta is building an agentic shopping tool for Instagram to rival TikTok Shop, planned for Q4 2026.

🔗 The Information
No image

#8NIST Evaluates DeepSeek V4 Pro — Capable but 8 Months Behind US Frontier

NIST's CAISI published its evaluation of DeepSeek V4 Pro, the 1.6-trillion-parameter open-source MoE model released in late April. The verdict: it is the most capable PRC AI model CAISI has evaluated to date, and more cost-efficient than GPT-5.4 mini on 5 of 7 benchmarks, but its overall capabilities lag the US frontier by roughly 8 months. The model activates only 49B of its 1.6T parameters per token and supports 1M-token context.

🔗 NIST
No image

#9IBM Think 2026: "Bob" Hits 80K Developers, watsonx Orchestrate Goes Multi-Agent

IBM used its Think 2026 conference to unveil its enterprise agentic AI stack. IBM Bob, its AI-powered development partner, is now generally available and used by 80,000 IBM developers with a reported 45% productivity boost. The next-gen watsonx Orchestrate enters private preview as a multi-agent control plane, letting organizations deploy and govern agents from any source with consistent policy enforcement.

🔗 IBM
No image

#10Microsoft: Global AI Adoption Hits 17.8% of Working-Age Population

Microsoft's latest Global AI Diffusion Report, published today, shows AI usage rose from 16.3% to 17.8% of the world's working-age population in Q1 2026. The UAE leads at 70.1%; the US climbed to 21st place at 31.3%. Notably, AI adoption is accelerating fastest in Asia — South Korea, Thailand, and Japan — driven by improving model performance in Asian languages. US software developer employment also rose 4% YoY through March 2026.

🔗 Microsoft
No image
Kilroy’s Daily Briefings · 5/7/2026