← Kilroy’s Daily Briefings
🧠 AI News PM

AI News Afternoon Briefing — May 2, 2026 at 3:00 PM

🧠 AI News PM5/2/2026🕐 3:00 PMAudioPM edition

Top stories, ranked by relevance.

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

#1Pentagon Signs Classified AI Deals With 7 Big Tech Firms, Freezes Out Anthropic

The Defense Department announced deals with SpaceX, OpenAI, Google, Microsoft, Nvidia, AWS, and Reflection AI to deploy their models on classified military networks, with Oracle added hours later. Anthropic remains locked out after the Trump administration labeled it a "supply chain risk" over the company's insistence on safety guardrails for military AI use. A federal judge blocked the government's ban, but a D.C. appeals court let the Pentagon's designation stand; White House talks with CEO Dario Amodei continue.

🔗 CNN
No image

#2Musk v. Altman Trial Week 1: Musk Admits xAI Distills OpenAI's Models

The first week of Elon Musk's $150B lawsuit against Sam Altman wrapped in Oakland with Musk testifying over three days that Altman and Brockman "stole a charity" by converting OpenAI into a for-profit engine. The bombshell: Musk admitted under cross-examination that his own company xAI uses OpenAI's models to train Grok, drawing audible gasps. The judge shut down Musk's existential-risk framing, noting he builds in the same space he warns against, and trial documents revealed Shivon Zilis served as Musk's covert liaison inside OpenAI.

🔗 MIT Technology Review
No image

#3Meta Stock Drops 9% Despite Earnings Beat as AI Capex Raised to $125–$145B

Meta beat Q1 expectations with $56.3B revenue and $10.44 EPS but saw its stock crater after raising full-year AI capital expenditure guidance to $125–$145 billion — nearly double last year's $72B. A $5B one-time tax benefit inflated the headline number, and free cash flow fell to $13.2B as $19B of $32B in operating cash went straight to AI infrastructure. Wall Street's message is clear: show the ROI or the spending story becomes a liability.

🔗 Yahoo Finance
No image

#4Google Signs Pentagon Classified AI Deal Despite 600+ Employee Revolt

Google proceeded with a classified military AI agreement with the Pentagon just hours after more than 600 Google and DeepMind employees publicly demanded CEO Sundar Pichai reject the deal. A senior DeepMind research scientist said he felt "incredibly ashamed," and the letter warned that classified workloads make oversight impossible. Google quietly removed its pledge not to pursue weapons or surveillance AI from its published AI Principles back in February 2025.

🔗 Washington Post
No image

#5Huawei Targets $12B in AI Chip Revenue as Ascend 950PR Enters Mass Production

Huawei expects AI chip revenue to hit approximately $12 billion in 2026, up 60% from $7.5 billion in 2025, driven by surging orders for its Ascend 950PR processor. ByteDance alone committed over $5.6B in Ascend purchases, with Alibaba and Tencent also placing large orders. Production targets are 750,000 units but demand is expected to outstrip supply — China's domestic AI chip stack is forming fast as Nvidia remains locked out by export controls.

🔗 Startup Fortune
No image

#6China Blocks Meta's $2B Acquisition of AI Startup Manus

Beijing ordered Meta to unwind its $2 billion acquisition of Manus, an agentic AI startup founded by Chinese engineers and based in Singapore. Chinese officials called the deal a "conspiratorial" attempt to hollow out China's tech base, and the decision was elevated to Xi Jinping's National Security Commission. Meta has indicated it will comply. The block underscores how AI acquisitions are now treated as national security matters on both sides of the Pacific.

🔗 CNBC
No image

#7Cloudflare and Stripe Launch Open Protocol for Autonomous AI Agent Deployment

Cloudflare and Stripe released a jointly designed protocol enabling AI agents to create accounts, start paid subscriptions, register domains, and deploy applications to production without human intervention. Stripe defaults to a $100/month spending cap per provider, and initial partners include Vercel, Supabase, and Sentry. This is the clearest signal yet that infrastructure companies are building for a world where AI agents are paying customers, not just tools.

🔗 Cloudflare Blog
No image

#8Chinese Court Rules Companies Cannot Fire Workers to Replace Them With AI

Courts in Hangzhou and Beijing ruled in two separate cases that AI adoption is a voluntary business strategy, not an unforeseeable change that justifies termination. In the lead case, a QA supervisor was fired after his role was automated by LLMs; the court found the employer illegally shifted the risks of technological change onto the worker and ordered compensation. This sets the first major legal precedent globally on AI-driven layoffs.

🔗 Bloomberg
No image

#9OpenAI Developing AI-First Smartphone With Qualcomm and MediaTek

OpenAI is building a smartphone designed around AI agents that replace traditional apps, with chips co-developed with Qualcomm and MediaTek and manufacturing by Luxshare. Specs are expected to be finalized by late 2026 or early 2027, with mass production in 2028. OpenAI's rationale: only by controlling both OS and hardware can it deliver a comprehensive AI agent service that captures the user's full real-time context.

🔗 TechCrunch
No image

#10Q1 2026 AI Venture Funding Hits $242B — 80% of All Global VC

Crunchbase data shows investors poured $300 billion into startups globally in Q1 2026, with $242 billion — a staggering 80% — flowing to AI companies. That is more than double all foundational AI funding in the entirety of 2025. Capital is still available but increasingly concentrated in companies that can prove speed, category control, and a path to market dominance.

🔗 Crunchbase News
No image
Kilroy’s Daily Briefings · 5/2/2026