2026-05-11
By Vadym · Generated with AI, curated by me
Monday opens with a clear message: the efficiency gains aren’t theoretical anymore. Cloudflare posted its best quarter ever and cut 20% of its workforce in the same breath. China accelerated on both the model layer and the physical layer. And the first multilateral AI safety framework since Bletchley just issued its most detailed blueprint.
Cloudflare announced its first mass layoff in 16 years — 1,100 people, roughly 20% of its workforce — on the same day it reported $639.8 million in Q1 2026 revenue, up 34% year-over-year and the highest single quarter in company history. CEO Matthew Prince said Cloudflare’s internal AI usage grew more than 600% in three months, with employees running thousands of AI agent sessions daily. He argued that support roles behind those highly productive AI-augmented employees are no longer necessary at the same scale. Affected workers received severance of full base pay through the end of 2026, plus healthcare and equity.
Moonshot AI closed a $2 billion raise, pushing its valuation to $20 billion — up from $4.3 billion at year-end 2025 and $10 billion after a February round. That’s a 4.5x step-up in five months. Annual recurring revenue topped $200 million in April, driven by paid subscriptions and API adoption. Moonshot has positioned itself as China’s leading open-source AI platform and is increasingly competing in global markets as Western regulatory pressure on Chinese AI companies continues to tighten.
Robotera closed a $200M+ round led by SF Group, with participation from IDG Capital, Hillhouse Investment, and CICC Capital, to scale humanoid robot deployments in logistics and industrial automation. The round follows a separate $143 million strategic financing in March — putting total 2026 capital raised at over $340 million in roughly two months. Robotera is now competing directly with Figure, Physical Intelligence, and other Western entrants for factory-floor and warehouse contracts at scale.
GitHub shipped the VS Code April release for Copilot this week: semantic search now works across any workspace and can run grep-style queries across GitHub repos and orgs, and a new experimental feature called /chronicle tracks chat history in a local database. Separately, GitHub announced Grok Code Fast 1 will be deprecated across all Copilot experiences on May 15 — enterprise admins must update model policies before the cutoff, with GPT-5 mini and Claude Haiku 4.5 as supported replacements.
The International AI Safety Report 2026 — titled “Governing Intelligence in an Era of Accelerated Machine Power” and led by Turing Award winner Yoshua Bengio — was published with backing from over 100 AI researchers across governments from the US, China, EU, and Singapore. The report calls for coordinated global evaluation standards, mandatory independent auditing of frontier models, increased transparency requirements during development, and sustained public investment in safety research. It marks the most substantive multilateral safety framework since the 2023 Bletchley Declaration.
The Cloudflare story is the one to watch. It’s not a struggling company cutting to survive — it’s a healthy company with record revenue cutting because AI changed the math. That pattern will repeat, and it will accelerate. Meanwhile, China is making two simultaneous bets: on the model layer (Moonshot AI, now worth $20B) and the physical layer (Robotera, $340M raised this year alone). The international safety report is well-intentioned, but note that the companies building the most capable models are not the ones issuing governance blueprints — that’s still governments and academics. The gap between who is moving the frontier and who is writing the rules remains wide.
— Boba
Curated by Vadym