← Back to newsletters

AI Daily Brief — Sat May 23

2026-05-23

By Vadym · Generated with AI, curated by me


Listen to this issue

The weekend is quiet. The numbers aren’t. Anthropic printed its first profit the same week Meta told 8,000 people their roles are being absorbed by AI. The institutionalization isn’t coming — it’s mid-sentence.


Headlines & News
Industry

Anthropic Projects First Operating Profit of $559M as Q2 Revenue Hits $10.9B

Anthropic shared internal investor projections showing $10.9 billion in Q2 2026 revenue — a 130% jump from Q1’s $4.8 billion — alongside a $559 million operating profit, the company’s first ever. The figures were disclosed ahead of a fundraising round, per The Wall Street Journal. Claude Code alone crossed $1 billion in annualized revenue within six months of launch, driven almost entirely by enterprise developers. Full-year profitability is not guaranteed due to planned increases in compute and model training spend.

Source: CNBC

Workforce

Meta Cuts 8,000 Jobs and Redirects 7,000 Workers Into AI-Focused Teams

Meta began notifying roughly 8,000 employees of layoffs — approximately 10% of its workforce — in a restructuring the company is framing explicitly around AI investment. At the same time, 7,000 workers are being redirected into newly created AI-focused units including Applied AI Engineering, Agent Transformation Accelerator XFN, and Central Analytics. Meta’s 2026 capex guidance runs $125–145 billion, more than double 2025’s outlay. Further cuts later in the year remain possible.

Source: NBC News

Policy

Trump Delays AI Pre-Release Review Order After Tech Exec Pushback

The White House postponed a planned executive order that would have established a voluntary framework giving federal agencies up to 90 days to review advanced AI models before public release. The signing had already been delayed multiple times; Trump said he “didn’t like certain aspects.” Musk, Zuckerberg, and venture capitalist David Sacks all opposed the measure — despite OpenAI reportedly backing it. No new signing date has been set.

Source: Bloomberg

Coding

ServiceNow Build Agent Goes GA Inside Cursor, Windsurf, Claude Code, and GitHub Copilot

ServiceNow made Build Agent generally available in ServiceNow Studio and extended core skills into Cursor, Windsurf, Claude Code, and GitHub Copilot. Developers can now orchestrate ServiceNow platform capabilities from any of these coding environments with full enterprise context and built-in governance — no tool switching required. The announcement turns every major coding IDE into a ServiceNow deployment surface, a significant distribution play for the enterprise workflow layer.

Source: ServiceNow Newsroom

Research

Anthropic Adds Self-Hosted Sandboxes and Private MCP Tunnels to Claude Managed Agents

Anthropic shipped two new deployment architecture features for Claude Managed Agents: self-hosted sandboxes (public beta) and MCP tunnels (research preview), unveiled at the Code with Claude developer conference in London. Self-hosted sandboxes let organizations run tool execution inside their own infrastructure — keeping sensitive files and services fully private — while Anthropic’s orchestration layer stays on-platform. MCP tunnels route agent connections to internal services over private networks without public internet exposure.

Source: InfoQ


Analysis

Takeaway

Three of today’s stories share an axis: capital and jobs are being restructured around AI, and no one is asking permission. Anthropic’s first profit is built on enterprise developers; Meta’s layoffs are funded by the same thesis. The White House failed — again — to establish even voluntary pre-release review, blocked by the executives who benefit most from moving fast. ServiceNow and Anthropic’s infrastructure plays are the quiet version of the same story: whoever controls the deployment architecture controls the enterprise. The news this weekend isn’t flashy. It’s structural.

— Boba


Curated by Vadym