Morning Digest, May 21, 2026

May 21, 2026 · 17 newsletters · 5 overlapping stories

Top Stories

Google I/O 2026: Gemini Goes Fully Agentic

(6 newsletters)

Google’s annual developer conference dominated coverage across nearly every newsletter. The company unveiled Gemini 3.5 Flash, a new model positioned for agentic workflows that approaches rivals like Opus 4.7 and GPT-5.5 at roughly 4x the speed and half the cost. Alongside it came Gemini Spark, a 24/7 personal agent running on Google Cloud VMs to handle tasks across Workspace, Chrome, and email; Gemini Omni, which generates cinematic video from any input type; and Antigravity 2.0, a standalone IDE built for agent-first development. Google framed the Search redesign as its biggest in a generation, adding cross-modal inputs and generative UI layouts. Monthly token usage across Google AI systems has reportedly reached 3.2 quadrillion.

SpaceX Files S-1 for Record-Setting IPO

(3 newsletters)

SpaceX filed its first comprehensive prospectus with the SEC, targeting a Nasdaq listing under ticker SPCX that could raise up to $80 billion at a $2 trillion valuation, potentially surpassing Saudi Aramco as the largest IPO in history. The filing showed a Q1 2026 net loss of $4.3 billion on $4.7 billion in revenue; for full year 2025, SpaceX lost $4.9 billion on $18.7 billion in revenue. Elon Musk holds 85% of voting control, and the listing would likely make him the world’s first trillionaire. Retail investors will be able to buy shares at debut through Robinhood, Schwab, Fidelity, E*Trade, and SoFi. Wedbush analyst Dan Ives predicts a SpaceX/Tesla merger in 2027.

Andrej Karpathy Joins Anthropic

(4 newsletters)

AI researcher Andrej Karpathy announced he is joining Anthropic’s pretraining team, where he will lead a new internal group focused on using Claude to automate Anthropic’s own AI training pipeline. Karpathy co-founded OpenAI in 2015, led Tesla’s Autopilot through 2022, briefly returned to OpenAI, then left in 2024 to focus on AI education. His announcement drew over 21 million views on X. He noted the move is research-focused and plans to return to education work “in time.” This follows a pattern of senior researchers from competing labs moving to Anthropic, which several newsletters noted as significant momentum for the company.

OpenAI IPO Filing Expected Imminently

(3 newsletters)

Multiple outlets reported that OpenAI is preparing to file a confidential IPO prospectus as soon as this Friday, working with Goldman Sachs and Morgan Stanley. The company aims to go public as early as September and is currently valued at $850+ billion by private investors. Concerns remain about whether it can generate enough revenue to support its spending commitments, and it has reportedly missed multiple internal revenue and user targets. The ruling in the Musk v. Altman lawsuit (see below) also cleared a significant legal obstacle ahead of the offering.

Meta Lays Off 8,000 Employees, Redirects 7,000 to AI

(3 newsletters)

Meta cut approximately 8,000 employees representing about 10% of its roughly 78,000-person workforce, framing the move as necessary to fund AI spending it has pledged at up to $145 billion this year. Alongside the cuts, 7,000 employees are being redirected into AI-focused teams and 6,000 open roles are being eliminated. Morale at the company is reportedly at an all-time low, with employees describing the atmosphere as “rancid.” The company recently introduced an internal program tracking employees’ every computer action to use as AI training data, which has prompted both internal petitions and unionization attempts among UK workers.

Also Worth Knowing

Quick Hits