0>1.software

// digest

Daily Digest

The stories in AI and software that matter, once each weekday, with a take. Read it or listen to the daily audio edition.

August 4, 2026

Daily Digest: August 4, 2026

White House to Host OpenAI, Anthropic, Google to Review New Voluntary AI Safety-Testing Framework

On August 4, the Trump administration hosts OpenAI, Anthropic, and Google at the White House to unveil a finalized voluntary AI safety-testing framework stemming from a June executive order. The framework lets the government access frontier models for up to 30 days before wider release, and stops short of a mandatory licensing regime.

Voluntary pre-release access sounds substantial until you ask what the government actually does with those 30 days when there's no enforcement mechanism behind it.

Valar Atomics Raises $1B Series B Led by Sequoia to Mass-Produce Nuclear Reactors for AI Data Centers

Valar Atomics raised $1B in Series B equity led by Sequoia, plus a separate $200M credit facility, at a $6B valuation to scale factory production of small helium-cooled reactors. The raise follows the company powering an Nvidia AI chip with fission and announcing a 30MW nuclear-powered AI facility with Nvidia in Utah.

Power is turning into the real constraint on AI scaling, and investors are now pricing reactor manufacturing like it's a compute company.

Shield AI's Hivemind Completes First Water-Based Autonomous Swarm Test in Taiwan

Shield AI and Thunder Tiger ran Hivemind's first water-based autonomous swarm test in Pingtung, Taiwan, coordinating two SeaShark unmanned surface vessels. Using radar and AIS data, the vessels autonomously detected and jointly escorted a simulated intruder out of a search area, marking the software's first multi-asset maritime teaming test.

Running this specific test in Taiwan waters is the message as much as the tech. Autonomous vessel teaming against an intruder isn't a hypothetical scenario there.

Horizon3 Raises $250M Series E at $2B+ Valuation to Lead the 'AI vs. AI' Cybersecurity Era

Autonomous-penetration-testing company Horizon3 closed a $250M Series E co-led by NightDragon and NEA, more than tripling its valuation to over $2B. The company plans to expand its NodeZero AI-hacker platform into federal and international markets.

A pentesting startup tripling its valuation on an 'AI vs AI' pitch shows how fast attacker-side automation is becoming a budget line defenders have to match.

OpenAI, Anthropic, Meta, and Thinking Machines Locked in Escalating AI Talent War

Axios reports escalating competition among OpenAI, Anthropic, Meta, and Thinking Machines for top AI researchers. Recent moves include Google's Noam Shazeer joining OpenAI and Nobel laureate John Jumper joining Anthropic.

When a Nobel laureate is a transfer target, scientific prestige has become a recruiting asset alongside engineering talent.

Pentagon inks $3B framework agreement for Patriot, THAAD components

The Pentagon signed a $3B framework agreement covering Patriot and THAAD components. The Patriot solid rocket motor deal with Northrop Grumman is valued at $2B, and the THAAD component deal at $1B.

This is a framework agreement locking in Patriot and THAAD component supply from Northrop Grumman in one package, splitting three billion dollars roughly two to one between the two programs.

// subscribe

Get the digest in your inbox

One email each weekday. The stories that matter, with a take. No spam, unsubscribe anytime.

// archive