AI Told These Ex-Google Guys How to Build a Billion-Dollar Weapons Business

Glass Epoch DeskOriginal digestSource: WIRED AI

AI-assisted digest, checked against the original source before publishing. How we work

Ex-Google and ex-Apple executives used an AI model to pick their weapon system — and now run a defense firm valued over $1 billion.
AI Told These Ex-Google Guys How to Build a Billion-Dollar Weapons Business
Photo: visuals · free stock license

Chad Steelberg, former CEO of enterprise AI company Veritone, and Andy Baynes, a former Apple executive and Google alumnus, have turned AI-assisted analysis into a defense business. Their firm, Tiberius Aerospace, announced $85 million in Series A financing this week, led by European investment firm Lakestar and national-security-focused Rochefort Asset Management, with Ukraine's UA1 also investing. The company says its defense-as-a-service business is valued at more than $1 billion.

The approach began in Steelberg's private research lab, incorporated in 2021 to explore ideas spanning drones, electric vertical takeoff aircraft, and quantum computing. After Russia's 2022 invasion — shortly after Steelberg had taken in Ukrainian refugees at his French chateau — he redirected the lab's AI and aerospace engineers to model the war's variables and identify what could shift the outcome. The model ruled out better defensive technology, arguing it would only harden the stalemate, and instead recommended longer-range artillery munitions over another drone.

The result is Sceptre, a round fired from a 155-millimeter howitzer, the most common artillery piece donated to Ukraine. Tiberius claims it travels farther than current rounds, reaches supersonic speeds, and flies beyond jamming range. The company announced a successful first test fire from a NATO-standard 155-mm howitzer in April. Sam Cranny-Evans of the Royal United Services Institute told WIRED that added range would let Ukrainian gunners operate beyond Russian drone attacks, calling any technology that returns howitzers to combat valuable.

Tiberius is now working with the US Army to mature the system and has announced a $500 million technology collaboration contract with UK partners to manufacture, license, and distribute the munition. A missile in development is expected to begin generating revenue next year. Baynes joined as chief strategy officer in 2023 after spending roughly six months in Budapest coordinating refugee evacuation and equipment transport into Ukraine; he previously bought a local life-science company in Kyiv in 2016 and has called the Western response to the invasion inadequate.

The internal AI engine has been commercialized as GRAIL (Generative Real-Time Artificial Intelligence for Lethality), which Steelberg says he is most excited about. The platform simulates weapons against different targets, automates supply chain management and compliance, and is intended to enable continuous design updates by decentralizing manufacturing. Whether GRAIL can deliver on that promise — and whether AI-driven design cycles translate into faster fielded capability — remains unproven at scale, as does the pace at which partnerships with the US Army and UK partners convert into deployed systems.

Key questions ahead center on that conversion: Sceptre's maturation timeline with the US Army, the missile's revenue target for next year, and whether decentralized manufacturing actually shortens upgrade cycles for licensed weapon designs.

Key facts

  • • Tiberius Aerospace raised $85 million in Series A financing, led by Lakestar and Rochefort Asset Management.
  • • The firm values its defense-as-a-service business at more than $1 billion.
  • • Sceptre is a 155-mm artillery round claiming supersonic speeds and range beyond jamming.
  • • Tiberius announced a $500 million UK technology collaboration to manufacture and distribute the munition.
  • • Tiberius fired Sceptre from a NATO-standard 155-mm howitzer for the first time in April.

Why it matters

Builders in dual-use and defense tech should note that AI-driven market and lethality analysis is now shaping product selection pre-build, not just design optimization. Expect more software-first entrants competing on iteration speed against traditional primes, with government partnerships and licensing deals as the near-term revenue path rather than product sales alone.

Spotted an error? Send a correction.

Get the day's AI news in one email

The Daily Epoch · 60-second read · unsubscribe anytime

Related

Agents
MIT Technology Review2 min read

Building AI that can work across the enterprise

SAP and Infosys executives argue that scaling AI agents forces enterprises to rebuild data foundations, governance, and operating models — not just add tools.