ANALYSIS

Google DeepMind Dev Ports Command & Conquer to iOS With Claude Code

M Marcus Rivera Jul 6, 2026 3 min read
Engine Score 7/10 — Important

tier-1 analysis

Editorial illustration for: Google DeepMind Dev Ports Command & Conquer to iOS With Claude Code
  • Ammaar Reshi, Lead Product and Design for Google AI Studio, ported Command & Conquer: Generals Zero Hour (2003) to iPhone and iPad using Anthropic’s Claude Code with Fable 5.
  • The game runs natively on ARM64 with no emulator; the graphics pipeline translates DirectX 8 to Apple’s Metal API through several intermediate steps.
  • The first build took about 40 minutes, followed by “a few hours” of debugging; Reshi burned his entire Claude Max quota over two days.
  • The full source code is on GitHub as open source — game assets not included, so a Steam copy (about $5) is required.

What Happened

A Google DeepMind developer ported the 2003 real-time strategy game Command & Conquer: Generals Zero Hour to iPhone and iPad using a competitor’s tools, as reported by The Decoder on July 5, 2026. Ammaar Reshi, Lead Product and Design for Google AI Studio, used Anthropic’s Claude Code with the Fable 5 model; the game runs natively on ARM64 with no emulator, and campaign, skirmish, and “Generals Challenge” modes all work with touch controls.

Why It Matters

Porting a two-decade-old DirectX PC game to a modern mobile platform is the kind of project that historically consumed engineering months; Reshi reports the first working build took about 40 minutes, followed by “a few hours” of debugging. It’s a concrete, verifiable demonstration of what current agentic coding tools can do on real legacy-code problems — with the full source published openly on GitHub so others can check the work.

The cross-company angle drew attention too. Asked why a Google AI Studio lead used Anthropic’s tools, Reshi replied that “you can love the AI space and respect the competition while still being fully focused on building the best answer. It’s a long game.”

Technical Details

The port’s graphics pipeline translates DirectX 8 calls to Apple’s Metal API through several intermediate steps, and the binary runs natively on ARM64 rather than under emulation. Reshi documented every bug and fix in an engineering log published alongside the code. The effort wasn’t free of cost or rough edges: over two days he burned through his entire Claude Max quota, and on iPads the game can crash during long sessions due to high memory usage. Game assets are not included in the repository, so trying it requires a personal copy of the game, available on Steam for about $5.

Who’s Affected

For developers, the project is a reproducible reference case for AI-assisted legacy porting — source, engineering log, and all. For the game’s community, it’s a playable native iOS version of a 2003 classic, contingent on owning the original assets. For Anthropic, a senior Google AI figure publicly choosing Claude Code and Fable 5 for a showcase project is a notable endorsement in the intensifying agentic-coding competition among Anthropic, OpenAI, and Google itself.

What’s Next

The code is open source on GitHub for anyone to build on, and the documented iPad memory crashes mark the obvious next fix for contributors. The stated caveats hold: this is one developer’s account of timing and effort, and the “few hours” figure covers the debugging pass, not a production-hardened release. As a benchmark of what agentic coding can achieve on legacy software in mid-2026, it sets a public, checkable bar.

Share

Enjoyed this story?

Get articles like this delivered daily. The Engine Room — free AI intelligence newsletter.

Join 500+ AI professionals · No spam · Unsubscribe anytime