The news: Sundar Pichai announced that 75% of Google's new code is now AI-generated, up from 50% last fall. The company also created a "strike team" to improve its coding models and catch up to Anthropic, whose Claude Code writes 70-90% of its code.

What actually happened:

Google is publicly admitting it's behind in the AI coding race. The 75% figure sounds impressive until you realize it's a defensive metric, not a leading one. Anthropic hit similar numbers months ago. Google isn't winning -- it's catching up.

The "strike team" memo from Sergey Brin reportedly demands that "every Gemini engineer must be forced to use internal agents for complex, multistep tasks." Forced adoption is not a sign of product-market fit. It's a sign of internal panic.

Why this matters:

  1. Google's AI coding moat is nonexistent. DeepMind has world-class researchers. Google Cloud has enterprise distribution. But Anthropic's Claude Code has developer love. Google is learning that building great models doesn't automatically mean developers will choose your tools.

  2. The 75% number is misleading. AI-generated code isn't the same as AI-shipped code. Google has massive internal infrastructure, legacy codebases, and review processes. Most of that 75% requires heavy human modification. The metric measures volume, not value.

  3. Enterprise vs. indie developer split is widening. Google's Workspace AI updates (Chrome Enterprise AI monitoring, Gmail AI Overviews for business, Meet AI notetaking for in-person meetings) show where its real focus is: selling to IT departments, not winning hacker hearts. Anthropic owns developers. Google owns procurement.

  4. The RAM shortage warning is the buried lede. Google quietly acknowledged that memory constraints could bottleneck AI development "for years." If the infrastructure layer can't keep up with model demand, the coding race becomes a supply chain war.

The bottom line:

Google isn't losing the AI coding war because its models are bad. It's losing because it treated developer tools as an afterthought while Anthropic built a cult. The 75% figure and strike team are damage control, not momentum. In the AI coding race, being big isn't the advantage it used to be.