Google just missed its own deadline for Gemini 3.5 Pro. Four senior researchers left in the same week. Your Q3 model evaluation just broke.

Google promised Gemini 3.5 Pro would ship "next month" at I/O in May.

That month passed. The model didn't.

As of June 27, Gemini 3.5 Pro remains locked in a limited enterprise preview on Vertex AI. Google has confirmed to partners that general availability will not come before July.

This is the second major AI delivery miss Google has recorded in 2026.

But the delay isn't the story. The timing is.

Between June 21 and 27, four senior Gemini researchers announced departures. Noam Shazeer — VP of Engineering, co-author of "Attention Is All You Need," the paper that invented the transformer — is joining OpenAI. John Jumper — Nobel laureate for AlphaFold, the work that proved AI could solve real science — is joining Anthropic. Two additional senior Gemini architects followed them out the door.

Alphabet's market cap dropped $269 billion in the same week.

Here's what this means for your infrastructure plan: If you're a CIO deciding what model stack to standardize on for H2 2026, July is not abstract. Procurement calendars are real. Security reviews take weeks. Developers build habits around whatever tool is available when the budget clears.

OpenAI shipped GPT-5.6 Sol on June 26. Anthropic is gaining ground on coding. Google just told you they can't ship on time — twice in one year.

Audit your model evaluation timeline today. If your Q3 plan depends on Gemini 3.5 Pro being available in June, that plan is already broken. Build your workflow on a model that ships today. Upgrade when something better actually arrives.

The next great model is always one month away. Your automation shouldn't have to wait for it.

SOURCE: https://www.businessinsider.com/google-3-5-pro-july-release-tokens-ai-agents-model-2026-6
VERIFIED: Business Insider (June 24), FAQ.com (June 28), Startup Fortune (June 25), Crypto Briefing (June 24)
SIGNAL: Google's second delivery miss in 2026, combined with a talent exodus to OpenAI and Anthropic, directly impacts enterprise model evaluation timelines for H2 2026. CIOs standardizing on a model stack now face a credibility gap they cannot ignore.