Google Delays Gemini 3.5 Pro Launch After Coding Falls Short, Wiping Out Nearly $200 Billion in Alphabet Value

Google's flagship AI model missed its promised June rollout after coding performance fell short internally, and Alphabet shares dropped 4.4% on the report, erasing about $200 billion in market value.

Verified pipeline

Sources: 3 Publisher: signed Contributor: signed Hash: 28c91c23ce View

Editor's Note · July 19, 2026

Correction:

The article quotes 9to5Google as writing, "Google has delayed the launch of Gemini 3.5 Pro beyond its originally promised June timeline." This exact sentence does not appear in the cited 9to5Google article. The underlying fact is accurate and corroborated by both 9to5Google ("That deadline has passed with no update on when to expect it") and PYMNTS ("Google is months behind schedule"), but it should not have been presented as a direct quotation from 9to5Google.

Correction:

The article quotes Investing.com as attributing part of the delay to Google's "complex organizational structure—requiring multiple stakeholder approvals and cross-product integration across Search, Maps, and YouTube." Investing.com's actual wording is: "Google's structure—with multiple layers of stakeholders involved in preparing models for release and working to integrate AI across a vast product portfolio including Search, Maps, and YouTube—can cause delays." The substance is accurate; the article's version is a paraphrase, not a verbatim quote.

Overview

Google has delayed the public launch of Gemini 3.5 Pro, its flagship AI model, months past the June rollout the company had promised, after internal testing found the model’s coding performance fell short of expectations, according to 9to5Google. The report sent Alphabet shares down 4.4% on Thursday, erasing about $200 billion in market capitalization, according to Investing.com.

What We Know

At Google’s I/O 2026 developer conference on May 19, the company said in a blog post: “We’re also hard at work on 3.5 Pro. It’s already being used internally, and we look forward to rolling it out next month,” according to PYMNTS. Machine Herald previously reported on that same I/O keynote, where Google teased that the full Gemini 3.5 Pro model was already in internal testing.

That promised June rollout never happened. “Google has delayed the launch of Gemini 3.5 Pro beyond its originally promised June timeline,” 9to5Google reported, citing Bloomberg. The company is “taking time to try to improve [Gemini 3.5 Pro’s] capabilities, particularly in coding,” and in late June it “updated the data being used to train Gemini in an attempt to improve [coding] skills, but the results were disappointing,” per the same report. Investing.com similarly reported: “Late last month, Google updated the data being used to train Gemini to improve these skills, but the results were disappointing.”

The coding shortfall comes as rivals close the gap: “both OpenAI and Meta Platforms Inc. recently released new models that outpace Google’s current offerings in AI for writing code,” according to Investing.com.

Internal friction has also played a role. PYMNTS reported that “the delay has been caused in part by Google’s many layers of stakeholders involved in preparing models for release, the company’s efforts to make the 3.5 Pro’s skills in writing code more competitive with its rivals, and competing factions within Google each building their own AI coding tools.” Investing.com separately attributed part of the delay to Google’s “complex organizational structure—requiring multiple stakeholder approvals and cross-product integration across Search, Maps, and YouTube.”

AI-written code is already widespread inside the company: as of April, “75% of all new code at Google is now AI-generated and approved by engineers, up from 50% last fall,” 9to5Google reported, though it noted some Google engineers reportedly prefer human-written code to maintain company standards.

Responding to the report, Google confirmed it is “currently testing 3.5 Pro, an upgraded Flash model, and other models with partners,” identical wording carried by both 9to5Google and PYMNTS. The company added that it is “shipping quickly across a wide range of models while keeping them highly cost-effective for customers,” and, per PYMNTS, is “productively engaged with the U.S. government on model testing and broader frameworks.”

Alphabet’s stock “closed down 4.4% on Thursday, erasing $200 billion in market capitalization,” Investing.com reported. The outlet noted that Alphabet’s quarterly earnings report was scheduled for the following week, where investors anticipated updates on the revised Gemini 3.5 Pro timeline and broader AI strategy.

What We Don’t Know

Google has not announced a new public launch date for Gemini 3.5 Pro. It also remains unclear how long the current “testing with partners” phase will run or which partners are involved. Neither report detailed how the delay might affect Alphabet’s broader AI product roadmap beyond the immediate market reaction, and any further comment is expected only at the company’s upcoming quarterly earnings call.

Analysis

The roughly $200 billion single-day paper loss underscores how tightly investors now tie Alphabet’s valuation to its position in the AI model race. Google had positioned the Flash-tier Gemini 3.5 model, launched at I/O 2026, as evidence that its next generation was ahead of schedule, with the more capable Pro variant framed as an imminent follow-on. The gap between that May promise and the missed June deadline — compounded by a failed retraining attempt in late June — has instead become a visible marker of how much ground Google believes it still needs to close against OpenAI and Meta on coding, one of the metrics enterprise AI buyers weigh most heavily when choosing a model provider.