Alphabet drops as delayed Gemini model raises AI worries
Alphabet shares slid more than 4% after a report said Google’s Gemini 3.5 Pro is running months late while rivals roll out new AI models.
By Sal Moretti · Money Reporter
3 min read
Alphabet took a Wall Street hit Thursday after a Bloomberg report said Google’s next top-tier AI model, Gemini 3.5 Pro, is months behind schedule.
Shares of Google’s parent fell more than 4%, with GOOG down 4.43% and GOOGL off 4.44%, according to MarketWatch. The drop came as investors weighed whether Google is losing momentum in the fast-moving market for artificial intelligence tools.
Bloomberg reported that Gemini 3.5 Pro, described as Google’s strongest AI model so far, has been delayed while the company works to improve its performance. Alphabet Chief Executive Sundar Pichai said in an early June investor presentation that the model was expected later that month, but it has not been released.
The delay has caused frustration among Google engineers and AI researchers, Bloomberg reported. Some staff are worried Google could give up ground to competitors including Anthropic and OpenAI, according to that report.
Bloomberg also reported that Google’s task is complicated by the need to place AI across a wide set of products. In June, the company updated Gemini’s training data to improve coding performance, but the results have fallen short of expectations, according to Bloomberg.
Google pushed back with a statement focused on its release pace and testing pipeline.
“We’re shipping quickly across a wide range of models while keeping them highly cost-effective for customers,” a Google spokesperson said, according to MarketWatch. “We’re currently testing 3.5 Pro, an upgraded Flash model, and other models with partners, and we’re productively engaged with the U.S. government.”
The stock move added pressure to a wider tech selloff. MarketWatch reported that concerns over AI spending have weighed on technology shares in recent weeks as investors look for clearer returns on heavy investment. Bryan Wong, a portfolio manager at Osterweis Capital Management, told MarketWatch that investors are questioning how durable the business model is when the leading AI model can change every few months.
Google’s rivals have been busy. Anthropic temporarily shut off access to its Fable 5 and Mythos 5 models last month after the U.S. government flagged them as a security threat, MarketWatch reported. OpenAI limited access in June to its GPT-5.6 models, named Sol, Terra and Luna, to trusted partners, and made them available for wider use last week, according to the company.
Meta Platforms released Muse Spark 1.1 last week, which MarketWatch described as its first model touting near-frontier agentic coding abilities. SpaceX launched Grok 4.5, which the company said was specifically trained for coding and agents. Meta shares fell 2.46% and SpaceX shares fell 3.08%, according to MarketWatch ticker data.
Another fresh challenge came from Beijing-based Moonshot AI, which published Kimi-K3 on Thursday and said it was built for “long-horizon agentic coding.” Arena.ai, which compares AI model performance, said Kimi-K3 has beaten Anthropic’s Fable 5 in some areas.
Arena CEO Anastasios Nikolas Angelopoulos wrote on X that the release “may be the single biggest release of the year” and said it marked a point where open-source Chinese models had surpassed U.S. models.
This story draws on original reporting from MarketWatch.