Gemini 3.5 Pro delayed months; Alphabet shares slip 15% from peak

Alphabet faced an unexpected setback in its flagship AI roadmap when Gemini 3.5 Pro, promised for June, remained stuck in testing months past the deadline. The company had announced Gemini 3.5 Flash at its I/O developer conference in mid-May and said the more powerful Pro model would arrive a month later. Instead, Bloomberg reported on Thursday that Google had paused the launch to retrain the model and strengthen its coding capabilities—a critical feature for enterprise customers—but the results fell short of internal expectations. The company has not disclosed a new launch date, only saying it is "currently testing 3.5 Pro, an upgraded Flash model, and other models with partners."

The delay underscores a competitive pressure in AI development. Coding performance is arguably the main battleground for AI labs right now and a core service customers pay for. While OpenAI and Anthropic continue shipping new models, some inside Alphabet worry their rivals have surpassed Gemini in capability. The stakes are heightened by Alphabet's massive infrastructure investment: the company expects capital expenditures of as much as $190 billion（約30兆円） this year, much of it directed toward the hardware and systems behind its AI push. That spending assumes Gemini stays competitive at the frontier. A one-month slip might be manageable noise, but a pattern of delays could undermine confidence in the strategy.

Alphabet's stock price has already felt pressure, sliding from a peak of $408.61 earlier in the year to $346.77 on Friday—a decline of about 15%. Yet the company's business fundamentals remain robust, at least in the near term. In the first quarter, Alphabet's revenue climbed 22% year over year to $109.9 billion（約18兆円）, marking 11 consecutive quarters of double-digit growth. Google Cloud, the division driving AI adoption, delivered an acceleration: revenue jumped 63% year over year to $20.0 billion（約3.2兆円）, up from 48% growth in Q4 2025 and 34% in Q3 2025. The cloud segment's operating income roughly tripled year over year to $6.6 billion（約1.1兆円）. CEO Sundar Pichai reported that the company's cloud backlog nearly doubled from the prior quarter to over $460 billion（約74兆円） and that Gemini models were processing over 16 billion tokens per minute through direct customer use, up 60% in three months. Google Search & other revenue rose 19%, with management noting that search queries hit an all-time high. These metrics suggest customers have not paused their AI spending while waiting for Gemini 3.5 Pro.

The investment thesis for Alphabet therefore remains intact, at least for now. The company trades at about 25 times forward earnings, a multiple that assumes growth rates stay robust. If Gemini were to fall a full generation behind competitors, the AI demand underpinning that cloud backlog could become harder to defend, and the new AI-powered search features now driving usage could start to lag. Investors will get fresh evidence on Wednesday, July 22, when Alphabet reports second-quarter results. Analysts will scrutinize Google Cloud's growth rate and any timing guidance management provides for Gemini 3.5 Pro to determine whether this delay is a one-off setback or a sign of deeper execution risk.

Alphabet introduced Gemini 3.5 Flash in May at its I/O developer conference and promised the more powerful Pro version would follow in June. Instead, the company quietly pushed the model into extended testing, a notable slip in what was supposed to be a rapid release cadence following Gemini 3's launch in late 2025. Bloomberg's reporting reveals the snag: Google retrained the model to strengthen its coding abilities—a critical differentiator in enterprise AI—but the effort did not meet internal standards.

The delay carries real stakes. Alphabet is investing as much as $190 billion（約30兆円） in AI infrastructure this year, a capital commitment that depends on Gemini remaining competitive at the frontier of AI capability. Rivals are not waiting: some insiders at Google fear OpenAI and Anthropic have already leapfrogged Gemini on the models that matter most to paying customers. A single month's slip might be noise, but a pattern of delays would call into question whether Alphabet's massive infrastructure bet is justified.

Yet the company's Q1 financial results offer counterweight. Google Cloud revenue jumped 63% year over year to $20.0 billion（約3.2兆円）, the company's cloud backlog nearly doubled to over $460 billion（約74兆円）, and management reported search queries hit an all-time high. CEO Sundar Pichai said Gemini models were processing over 16 billion tokens per minute through direct customer use, up 60% in three months. These numbers suggest customers are not holding off contracts waiting for Gemini 3.5 Pro; demand for the company's current AI offerings remains robust. Alphabet trades at about 25 times forward earnings, a multiple that assumes continued growth. If Gemini were to fall a full generation behind, that growth story could erode. For now, however, one late model does not put Alphabet in that position. The risk remains worth monitoring closely.

No comments yet. Be the first to share your thoughts!