A rumor is circulating that Google plans to unveil its next major AI model, Gemini 3.5, at its annual developer conference, Google I/O, in May 2026.
While Google hasn't confirmed this, several strong signals make this rumor highly plausible. The evidence points to a clear strategy of accelerated development and a need to showcase significant progress.
First, Google has recently picked up the pace of its AI releases. They launched Gemini 3 in November 2025 and followed up with a preview of the more powerful Gemini 3.1 Pro in February 2026. This rapid, iterative cycle strongly suggests that a mid-year update, logically named '3.5', is next in line. The annual I/O conference has historically been Google's biggest stage for AI announcements, making it the perfect venue for such a reveal.
Second, the financial commitment is undeniable. Alphabet, Google's parent company, announced a massive capital expenditure (CapEx) plan of $175-185 billion for 2026, largely to build out its AI infrastructure. This is supported by Google Cloud's impressive 48% year-over-year growth and a $240 billion backlog, indicating strong customer demand for AI services. Furthermore, reports suggest Google is ramping up its own TPU chip production to compete with Nvidia, ensuring they have the computing power needed for the next generation of models.
Finally, the competitive landscape plays a crucial role. With rivals like OpenAI consistently releasing new models and raising the bar for AI reasoning capabilities, Google is under constant pressure to demonstrate its own advancements. I/O provides the ideal platform to respond to the competition and reassure investors and developers that it remains at the forefront of AI innovation. Although the 'Gemini 3.5' name is just speculation for now, the convergence of a clear timeline, massive investment, and competitive necessity makes a major AI announcement at I/O 2026 seem very likely.
- CapEx: Short for Capital Expenditure, which are funds used by a company to acquire, upgrade, and maintain physical assets like data centers, servers, and buildings.
- TPU: Tensor Processing Unit. A custom-designed computer chip developed by Google specifically for accelerating AI and machine learning tasks.
- Inference Cost: The cost associated with running a trained AI model to make predictions or generate outputs. Lower inference cost is crucial for deploying AI services at scale.