Google I/O 2026 to Highlight AI Models, Cloud Growth, and XR Glasses

In addition to Gemini and AI search, Android XR smart glasses, TPU chips, Google Cloud, and AI-driven shopping experiences may also be major highlights at this year’s Google I/O. Meanwhile, the market is awaiting further commercialization signals.

Over the past year, Alphabet's (GOOGL.O) stock price has risen 140%, with its cloud business growing faster than Amazon's (AMZN.O) and Microsoft's (MSFT.O). Yet just 18 months ago, the market widely believed that, despite Google's years of preparation for AI, OpenAI had taken the lead in defining the era of generative AI.

Today, Wall Street’s view of Alphabet has changed significantly. Investors are beginning to see it as one of the few companies capable of generating profits across multiple layers, including AI models, chips, cloud computing, search, and enterprise software.

As the Google I/O developer conference opened locally on Tuesday, market focus shifted from "whether Google is catching up in the AI race" to "whether Google can truly commercialize AI."

Lo Toney, founding managing partner at Plexo Capital and an early investor in Anthropic, said Google’s greatest advantage lies in its near-complete control over the entire AI technology stack, including chips, cloud infrastructure, models, and end-user applications.

Gene Munster, managing partner at Deepwater Asset Management, believes the greatest value of a complete technology stack is speed. By developing its own chips, securing power resources, and building data center capabilities, Google can deploy AI products more quickly.

The most anticipated content at Google I/O was still the Gemini model.

The market had previously speculated that Gemini 4 might be released, but Citibank believes that, given Gemini 3.1 Pro was only launched in February, Google is more likely to release an upgraded version such as 3.2 or 3.5 rather than a true generational update.

However, this is not just a matter of version—it also concerns whether Google will return to the forefront of the AI industry.

Mizuho Securities noted that if Gemini 4 is officially launched, it would signify Google rejoining OpenAI and Anthropic at the same competitive level; if it’s only a minor update, it would more closely resemble continued catching up.

In addition to the model itself, investors are also paying attention to Project Astra, Gemini Live, multimodal video understanding, and the AI tool invocation capabilities within Search, Gmail, Calendar, and Maps.

Gemini's user growth has significantly accelerated. According to Citi data, Gemini's app saw a 127% year-over-year increase in monthly active users in the U.S. in April, and Gemini Enterprise's monthly active users grew by 40% compared to the previous quarter.

Another core theme of this year's I/O conference is AI agents.

Google aims to elevate Gemini from a chatbot to the "operating layer" of the entire Google ecosystem, capable of not only understanding user needs but also directly completing tasks.

The conference agenda extensively covered agent-based programming, multimodal workflows, and robotics, serving as Google's response to OpenAI's Codex and Anthropic's Claude Code.

Tony said that one of the biggest future market opportunities could be AI agents and enterprise Copilots.

Google is also beginning to deploy AI agents in commercial scenarios. The market expects this year’s I/O conference to highlight agent-based shopping workflows, where Gemini not only answers shopping-related questions but can also complete purchases directly.

Google has recently continued expanding the Universal Commerce Protocol, adding new partners including Meta (META.O), Microsoft (MSFT.O), Stripe, Klarna, and Affirm (AFRM.O).

Sameer Samat, President of Google’s Android ecosystem, gave an example where he asked Gemini to plan a barbecue, including generating a menu, opening Instacart, adding items to a Safeway cart, and completing the notification.

Meanwhile, the commercialization of AI search has become one of the market's biggest concerns.

Citi data shows that AI-enhanced ads currently account for over 30% of search advertising spend. AI Max will replace Dynamic Search Ads in September and demonstrates stronger conversion capabilities.

However, AI search also brings significant issues. Mizuho noted that AI Mode reduces user navigation to external websites, with its estimates showing that 93% of searches do not leave the Google page, and the natural click-through rate for AI Overview queries has dropped by 15%.

This means Google must prove that AI search can not only enhance the user experience but also sustain advertising revenue.

Compared to search, Wall Street is now paying more attention to Google Cloud.

In Q1, Alphabet's cloud business grew 63% year-over-year, surpassing Azure and AWS. The cloud backlog reached $462 billion, up approximately 90% quarter-over-quarter, with half of it expected to be recognized as revenue over the next 24 months. Revenue from generative AI products increased by approximately 800% year-over-year.

Alphabet CEO Sundar Pichai said that an increasing number of customers are expanding their partnerships with Google. The number of single contracts signed in 2025 exceeding $1 billion has already surpassed the total from the previous three years.

Another key variable is the TPU. Google has previously disclosed that it plans to begin selling its self-developed AI chips to external customers in the second half of 2026 and further scale up in 2027.

Mizuho noted that the market's current key concerns include how TPU revenue will be recognized, the profit margin levels, and whether it will be included in backlog orders.

Munster estimates that the global AI chip market is currently around $500 billion annually, and even a small market share for Alphabet could represent a significant new revenue source.

Before Google I/O, Google's relationship with Anthropic also became a market focus.

Alphabet holds a significant stake in Anthropic, and the recently reported $200 billion cloud services commitment, if true, could represent a substantial portion of Google's future cloud revenue.

Meanwhile, Google has also committed up to $40 billion in investment in Anthropic. This model has raised market concerns about customer concentration, similar to previous discussions sparked by Oracle's (ORCL.N) partnership with OpenAI.

However, Tony views this more as a “hedge.” He notes that even if enterprise clients increasingly prefer Claude over Gemini in the future, Google will still benefit from ongoing demand for its underlying cloud infrastructure and TPUs.

Android remains a key platform for Google to showcase its AI capabilities.

The first beta version of Android 17 was released in February, with the official version expected to launch this summer. The major changes currently focus on AI features and interaction optimizations.

Among the more popular new features is “App Bubbles,” which allows users to run apps in floating windows.

Google Books is another key focus. Google demonstrates that users can natively run Android apps from their phones directly on Google Books, without needing to operate the phone.

Google also unveiled the AI-powered Magic Pointer feature, which uses contextual understanding to enable photo editing and other tasks with a single prompt.

The Android XR smart glasses are also expected to be a key hardware product at I/O.

Google demonstrated an Android XR concept product last year. This year, outsiders expect it to move closer to official commercialization.

Compared to Google Glass from 2013, the new generation of Android XR glasses have a design that more closely resembles ordinary eyewear and support features such as real-time translation, notification display, and Gemini Live.

Currently, smart glasses developed by Meta Platforms in collaboration with Ray-Ban have entered the market, and Samsung Electronics' Galaxy XR headset is also available.

The market also expects Google to reveal more demos, hardware details, and release dates at this year's I/O conference.