Sundar Pichai walks onto the Shoreline Amphitheatre stage at 10 a.m. Pacific today, and the bar he is being graded against has nothing to do with a single product. It has to do with $190 billion. That is Google’s projected 2026 capital spending number, the largest in the company’s history, and the I/O keynote is the one annual moment when Wall Street and developers get to see whether all that AI infrastructure has actually translated into surface area consumers can touch.
The expected lineup is the most ambitious in years. Gemini Omni, a unified multimodal framework that processes video, images, audio, and text in one pass, is the headline. A new premium laptop class called “Googlebook” running an AI-native OS designed for on-device Gemini inference is the hardware swing. Android XR smart glasses, built in partnership with Samsung and XREAL, are the form-factor bet that Mark Zuckerberg has been winning for two years on the Meta Ray-Ban side. None of these are leaks anymore. They are commitments.
Gemini Omni: The Multimodal Catch-Up
OpenAI has been shipping native multimodal in GPT-4o and its successors since 2024. Anthropic’s Claude family is now the enterprise default for AI coding work and has just claimed the top spot on the CNBC Disruptor 50. Gemini, on paper, has been at parity for a year on benchmark scores. In practice, the developer perception has been that Google’s API was the most awkward to integrate and the slowest to ship new modalities into production.
Gemini Omni is the answer to that critique. According to Yahoo Tech’s pre-keynote breakdown, the framework promises a single endpoint that ingests video frames, audio waveforms, document images, and text simultaneously, and returns structured output across any of those modalities. Veo, Google’s text-to-video tool, now generates native audio cues with the picture, closing the gap with Runway’s Gen-4 and Sora. Lyria, the music arrangement model, is moving from preview to pro-tier general availability. If the demos hold up under live conditions, the developer story shifts overnight.
The Googlebook Bet: Killing ChromeOS To Save Hardware
The Googlebook reveal is the most strategically risky move on the deck. Google is essentially telling the OEM ecosystem, the same Samsung, HP, Lenovo, and Acer partners who have built ChromeOS into a roughly $5 billion education-and-business hardware franchise, that the future premium tier runs on a different operating system. The new OS is described as “AI-native,” meaning Gemini Intelligence is not a layer on top, it is the file system, the search, the notification routing, and the application runtime.
That is the same architectural argument Microsoft made for Copilot+ PCs in 2024 and then quietly walked back when developers refused to rebuild apps around the NPU. Google is now making a more aggressive version of that bet. The premium Googlebook is supposed to undercut MacBook Pro pricing while offering on-device Gemini inference that Apple cannot match until its own M5-class silicon ships later this year. If Pichai can show even one credible enterprise customer using it for daily knowledge work, Wall Street will give Google credit for a new hardware revenue line. If the demo stumbles, the analyst questions about ChromeOS cannibalization will define the rest of earnings season.
Android XR: The Glasses Fight Google Cannot Lose Again
Smart glasses are the form factor where Google has the longest scar tissue. Google Glass shipped in 2013, became a cultural punchline, and got shelved. Eleven years later, Meta and EssilorLuxottica turned Ray-Ban Stories into a roughly 2 million-unit category and reset consumer expectations for what AI eyewear should feel like. Apple’s Vision Pro went the other direction with a $3,500 mixed-reality headset that has yet to find a mainstream buyer.
Android XR is Google’s third attempt and it is structured very differently from the first two. Samsung is the manufacturing partner. XREAL handles the optics. Gemini Omni runs the assistant layer. The pricing target is reported to land between Ray-Ban Meta and Vision Pro, in the $599 to $899 range for the consumer SKU, with an enterprise-only model at higher tiers. The strategic stakes are enormous because the next computing platform fight is not phone-versus-tablet. It is glasses-versus-watch-versus-pendant, and the company that owns the developer ecosystem for ambient AI captures the most valuable real estate of the next decade.
Why Wall Street Is Watching The Capex, Not The Demos
Alphabet shares are roughly flat year to date despite five quarters of solid earnings beats. The market is telegraphing a simple concern: $190 billion in 2026 capex is real money. Microsoft, Meta, and Amazon are spending at similar levels, but they have clearer near-term monetization stories. Microsoft is selling Copilot seats to every Fortune 500 IT budget. Meta is using AI to lift ad targeting and short-form video engagement. Amazon is monetizing AWS Bedrock and selling Anthropic compute. Google’s story is more diffuse, and the I/O keynote is where that diffuseness is supposed to resolve into specific product lines with specific revenue trajectories.
We laid out the full preview last week in our piece on how I/O 2026 has to justify Google’s $190 billion AI spend. The shareholder math has only sharpened since then. Bond yields ripping to 19-year highs make every dollar of capex more expensive in present-value terms. Anthropic’s enterprise wins and OpenAI’s continued consumer dominance mean Gemini cannot afford to look like the safe corporate choice that ships features twelve months late. According to Tom’s Guide’s live keynote coverage, Google is expected to lean hard on agentic Gemini demonstrations, the kind of multi-step workflow automation that Anthropic has been winning on.
What Is Not On The Stage
The absences are almost as telling as the headliners. There will be no Pixel 11 preview, no new mid-range Pixel-A phone, and no foldable refresh. Google is consolidating its consumer-attention budget around AI surfaces and giving up the iterative phone news cycle to Apple’s June WWDC. That is a clear strategic concession. The Pixel hardware business has never scaled past niche, and Google is using I/O to argue that the platform itself, Android plus Gemini plus XR, is the durable franchise, not any individual handset.
The Real Test
I/O has historically been graded on the polish of the demo and the loudness of the developer-room reaction. Today, with $190 billion at stake, it gets graded on a different scale. Three questions matter. Does Gemini Omni actually ship to enterprise customers inside 90 days? Does the Googlebook attract at least one Fortune 100 corporate IT commitment before year-end? Do the Android XR glasses reach retail by November, before the holiday window and before Meta’s next Ray-Ban refresh?
If the answer to all three is yes, Alphabet has its first cohesive AI consumer story in the post-ChatGPT era, and the capex narrative reframes from “Google spending blindly” to “Google spending early on the next platform shift.” If the answer to any of them is no, the I/O 2026 keynote becomes the moment Wall Street pulled the AI premium out of GOOG. That is the bar. The demos start in less than an hour.