Google is rebuilding its entire product suite around AI as the foundation.
At its annual developer conference in Mountain View, Google marked a decade of AI-first ambitions by unveiling a suite of autonomous agents, a reimagined search experience, and consumer hardware that places artificial intelligence not as a feature but as the foundation of daily life. The company's 'agentic Gemini era' is less a product announcement than a philosophical repositioning: AI that acts on your behalf, continuously, whether your device is on or not. With 900 million Gemini users, infrastructure spending nearly doubled, and a premium subscription repriced to challenge OpenAI directly, Google is signaling that the era of asking machines questions is giving way to the era of delegating decisions to them.
- Google is no longer adding AI to its products — it is rebuilding Search, Gmail, Chrome, and even eyewear from the ground up with AI as the structural core.
- Gemini Spark, a personal agent that works autonomously in the background across your calendar, inbox, and documents, represents the sharpest edge of that ambition — and the sharpest question about user trust.
- The price of Google's premium AI tier dropped from $250 to $100 a month, a direct strike at OpenAI's market position that reframes the subscription AI wars as a race to the floor.
- Smart glasses shipping this fall through Samsung and Warby Parker put Google back in a hardware category it once abandoned, now chasing the cultural foothold Meta's Ray-Ban glasses have quietly built.
- The real test is not whether these products impress in demos — it is whether autonomous agents actually change daily behavior, or quietly become the most sophisticated tools most people never use.
Google used its I/O 2026 keynote at the Shoreline Amphitheatre to declare the arrival of its 'agentic Gemini era' — a phrase that carried real weight given the products behind it. CEO Sundar Pichai framed the moment as the culmination of a decade-long commitment to AI-first thinking, noting that Gemini now serves 900 million monthly users, more than double the figure from a year ago. Alphabet is directing between $175 and $185 billion in capital expenditure toward AI infrastructure in 2026, nearly double the prior year. DeepMind's Demis Hassabis added that artificial general intelligence is 'just a few years away,' a claim that reflects the company's internal confidence even if the broader research community remains skeptical.
The headline product is Gemini Spark, a personal AI agent that runs continuously on Google's servers and keeps working even when your device is off. In its on-stage demo, Spark planned a neighborhood block party by pulling RSVPs from Gmail, building a tracking spreadsheet, and drafting follow-up emails — all without manual input. High-stakes actions still require user approval, but the labor happens in the background. Spark rolls out to AI Ultra subscribers in the US next week, with a Chrome version arriving later this summer and a live activity display called Android Halo coming to Android phones.
Google Search is undergoing its most significant redesign in decades. AI Mode, already at one billion monthly users, is merging with traditional results into a single unified experience. The search box now accepts images and video natively, and this summer AI agents will handle complex multi-step queries in the background, surfacing answers over time rather than instantly. Google's own framing was unambiguous: 'Google Search is AI Search.'
Elsewhere, Google announced Universal Cart for AI-powered shopping, Gemini Omni for multimodal content creation, and Antigravity 2.0 — a standalone coding agent desktop app that is twelve times faster than its predecessor and capable of managing multiple autonomous coding tasks simultaneously. A live demo showed it building a basic operating system and running Doom on it for under $1,000 in compute.
On hardware, Google confirmed its first Android XR audio glasses will ship this fall through Samsung, Warby Parker, and Gentle Monster, offering both audio-only and in-lens display variants compatible with Android and iPhone. Pricing also shifted: Google AI Ultra dropped from $250 to $100 per month, bringing it in line with OpenAI's ChatGPT Pro and bundling 20 terabytes of storage, YouTube Premium, and early Spark access.
What I/O 2026 ultimately revealed is a company that has stopped treating AI as an enhancement and started treating it as the architecture. Whether that transformation takes hold in people's actual lives — or remains a series of remarkable demonstrations — is the question that will define Google's next chapter.
Google spent May 19 declaring itself officially in its "agentic Gemini era," and the company backed the claim with real products. At I/O 2026, held at the Shoreline Amphitheatre in Mountain View, the search giant unveiled autonomous AI agents, a redesigned search experience, smart glasses shipping this fall, and a price cut on its premium AI subscription that signals a direct challenge to OpenAI's market position.
CEO Sundar Pichai opened the keynote by marking a milestone: a decade since Google committed to becoming an AI-first company. The shift he described isn't incremental. It's about AI that takes action on your behalf, not just answers questions when you ask. The numbers he shared suggested the bet is paying off. Gemini now has 900 million monthly active users—more than double the 400 million from a year prior. Google is processing 19 billion AI tokens per minute across its products. And Alphabet is spending between $175 billion and $185 billion on capital expenditure in 2026, nearly double the prior year's $91.4 billion, almost entirely directed at AI infrastructure. DeepMind CEO Demis Hassabis added that artificial general intelligence—AI capable of performing any intellectual task a human can—is "just a few years away," a claim that signals how seriously the company's leadership views the current moment, even if the broader research community would push back on the timeline.
The centerpiece product is Gemini Spark, a personal agent that runs continuously on Google Cloud servers and keeps working even when your device is off. Unlike a chatbot you summon, Spark acts autonomously. In the on-stage demo, it planned a neighborhood block party by pulling RSVPs from Gmail, building a tracking sheet in Google Sheets, and drafting follow-up emails to neighbors who hadn't responded—all without a single button press. High-stakes actions like sending emails or spending money still require your approval, but the work happens in the background. Spark rolls out to AI Ultra subscribers in the US next week, with a Chrome version arriving later this summer. On Android, a new interface called Android Halo will display live agent activity at the top of your screen.
Google Search is undergoing its most significant overhaul in decades. AI Mode, which has already crossed 1 billion monthly users, is merging with traditional search results into a single unified experience. You can now move between conventional links, AI-generated answers, and follow-up conversations without losing context. The search box now accepts images and video natively and uses AI to suggest queries as you type. This summer, AI agents are coming to Search itself—they'll handle complex, multi-step questions in the background, with results surfacing over time rather than all at once. Google's messaging was direct: "Google Search is AI Search."
Beyond search and agents, Google announced Universal Cart, an AI shopping assistant built into Search and the Gemini app that surfaces price drops, deal alerts, and product compatibility warnings automatically. Gemini Omni, a new model family built for creation rather than conversation, accepts image, audio, video, and text as input and outputs video grounded in real-world knowledge. Antigravity 2.0, Google's answer to Claude Code, is now a standalone desktop app available globally today—it's 12 times faster than the previous version and designed to manage multiple autonomous coding agents at once. The live demo showed it building a basic operating system from scratch and running Doom on that OS using less than $1,000 in AI compute.
Google also confirmed its first Android XR audio glasses will arrive this fall, available in two tiers: audio-only with cameras, microphones, and speakers, or an optional in-lens display variant. Partners include Samsung, Warby Parker, and Gentle Monster. Both Android and iPhone users can connect by saying "Hey Google" or tapping the frame to access Gemini for navigation, real-time translation, and message management. This marks Google's first major move into smart glasses, a category Meta's Ray-Ban glasses have dominated.
Pricing shifted significantly. Google AI Ultra now costs $100 per month, down from $250—a calculated move that brings it in line with OpenAI's ChatGPT Pro and undercuts the competition. The plan includes 5 times the usage limits of the $20 AI Pro tier, 20 terabytes of storage, YouTube Premium, and early access to Gemini Spark. The old $250 tier drops to $200 with the same features.
What emerges from I/O 2026 is a company no longer betting on AI as a feature bolted onto existing products. Google is rebuilding its entire suite—Search, Gmail, YouTube, Docs, Chrome, Workspace, your phone, and now your glasses—around AI as the foundation. Whether that's a bold evolution or an overreach will depend on what happens next: whether Spark, Omni, and the glasses actually change how people work and live, or whether they remain impressive demonstrations that most users never meaningfully engage with.
Notable Quotes
We are firmly in our agentic Gemini era.— Sundar Pichai, CEO of Google
Google Search is AI Search.— Google, on the redesigned search experience
The Hearth Conversation Another angle on the story
Why does Google need a 24/7 agent running on servers somewhere? Isn't that just a more complicated way to do what I already ask my phone to do?
The difference is that Spark doesn't wait for you to ask. It watches your inbox, sees that you're planning something, and starts doing the work in the background. You don't have to remember to ask it to draft emails or build a spreadsheet. It just happens.
That sounds invasive. How do you stop it from doing something you don't want?
For routine tasks—organizing information, drafting messages—it just does it. But anything that costs money or sends something out into the world requires your approval first. It's a middle ground between full autonomy and needing to click a button for everything.
The smart glasses seem like the riskier bet. Meta's been selling Ray-Bans for years. Why would Google's version be different?
Google's glasses are arriving with Gemini built in from day one. You can translate languages in real time, get navigation without looking at your phone, manage messages by voice. It's not just a camera on your face—it's an interface to an AI that knows what you're doing.
But will people actually wear them? Or is this another Google Glass moment?
That's the real question. The technology is better now, the use cases are clearer, and they're coming from a company with actual distribution. But "better" doesn't guarantee adoption. It depends on whether people feel like they need an AI assistant on their face all day.
The price cut to $100 for Ultra feels aggressive. Is Google panicking about OpenAI?
Not panicking—competing. Google has 900 million Gemini users already. They're not trying to convince people to switch from OpenAI. They're trying to make it obvious that their top tier is worth the same price, and it comes with YouTube Premium and 20 terabytes of storage. It's a move to make the comparison easy.