Google slashed AI subscription prices, launched Gemini 3.5, unveiled autonomous search agents, a new video model, and audio smart glasses at I/O 2026, signalling a structural shift in enterprise AI competitive dynamics.

Key Highlights

  • Google launched Gemini 3.5 Flash and cut its AI Ultra subscription to $200 per month, adding a new $100 developer tier.
  • Autonomous AI agents embedded in Search can now execute purchases, monitor ticket availability, and manage schedules in real time.
  • Gemini Omni, a new multimodal video model, positions Google in the emerging world-model segment of generative AI.
  • Audio smart glasses developed with Samsung, Warby Parker, and Gentle Monster are set for an autumn 2026 launch.
  • Gemini's monthly active user base has more than doubled year-on-year to 900 million, with AI Overviews reaching 2.5 billion users.

A Strategic Pivot, Not Just a Product Cycle

Alphabet (Nasdaq: GOOGL) used its annual I/O developer conference in Mountain View, California to make a pointed argument to enterprise customers: competitive AI capability does not have to come at a premium price. CEO Sundar Pichai claimed Google's models can deliver performance comparable to frontier rivals at up to one-third of the cost, and that large organisations could save more than one billion dollars annually by migrating to Google's AI stack.

The claim arrived at a structurally significant moment. OpenAI and Anthropic are both preparing for public listings and intensifying their pursuit of high-value enterprise contracts. Google's pricing move is not simply a consumer promotion. It is a Capital allocation signal aimed at shifting procurement decisions before its rivals can anchor Market Share at higher price points.

Subscription Repricing and the Enterprise Calculus

Google reduced its top-tier AI Ultra plan from $250 to $200 per month and introduced a $100-per-month tier for developers and professional users. Pichai offered a stark illustration of current AI spending pressures: many large enterprises have already exhausted their annual token budgets before the halfway point of the year.

The repricing compresses margins in the short term but serves a longer-term platform capture strategy. Enterprises that build workflows around Google's model infrastructure face meaningful switching costs over time, making early price concessions a defensible use of capital for a company that reported $402.8 billion in total Revenue in 2025 and has committed between $180 billion and $190 billion in Capital Expenditure for 2026.

Gemini 3.5, Antigravity, and the Coding Market

Google launched Gemini 3.5 Flash, optimised for coding and automated tasks, with Gemini 3.5 Pro expected to follow in June. The model family powers a refreshed version of Antigravity, Google's coding assistant, which competes directly with Anthropic's Claude Code, the current Market Leader in AI-assisted development tooling.

Google's positioning in the coding segment is reinforced by talent acquisitions, including key hires from AI code generation startup Windsurf last year. The coding assistant market is strategically important: developer adoption tends to precede enterprise platform consolidation, making it a leading indicator of longer-term infrastructure spend.

Agents Enter Search, and the Consumer Stack

Beyond pricing and models, the more structurally significant announcement may be the integration of autonomous AI agents directly into Google Search. These agents can complete purchases, track ticket availability, and build schedules in real time without users leaving the search interface. Pichai framed the move as a reinforcing dynamic: AI-powered Search features drive higher Search engagement, protecting the Advertising revenue base that remains Alphabet's core Earnings engine.

Gemini Spark, a new agent pulling data across Gmail, Chrome, and YouTube, extends this integration into productivity workflows. The company also introduced AI-generated visuals and code directly within search results to address scientific and technical queries.

Liz Reid, who leads the Search team, described the changes as the most consequential transformation of Search since its inception. Nick Fox, overseeing the Search and Ads unit, called it the biggest reinvention of the search box in 25 years. These are not incremental claims, they reflect a deliberate effort to reframe Search as an AI-native platform rather than a legacy product with AI features layered on top.

Gemini's monthly active user base now stands at 900 million, more than doubling from 400 million a year ago. AI Overviews has reached 2.5 billion monthly users, and AI Mode is at approximately one billion.

Gemini Omni and the World-Model Ambition

Google also announced Gemini Omni, a new multimodal video model framed as the successor to Nano Banana, the image generator that attracted 13 million first-time users within four days of its September launch. DeepMind chief Demis Hassabis positioned Omni as a step toward a broader "world model" -- an AI system capable of simulating physical reality and generating any output from any input, starting with video.

The world-model concept remains early-stage but carries significant long-term implications for simulation, robotics, and scientific research applications, all of which represent substantial potential revenue pools beyond Google's current Business lines.

Hardware: Audio Glasses and the Wearables Race

Google announced its first audio smart glasses, built in Partnership with Samsung, Warby Parker, and Gentle Monster, targeting an autumn 2026 launch. The devices embed the Gemini assistant and deliver information through audio rather than a visual display, compatible with both Android and iOS.

A product demo showed the glasses ordering coffee via DoorDash, reading message summaries, adding calendar events, and providing navigation -- practical use cases designed to demonstrate daily Utility rather than speculative capability. The glasses can also take photos and apply AI image effects.

The competitive context is clear. Meta's Ray-Ban smart glasses sold seven million units in 2025, establishing meaningful consumer proof of concept in the category. Apple, Snap, and Alibaba are all advancing competing products. Google's entry adds institutional scale and ecosystem depth to a wearables race that is increasingly tied to AI assistant adoption.

The Takeaway

Google arrived at I/O 2026 from a position of improving competitive standing. Its pricing action compresses the cost advantage that newer AI entrants have used to attract enterprise customers. Its agent integration deepens the lock-in value of the Google ecosystem. And its hardware push signals that the platform battle for AI will extend into wearables.

The key risks are execution, particularly whether Antigravity can close the gap with Claude Code in developer adoption, and whether the Margin impact of aggressive pricing can be absorbed within Alphabet's broader capital deployment framework. For institutional observers, the more important question is whether Google's scale advantages in data, distribution, and infrastructure are now compounding fast enough to structurally widen its lead.