Google Unveils USD 250 Monthly AI Subscription at I/O 2025

Google launched a new premium AI subscription at its annual developer conference, Google I/O 2025, pricing the AI Ultra plan at USD 249.99 (SGD 340) per month.

Credit: GOOGLE

The subscription offers early access to Google's most advanced Gemini models, experimental tools, 30TB of cloud storage and YouTube Premium.

This makes it one of the most expensive flat-rate AI plans on the market, surpassing OpenAI’s and Anthropic’s USD 200 tiers.

Google is positioning Gemini as a standalone product, moving beyond its role as a background engine for Search and Workspace.

The company also introduced Android XR, a new extended reality platform supporting AR, VR and MR experiences.

Android XR will debut on Samsung’s Project Moohan headset, developed with Google and Qualcomm, and is expected to launch later this year.

The headset will feature deep integration with Gemini for real-time assistance in virtual environments.

Google also previewed smart glasses developed with Gentle Monster, Xreal and Warby Parker, offering features like live translation and object recognition.

These XR products place Google in direct competition with Meta, which already markets Ray-Ban Meta smart glasses and is developing an AR headset codenamed Orion.

In AI image generation, Google introduced Imagen 4, its most advanced text-to-image model, capable of rendering fine details and generating legible text within images.

Imagen 4 is designed for use in creating posters, greeting cards and slides that require precise typography.

Google also showcased Project Astra, a prototype multimodal agent that combines visual input, memory and audio for real-time contextual understanding.

Astra can describe environments, identify objects and maintain voice conversations, with some features expected to roll out to Gemini apps and Android later this year.

The new AI Mode in Google Search is now live in the US, replacing traditional link-based results with AI-generated summaries, explanations and citations.

Google is testing new ad formats within these summaries to maintain revenue from Search ads, which remain a major income source.

Gemini 2.5 Pro, developed by DeepMind, was also announced as Google’s most advanced AI model, outperforming GPT-4 in internal software development benchmarks.

It includes a Deep Think mode for step-by-step reasoning and a lightweight variant that runs 22% faster, optimised for mobile and edge devices.

Google also introduced Beam, a 3D video chat system developed from Project Starline, aimed at enterprise users and supported by HP hardware.

Beam uses depth sensors and AI rendering to simulate face-to-face interaction, aligning with Google’s vision for AI-enhanced remote collaboration.

While hardware announcements were limited, the event underscored Google’s commitment to embedding AI across its ecosystem and monetising Gemini as a core product.

---

• Google launches USD 249.99 (SGD 340) monthly AI Ultra subscription

• Android XR platform debuts with Samsung’s Project Moohan headset

• Imagen 4 improves AI-generated image quality and text rendering

Source: FORBES