Google I/O 2025 was a litmus test. Did Sundar Pichai deliver?

Google's flagship search product is also facing renewed pressure, with rivals like Perplexity and OpenAI's ChatGPT gaining ground amid a broader shift in how consumers seek information.

Additionally, the tech giant is facing the threat of a breakup by the US government, as antitrust regulators intensify their scrutiny of its dominance in search and online advertising, a move that could have a profound impact on the way it delivers products and services in the future.

Despite these challenges, Pichai, who is set to complete a decade as Google CEO this year, appeared brimming with confidence as he took the stage for the nearly two-hour long keynote on May 20.

His confidence likely stems from the fact that Google is now in a far stronger position than two years ago, when it was caught off guard by the rapid rise of OpenAI's ChatGPT despite pioneering much of the foundational research behind the current wave of AI innovation.

This move resulted in Google co-founder Sergey Brin returning to the company in 2023 to focus on AI development after stepping away from his leadership role in 2019. The firm has also accelerated its AI efforts over the past two years, launching the Gemini family of AI models, deeply integrating AI capabilities into its suite of products and introducing new AI-powered offerings.

"Normally, you wouldn't have heard much from us in the weeks leading up to I/O. That's because we'd be saving our best models for this stage. But in our Gemini era, we are just as likely to ship our most intelligent model on a random Tuesday in March or a really cool breakthrough like AlphaEvolve just a week before" Pichai said in his keynote.

"We want to get our best models into your hands and our products ASAP (As Soon As Possible) and so, we are shipping faster than ever," he added.

Google released the latest version of its flagship Gemini model, Gemini 2.5, in March and rolled out an updated version with significantly improved coding capabilities to help developers build richer, interactive web apps earlier this month.

The search giant also unveiled Android's biggest visual update in years just a week before the conference and subsequently announced AlphaEvolve, a Gemini-powered AI coding agent that can invent new complex algorithms.

Pichai said they have announced over a dozen models and research breakthroughs and released over 20 major AI products and features since the last I/O edition.

"We are in a new phase of the AI platform shift, where decades of research are becoming reality for people all over the world," he said.

What’s driving this rapid pace?

During a select media briefing, Pichai said that we are in a moment where AI, as a horizontal technology, applies across everything and that Google takes a long-term view and invests accordingly. Some ideas can come right away into products while some others take time, he said.

"One of the things I'm proud of is the depth and breadth of the research we are undertaking in Google DeepMind and Google Research," he said. Pichai cited examples such as Gemini Diffusion (an experimental text diffusion research model to achieve low latency), Waymo (autonomous ride-hailing service), and the company's quantum computing efforts.

At the heart of this rapid pace of innovation is Google's full-stack approach to AI, wherein the tech giant is investing in developing every layer of the AI technology stack: Research, product and platforms, models and tooling, and AI infrastructure.

"We are uniquely positioned as a company. While everyone else is dependent on one vendor, Nvidia, Google trains and serves its model on TPUs (Tensor Processing Units), and we are in the seventh generation of our TPUs" Mitesh Agarwal, managing director of the Asia Pacific region for Google Cloud, told Moneycontrol. Ironwood, the firm's seventh-generation TPU chip, is expected to be available to Google Cloud customers later this year.

Sameer Samat, President of the Android ecosystem at Google, said "There are very few companies - maybe we are the only ones in the world - that have that full stack of capabilities".

Google parent firm Alphabet plans to spend around $75 million in capital expenditure in 2025 to boost the company's AI efforts, Pichai said earlier this year.

Google’s AI upgrades in I/O 2025

AI was also front and centre at the conference with Google making a series of announcements for both developers and consumers.

These announcements ranged from new experiences and feature enhancements in the company's core Google Search product (AI mode, broader AI Overviews rollout, and new agentic shopping features) and its AI chatbot Gemini, to the launch of new video and image generation models (Veo 3, Imagen 4) and an update to its flagship Gemini AI models (Gemini 2.5 Flash, Gemini 2.5 Pro).

Google also launched a new AI filmmaking app Flow, an AI-first 3D video communications platform Google Beam (previously Project Starline), an asynchronous coding agent Jules, and an AI tool to generate UI designs from text prompts Stitch. Additionally, the company announced new partnerships with firms like Warby Parker and Gentle Monster to build AI-powered glasses based on Android XR, its mixed reality operating system.

Of all the announcements, what appeared to stand out most to attendees Moneycontrol spoke with, was the video generation model Veo 3. The model now allows users to generate videos with sound effects and background noise like traffic noises in the background of a city street scene, birds singing in a park, or dialogue between characters from a text prompt. This capability is currently missing in other leading AI video generators like OpenAI's Sora, Runway, and Pika.

Flow, an AI filmmaking tool that allows users to create cinematic clips, scenes and stories using Google's most advanced models - Veo, Imagen and Gemini - was also seen by many attendees as a serious challenger to software major Adobe's offerings.

Positioned as a tool built by and for creatives, Flow enables users to create videos using text prompts (text-to-video mode), upload or generate images to use as starting or ending frames (frames-to-video), or provide images as subject or style references (ingredients-to-video).

The tool also offers advanced camera controls including camera angles, movement, and perspectives and a scene builder that lets people edit, extend, and stitch together their video clips.

It appears that investors have also given a thumbs-up to Google's latest wave of AI updates, with the company’s shares rising for a second straight day on April 22, reportedly reaching a three-month high.

Monetising AI offerings

Google is also taking various measures to monetise its AI offerings at a time when investors are closely watching how major technology companies are turning their tens of billions of dollars in AI investments into profits.

During the conference, Google introduced a $250 monthly subscription plan called Google AI Ultra for its power users in the United States, giving early access to its experimental AI products and features, as well as the highest level of access to its AI models and services.

The search giant stated that it is also starting to test ads in AI Mode, the firm's AI search experience designed to take on rivals like Perplexity and OpenAI's ChatGPT.

Additionally, Google is expanding search and shopping ads on AI Overviews to desktop in the United States with plans to expand ads in AI Overviews in English to select countries on mobile and desktop later this year. The company first introduced ads in AI Overviews on mobile in the United States in October last year.

Pichai has previously stated that ads, cloud and subscriptions will be the primary avenues to earn money from its AI offerings.

At the media briefing, Pichai said the subscription plans allow the company to develop new products and features faster and get it in the hands of people. "Project Astra is a great example. It has come via Gemini Live for everyone. If we had something like Google AI Ultra last year, I think we would have shipped Astra faster," he said.

Project Astra is Google's research sandbox to explore the "bleeding edge" of Gemini's capabilities, including video understanding, screen sharing and memory.

Pichai mentioned that the tech giant is working "super hard" on bringing the technology to as many people as possible.

"Ads allow us to make it broadly available to everyone in the world. But with AI, especially frontier technology, it is expensive both to develop and to serve consumers," he said.

The Alphabet chief also highlighted the range of AI models that Gemini 2.5 offers, including Pro Deep Think, Pro, Flash, and Nano, as part of an effort to make it more accessible to people.

Pichai mentioned that Google will continue to invest in fundamental research, with intelligence, personalisation, and agents as its three focus areas over the next year.

“Our approach has always been pro-competitive, and I think we will have to earn our success on our merits," he said.