Three Big Things to Watch at Google I/O This Year
Google's annual developer conference, I/O, is set to kick off tomorrow in Mountain View, California. This year, the company arrives in a position that feels distinctly different from just twelve months ago. In 2025, Google was riding high on the success of Gemini 2.5 Pro, and the battle among top-tier foundation models seemed to come down to subjective preferences. But times have changed. Today, Google finds itself in a clear third place in the foundation model race, especially when it comes to coding capabilities—a key benchmark for modern AI. Rivals like Anthropic's Claude Code and OpenAI's Codex have set a new standard, and Google has even reportedly allowed some DeepMind engineers to use Claude to keep pace. Yet the conference isn't just about playing catch-up. Google still leads in areas like AI for science, and there are plenty of exciting developments to anticipate. Here are three things I'll be paying close attention to over the next two days.
1. An Attempted Coding Comeback
Google is taking its AI coding crisis seriously. According to reports from The Information, a new AI coding team has been formed at DeepMind, and the Los Angeles Times notes that Nobel laureate John Jumper—co-winner of the 2024 Chemistry Nobel for AlphaFold—is lending his expertise to the effort. This suggests that a major coding release could be unveiled at I/O, perhaps an update to the Antigravity agentic coding platform. However, don't expect a miracle. Even internal Google models and tools lag behind what's publicly available from rivals, and engineers were reportedly fighting for access to Claude Code just last month. Unless the company has made astonishing progress in weeks, Google likely won't reclaim the coding frontier immediately. But the commitment is clear, and this will be a key area to watch for signs of future competitiveness.
2. Science and Health Dominance
While coding may be a weakness, science is where Google DeepMind shines—it's the only frontier AI company to have earned a Nobel Prize. As large language models increasingly dominate the AI-for-science landscape, Google has only widened its lead. At I/O, expect to hear about tools like the AI co-scientist, described by a Stanford researcher as an "oracle" that formulates hypotheses and research plans in response to user queries. Another standout is AlphaEvolve, a system designed to accelerate evolutionary biology and drug discovery. These tools represent a different kind of AI leadership—one that doesn't rely solely on coding benchmarks but on tangible scientific breakthroughs. For health, Google will likely showcase how these models are being applied to protein folding, drug design, and personalized medicine. This is an area where Google isn't just competing—it's defining the cutting edge.
3. The AI Co-Scientist and Beyond
Beyond the headline-grabbing coding and science news, I'm eager to see what other AI innovations Google unveils. The AI co-scientist, which we briefly touched on, deserves its own spotlight. This tool goes beyond simple question-answering; it actively generates novel hypotheses and research plans, effectively acting as a creative partner for scientists. At I/O, Google may show how this system integrates with other products like Google Scholar or cloud services. Additionally, keep an eye out for updates to Gemini, the company's flagship model family, and how it's being woven into consumer apps like Search, Photos, and Workspace. Google has a history of surprising developers with new APIs and frameworks, and this year's conference could bring announcements that reshape how we think about AI assistants, code generation, and even creative tools. The next two days promise to be revealing for anyone tracking the future of artificial intelligence.
In conclusion, Google I/O 2026 comes at a pivotal moment. The company is fighting to regain its lead in coding while solidifying its status as a science powerhouse. Whether through a surprise coding upgrade, a deeper dive into scientific AI, or a broader vision for everyday AI tools, Google has much to prove and plenty to showcase. I'll be watching closely—and you should too.
Related Discussions