Gemini 3 Beats Rivals, $15B Anthropic Deal, and Pichai’s AI Bubble Warning

Google’s Gemini 3 Climbs the Leaderboards 🏆

Google just released Gemini 3, its newest and most powerful AI model, claiming it brings a “new era of intelligence” by beating competitors on most major tests. They also launched a new coding platform for AI agents called Antigravity.

The details:

• Gemini 3 (and its stronger version, Deep Think) set new records on difficult thinking tests like Humanity’s Last Exam and ARC-AGI-2, beating GPT-5’s previous high scores.
• The model now holds the top spot for scientific knowledge, maths, and understanding multiple types of data (like text and images), while being just slightly behind Claude Sonnet 4.5 in bug-fixing code tests.
• Gemini 3 is excellent at creating visual layouts and user interfaces on the fly, a feature Google is showing off in Search’s AI Mode.
• Google also launched Antigravity, a free coding platform that lets AI agents control browsers and run complex coding tasks by themselves.

Why it matters: With the long-awaited Gemini 3, Google has taken the performance lead from OpenAI for the first time in a while. With its huge resources and deep links across all of Google’s products, the company is making a very strong play as the industry moves toward more powerful, autonomous AI agents.

Microsoft, Nvidia Team Up with Anthropic 🤝

Microsoft and Nvidia just announced a massive new partnership with Anthropic, involving a combined investment of $15 billion. As part of the deal, Anthropic also promised to spend $30 billion on Microsoft’s Azure cloud services.

The details:

• Nvidia will invest up to $10 billion, and Microsoft will add up to $5 billion, pushing the value of Anthropic (the maker of the Claude AI model) close to $350 billion.
• Claude models will now be available through Azure AI Foundry, making them accessible across all three major cloud platforms (AWS, Google Cloud, and Azure).
• Nvidia and Anthropic will also work together to design new AI chips optimized specifically for Claude.
• Microsoft CEO Satya Nadella said this move supports his idea that the AI industry must reject a “zero-sum narrative,” where only one company wins.

Why it matters: This deal shows that Microsoft is serious about its “positive-sum future” vision, making a huge partnership with a competitor’s main partner (Anthropic is close with Google and Amazon). This circular economy—where investments fund cloud commitments which fund more chip development—is the new reality of the AI race. It also gives Microsoft a crucial second top-tier AI partner alongside OpenAI.

Sundar Pichai Warns of AI Investment ‘Irrationality’ ⚠️

Alphabet CEO Sundar Pichai told the BBC that while the growth in AI investment is “extraordinary,” he believes there are “elements of irrationality” in the current boom, suggesting a bubble might be forming.

The details:

• Pichai warned that “no company is going to be immune” if the AI bubble bursts, even though Google’s parent company, Alphabet, is well-positioned to handle a downturn.
• He compared the current moment to the dot-com boom and bust of the late 1990s, saying the industry can “overshoot” in its investment cycles.
• Pichai also warned about the “immense” energy needs of AI, which is delaying Google’s own climate targets and requires new global infrastructure.
• He remains positive that, like the internet, AI is a “profound” technology that will fundamentally change jobs, making those who learn to use AI tools better performers in their field.

Why it matters: When the CEO of the world’s leading AI developer (Google) and parent company of DeepMind speaks out, the market listens. Pichai’s warning confirms the fear felt across Silicon Valley that while the technology is real, the trillion-dollar investments and soaring valuations may not be sustainable, mirroring the famous “irrational exuberance” warnings that preceded the dot-com crash.

You May Also Like: Jeff Bezos’s New AI Startup, Grok 4.1 Overhaul, and Anthropic’s Job Loss Warning

Jigar Chaudhary

Jigar Chaudhary is the Editor-in-Chief at UrviumAI, where he oversees coverage of artificial intelligence news, tools, and in-depth studies. With over 5 years of experience analyzing AI and robotics, he focuses on maintaining high editorial standards, accurate reporting, and clear explanations to help readers understand how AI is shaping the future.