OpenAI just beat the world's richest man in court while Google proved it's no longer playing catch-up in the agent race.
The Summary
- Elon Musk's $150 billion lawsuit against OpenAI and Sam Altman was dismissed, ending his attempt to force the company back to its nonprofit roots
- Google unveiled Gemini 3.5 and a new AI agent called Gemini Spark at IO 2026, directly challenging ChatGPT and Claude's dominance
- A new model called Omni can turn images, audio, and text into video, expanding multimodal capabilities beyond static outputs
- The dual announcements mark a rare week where legal drama took a back seat to actual product velocity in the AI race
The Signal
The court dismissal of Musk's case removes the last credible legal threat to OpenAI's for-profit transformation. Musk argued the company had violated its founding mission by becoming a commercial entity worth over $80 billion. The judge disagreed. This wasn't just about money or principle. It was about whether early backers could bind a company to its founding structure forever, even as the technology and capital requirements evolved beyond anyone's 2015 projections.
The timing matters. OpenAI can now move forward without looking over its shoulder, just as Google showed up with Gemini 3.5 and an agent platform designed to actually build things. Not summarize. Not chat. Build. Gemini Spark is positioned as a direct agent competitor, the kind of tool that writes code, chains API calls, and automates workflows without constant human prompting.
"Google proved it's no longer playing catch-up in the agent race."
This is the first Google IO in three years where the company didn't feel like it was announcing vaporware to match OpenAI's last release. The products are live. The demos worked. Omni's ability to generate video from multimodal inputs puts Google ahead in the content generation arms race, where OpenAI's Sora still hasn't shipped broadly. If you can describe a scene in text, hum a soundtrack, and upload a reference image, Omni stitches it into video. That's not a research preview. That's a product.
The competitive dynamic just shifted. For two years, OpenAI set the pace and everyone else responded. Now Google has a credible agent story and shipping multimodal tools while OpenAI fights in court and ships incremental updates to GPT-4. The talent war, the product war, and the enterprise sales war are all heating up at once.
The Implication
If you're building with AI agents, watch what Google ships in the next 90 days around Gemini Spark. The agent landscape was starting to standardize around OpenAI's function calling and Anthropic's tool use. Google just threw a third framework into the mix with actual enterprise distribution through Workspace. The companies that build agent tooling need to support all three, or pick a horse and hope their bet wins.
For everyone else, the Musk case closing means the nonprofit-to-profit pathway is now legally tested. Expect more mission-driven AI startups to plan similar transitions from day one, with legal structures that allow pivots when the technology demands scale that donations can't fund. The "open" in OpenAI is dead. Long live OpenAI.