Google Leads Generative and Agentic AI Patent Applications

Google leads AI patent applications in both generative AI and the emerging field of agentic AI, according to an IFI Claims analysis covering the 14 months from February 2024 through April 2025.

The data shows Google overtaking former leader IBM in US generative-AI patent applications. Google and Nvidia also topped the US ranking for patents related to AI agents, followed by IBM, Intel and Microsoft.

Google’s Position in AI Patents

Google ranked first in US applications across overall AI, generative AI and agentic AI. Microsoft placed second in US generative-AI applications, while Nvidia’s position was especially strong in agentic systems.

• US generative AI: Google and Microsoft led, ahead of IBM.
• US agentic AI: Google and Nvidia led, followed by IBM, Intel and Microsoft.
• Global generative AI: Google ranked first, with six of the top ten positions held by Chinese companies or universities.
• Overall global AI: Google led, followed by Huawei and Samsung.

AI Patent Applications Are Rising Fast

IFI Claims counted 51,487 US patent applications related to generative AI in the latest year, a 56% increase. Granted US patents in the category rose 32%.

Generative AI accounted for 17% of AI-related US patent applications, while agentic AI represented 7%. Globally, the shares were 23% for generative AI and 5% for agentic AI.

China’s Role in the Global Rankings

The global list shows that the patent race is not only between large US technology companies. Three Chinese universities appeared in the top ten for agentic-AI patents, and Chinese companies or universities occupied six of the top ten places in generative AI.

That distribution reflects the scale of Chinese research and patent activity. It does not necessarily reveal which organisation has the strongest deployed model, but it does show where institutions are seeking legal protection for technical work.

Why Meta and OpenAI Were Missing

Neither Meta nor OpenAI appeared in the top ten. Meta has emphasized open-source development, while OpenAI has said it intends to use patents defensively. Different patent strategies can therefore make two companies with major AI products look very different in a filing-based ranking.

What Patent Leadership Does and Does Not Mean

Patent applications are evidence of research focus and an attempt to protect inventions, but they are not a direct scorecard for product quality or commercial success. Applications can take years to become granted patents, and broad portfolios may include ideas that never reach users.

Google’s lead still matters because it spans foundational generative systems and agents that plan, use tools and complete multi-step tasks. Those are two of the most commercially important directions in AI development.

Sources: Axios and the IFI Claims analysis. For related product coverage, read Jio’s TeleFrame AI device.