OpenAI is not patenting the intelligence itself, but rather the products built around that intelligence. A detailed analysis of the company's entire patent portfolio shows that none of its 69 patents protect how models are built, trained, or made to behave. Instead, they protect the visible machinery that customers see: the way ChatGPT writes inside a document, how Sora converts text to video, and how agents click around a screen.

What Is OpenAI Actually Protecting in Its Patents?

When you line up OpenAI's patents against its product launches, they read like an annotated roadmap. The company filed its Canvas writing-tool patent three days before Canvas appeared, and its Sora patent the day before it announced Sora. Four of OpenAI's video patents share the priority date of February 14, 2024, the day before the company announced Sora to the world.

This product-focused approach mirrors how pharmaceutical companies operate. Drug makers often cannot patent the basic chemistry of a medicine, so they patent the auto-injector pen that delivers it. OpenAI does the same thing with artificial intelligence. The company leaves the science unpatented, either keeping it as a trade secret or publishing it years ago, which permanently bars anyone, including OpenAI itself, from patenting it later.

How Does OpenAI's Patent Portfolio Actually Break Down?

The 69 patents are not one cohesive portfolio, but rather three separate collections stitched together. OpenAI filed 49 of them itself, but filed nothing at all before mid-2022. The rest it acquired: 16 patents came along with a hardware startup called Rain, and 4 more came with a search-database startup called Rockset. Every patent in OpenAI's name that predates 2022 came from one of those purchases, not from OpenAI's own labs.

This means what appears to be a decade of invention is really about three years of internal development, combined with two companies' older work. The portfolio also contains some inflation: the 69 patent families describe only about 34 to 36 distinct inventions. OpenAI files the same invention several ways, which inflates the count by roughly a third.

• Internal Patents: 49 patents filed by OpenAI itself, with the earliest priority dates starting in mid-2022
• Rain Acquisition: 16 patents acquired along with the hardware startup Rain
• Rockset Acquisition: 4 patents acquired along with the search-database startup Rockset
• Actual Distinct Inventions: Only 34 to 36 unique inventions across all 69 patent families

What Science Did OpenAI Choose Not to Patent?

The absences in OpenAI's patent portfolio are as revealing as what is included. There is not a single patent in the entire portfolio on a model's architecture, on a pre-training recipe, or on an alignment method. The science that makes the models work is nowhere in the fence.

OpenAI's most influential ideas are owned by no one. The company published reinforcement learning from human feedback (RLHF), a technique that trains AI systems using human feedback, and its scaling research without patenting them. This makes them permanent prior art against everyone, including OpenAI itself. Prior art is earlier public knowledge that blocks a later patent. By publishing these foundational methods, OpenAI ensured that no competitor, including itself, could ever patent them.

What Unusual Details Appear in OpenAI's Patents?

Some of the most specific details in OpenAI's patents reveal how granular the company's product protection strategy has become. OpenAI tried to patent a price: its prompt-caching patent claims the automatic 50 percent discount on reused input as protected subject matter, so the exact number on its own price sheet became part of the invention. The company also wrote a literal web address into a patent. The path.well-known/ai-plugin.json, the exact spot a website would post its instructions for an AI, appears in the plugins filing.

Additionally, one made-up phrase, "generative response engine," runs through many of OpenAI's patents. It is the in-filing alias for the company's own model, which lets the claims describe what ChatGPT does without ever naming ChatGPT.

How Does This Compare to Anthropic's Patent Strategy?

OpenAI's approach is not unique among frontier AI companies. OpenAI's chief rival, Anthropic, owns about 17 patent families, and independent analysis confirms it originally filed only about 1 of them. Nearly all the rest are older patents it acquired, and the patent-office records name IBM on the oldest ones, filed years before Anthropic existed. Anthropic's single homegrown patent covers a computer-use agent, which is to say a product.

Two labs at the top of the field, opposite in how much they patent, land in exactly the same place: the science stays unpatented, and the product gets the fence. This reveals a broader shift in how frontier AI companies are building defensible competitive positions.

What Does This Mean for OpenAI's Public Offering?

For anyone weighing OpenAI as an investment, the patent portfolio matters in a concrete way. OpenAI confidentially filed to go public in June 2026, and when a company reportedly worth about 852 billion dollars files to go public, buyers of the stock will ask what it actually owns. As of July 2026, the honest answer on patents is one product-tracking software estate that OpenAI built in about three years, plus two smaller estates it bought. That is a narrower and more specific asset than the hype implies, and it is also a legible one.

The patents map the exact layer OpenAI believes it can defend, and they hint at how a shield built for defense could be picked up as a sword later. Patents stay hidden for about 18 months after they are filed, so everything OpenAI has filed since roughly early 2025 is invisible today, and even the portfolio that can be seen is already a year and a half out of date.

How to Understand OpenAI's Patent Strategy

• Focus on Products, Not Science: OpenAI patents the visible machinery customers interact with, such as how Canvas writes and edits documents or how Sora generates video from text, rather than the underlying model architecture or training methods
• Timeline Alignment with Launches: The company files patents immediately before product announcements, with Canvas patented three days before launch and Sora patented the day before its announcement, showing deliberate product protection timing
• Acquisition-Heavy Portfolio: More than half of OpenAI's patents came from acquiring Rain and Rockset, meaning the company's internal patent development only spans about three years rather than a decade
• Science Remains Unpatented: Foundational AI research like reinforcement learning from human feedback and scaling insights are published without patent protection, making them permanent prior art that no company can later patent

OpenAI filed the training pipeline behind computer-use agents in April 2023, roughly two years before "computer use" became a product category anyone was racing to fence. This early filing suggests the company was thinking about product defensibility long before the feature became a competitive battleground.