An analysis of more than 1,000 patents that Google bought from IBM offers a glimpse inside the search giant’s increasingly frantic efforts to protect its Android mobile operating system against legal attacks from competitors.
IPVision, which makes patent-analyzing software, says that the 1,029 patents that Google bought from IBM in July contain little that the company could use to either attack its competitors or defend its own products.
Bundles of patents covering computing—especially mobile computing—technology have become a hot property in recent months. Apple, Nokia, Microsoft, and others have used them to extract money from competitors, or even to block those competitors’ products from being sold. This year, Apple successfully prevented the sale of some Samsung devices in much of Europe, while Microsoft has used patents to extract millions of dollars in licensing fees, from companies including Samsung and HTC, for using Google’s “free” Android operating system.
In August, Google announced it planned to buy Motorola’s cell-phone business, Motorola Mobility, for $12.5 billion. That company’s patent portfolio was key to the deal—Google CEO Larry Page said it would “better protect Android from anticompetitive threats from Microsoft, Apple, and other companies.”
Hoo-Min Toong, cofounder of IPVision, says, “It’s common knowledge that Google is patent poor. Not only does it need to gain patents in their own product areas, but also in defending themselves against other claims.”
Google is young compared to many of its rivals, without decades of research and patent filings, unlike competitors such as Apple. Toong says Google’s current stock of patents is not only small, but it also relates to fundamental search technologies and techniques, rather than competitive areas such as mobile software.
At first glance, the patents Google bought from IBM look good. The U.S. patent office maintains a set of subject categories used to sort patents, and most of those acquired from IBM spread across the “700 series,” where new software ideas and techniques are to be found.
This image, created by IPVision, shows how that contrasts with Google’s prior patent portfolio, which is concentrated in just one category related to database and information retrieval techniques used in search. “This looks good in that I’m getting patents that strengthen my position and could be used against other companies,” says Toong. “But if you look under the covers, things aren’t so good.”
Patent filings must reference previous, related patents, making it possible for IPVision to calculate whether a collection of patents contain clusters that are highly connected—what is termed a “patent thicket.” Such clusters, combined with strong “families” of patents filed on different aspects of the same technology, make it easier to mount a bulletproof assertion in court that a competitor’s technology infringes on a company’s intellectual property.
“When you look for clusters, this collection from IBM actually seems to be one-off patents that aren’t related,” says Toong. “They got a lot of patents that were scattered.” Companies that make good licensing revenue from patents generally have strong clusters and families of patents, says Toong, who adds, “the IBM portfolio Google acquired has neither.”
Another image from Toong’s analysis shows that only around 10 percent of the collection bought from IBM is linked to any other patents in the collection.
“From this analysis, we can see there’s a whole range of things that Apple built off these patents, and perhaps even stronger connections to Microsoft patents, but the IBM portfolio appears to consist of isolated patents,” says Toong. “To assert that those later Microsoft and Apple patents infringe on the earlier ones, you really would want strong clusters and patent families; Google seems to have gotten a grab bag, although of course additional due diligence would need to be performed [to confirm that].”
Two weeks ago, it emerged that Google had purchased a second set of IBM patents, but Toong says it is too soon to have exhaustively analyzed that collection.
Toong expects Google and competitors in the mobile space to look for other chances to buy up mobile technology patents. Eastman Kodak’s many imaging patents are one possibility, as imaging has become a key part of the smart phone. “Nokia and RIM both have large portfolios in the wireless space and have businesses that appear to be in jeopardy,” says Toong, “so they are also candidates for buyers looking for patents.”
However, Toong says, a really smart company should look further ahead to the next technological fight. “This is going to repeat in other spaces where technology is converging fast,” he says. IPVision’s analytics suggest to him that patents around alternate methods of controlling computers such as gestures and voice could soon become very valuable. He cited how heavily invested Microsoft is in the Kinect gesture control system, which is used for games today, but which the company plans to use more broadly in the future.
When reached for comment, Google would only confirm that it had indeed acquired a patent collection from IBM, and state that Google is actively working to encourage reform of the patent system to prevent wrangling over patents from hindering competition in new areas of technology.
Lewis Lee, cofounder of IPStreet, a startup that also analyzes patents, argues that making it easier for companies to assess the value of patents prior to buying them can also help. “We’re seeing patents become a business asset that is traded, like property or stocks, but we lack any way to value them or know if they are worth more or less than others,” says Lee. “Not even Google will know whether the huge amount it paid for Motorola’s patents is reasonable.”
Lee thinks that the tools offered by his company, and those from IPVision, could help to change that by making it possible for a small inventor or company to prove the value of a patent. Making patents more tradeable could diffuse some of the tensions distracting Google and others from working on technology, claims Lee.
However, holders of large patent portfolios could still use those portfolios to force aggressive licensing terms on their competitors, Lee concedes. “Changing that will need political action on reform,” he says.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.