Big Tech Bought 1,200 Startups, and Patent Power Concentrated From 4 Percent to 17 Percent

Big Tech acquisitions are concentrating patents, study finds

A study of 1,200 Big Tech startup acquisitions has put numbers on something regulators have suspected for years: repeated buying of startups concentrates patent power. The Big Tech acquisitions study, published as a CEPR discussion paper in 2026 and summarized by VoxEU on October 20, found that patent citations rose 28 percent after acquisitions, while the eight large acquirers' combined share of patents in the studied technology fields climbed from about 4 percent to 17 percent. That is a 13-point jump in concentration, driven by deal after deal.

The study matters because merger review usually examines one deal at a time. A single acquisition of a small startup rarely looks threatening: the target is tiny, the market share change is small, and the deal sails through. But the study shows the cumulative effect is different. Dozens of small deals, each harmless in isolation, add up to a patent portfolio that shapes who can build what, on what licensing terms, and with what negotiating power. Regulators are starting to notice the pattern, and this paper gives them evidence.

The researchers also found a nuance that complicates any simple narrative. When more inventors leave after an acquisition, their previous patents get cited more by the companies hiring them. Where the acquirer retains more inventors, a larger share of citations comes from the acquiring company itself. In other words, acquisitions do not just transfer patents. They move people, and people carry knowledge that shows up in later innovation, wherever they land.

Why serial acquisitions escaped scrutiny

Serial acquisition strategies work because each individual deal falls below the thresholds that trigger serious review. A large company can buy a ten-person startup with an interesting patent portfolio and a promising team, and no regulator blinks. Do that fifty times across a decade and you own a meaningful slice of a technology field's intellectual property, plus the teams that created it. The 2023 US merger guidelines identified serial acquisitions as an area of antitrust scrutiny, but enforcement has lagged behind the practice.

The patent angle makes the problem sharper than market share alone. Patents are exclusionary by design: they let the owner stop others from using an invention, or charge for the privilege. When a handful of companies hold 17 percent of patents in a field, up from 4 percent, smaller firms face a licensing landscape shaped by their largest competitors. That affects what gets built, what gets funded, and which startups can credibly threaten an incumbent.

This is not an argument that all acquisitions are harmful. Startups need exits, and acquisition is a legitimate one. Many acquired teams do their best work inside a larger company, with resources they could never raise alone. The study's point is narrower and more useful: evaluate the pattern, not just the deal. A regulator that approves fifty small acquisitions without asking what they add up to is not really reviewing the market. It is stamping paperwork.

What this means for startups and regulators

For startup founders, the study cuts both ways. On one hand, it validates the acquisition path: Big Tech buyers clearly value startup innovation, and the citation gains suggest acquired technology gets used and built upon. On the other hand, a more skeptical review climate could slow exits or reduce prices, especially for deals in fields where the buyer already holds significant patent share. Founders should understand that their acquirer's deal history is now part of the regulatory picture, not just their own company's story.

For regulators, the practical step is cumulative review: tracking a buyer's acquisitions in a technology area over time and treating the portfolio as the unit of analysis. That requires data infrastructure and patience, since the pattern only becomes visible across years. But the alternative is the status quo, where concentration grows deal by deal while each deal is judged harmless. The study gives that approach a number: 4 percent to 17 percent, with citations up 28 percent.

For everyone else, the study is a reminder that competition policy is about structure, not slogans. Markets do not stay competitive by accident. They stay competitive when someone watches the cumulative effects of individually reasonable decisions. The same logic applies beyond patents: the Schneider Electric purchase of PTC for $23.7 billion and Salesforce's $5.66 billion Braze deal are large, visible transactions that get proper review. The quiet danger is the hundred small deals nobody reviews at all.

The next test will be whether regulators turn this evidence into practice. Guidelines already mention serial acquisitions. The study supplies the empirical backing. What remains is the institutional will to look at a company's tenth small acquisition in a field and ask what the first nine already bought. Until that happens, the patent concentration trend the researchers documented is likely to continue, one harmless-looking deal at a time.