Debates over Big Tech M&A often focus on “killer acquisitions,” where a dominant firm buys a startup in order to eliminate a future rival. A more subtle and, in some ways, more difficult problem is that innovation can lose momentum after the deal closes even when the buyer did not set out to destroy the target. In those cases, the startup is not killed in the literal sense. Its technology is absorbed, but the full innovative potential of the technology is never realized.

That matters because acquisitions are often justified as a way to accelerate innovation, scale promising technologies, and give startups the resources they need to grow. The empirical picture is more complicated. Recent work from the OECD suggests that startup innovation can decline after acquisitions, and acquirers do not necessarily compensate by increasing their own innovation efforts. That finding points to a different kind of harm: not deliberate suppression, but the loss of innovative momentum inside a larger organization. The startup’s knowledge does not disappear immediately. Rather, it becomes harder to develop, harder to recombine, and easier to underuse.

This is the core of what I call a slack acquisition. The term does not refer to a deliberate plan to eliminate a competitor. It refers to acquisitions in which valuable startup knowledge is brought inside the acquirer, but then underused, delayed, or poorly integrated. The problem is not always foreclosure. It is stagnation. A startup may have promising technology, a strong development path, or a novel idea that is still evolving. Once acquired, however, that idea may be folded into a larger organization whose incentives, internal processes, or strategic priorities are not well aligned with developing it further. The acquisitions of Danger by Microsoft, Bump by Google, Lala by Apple, Yap by Amazon, and the more recent waves of Big Tech buying in the metaverse and AI illustrate different shades of slack acquisitions, ranging from outright shelving to partial integration and the underuse of acquired innovation.

That distinction matters for competition law. Traditional legal constraints on mergers are very good at identifying price effects, concentration effects, and clear exclusionary conduct. They are less effective at addressing acquisitions that weaken innovation in a subtler way. If a startup’s research output, patenting activity, or product development declines after a deal, but the acquirer does not pick up the slack, the competitive harm may still be real even if there was no intent to kill a rival. In that sense, the law may be looking in the wrong place if it focuses only on whether the acquisition was meant to be anticompetitive.

The digital economy makes this issue especially important. Startups in digital markets often hold modular forms of knowledge: code, design, technical methods, data practices, product ideas, and user-interface tools that can be recombined in many different ways. Their value is not only in what they currently sell, but in the future innovation paths they may open. That makes digital acquisitions different from many traditional asset purchases. If a startup’s knowledge is absorbed into a large organization and then left underdeveloped, the market loses more than one firm. It loses an engine of experimentation, recombination, and spreading knowledge.

That is also why the problem can be hard to see. A deal may look successful from the outside because the acquirer announces how it will combine complementary companies, create, synergy, or expand the reach of products and services. Yet the real test is not whether the buyer can narrate a strategic rationale. The real test is whether the acquired capacity for innovation will continue to develop. . In some cases, the acquired work gets trapped between departments, diluted by bureaucracy, or simply pushed aside by more urgent priorities.

This does not mean that all acquisitions are bad for innovation. Many produce genuine benefits. Some startups do gain access to capital, scale, and infrastructure they could not have built alone. Some technologies do become more useful after acquisition because the buyer has the resources to commercialize them properly. That is why the issue is subtle. The law cannot assume that every transaction is either clearly procompetitive or clearly anticompetitive. Some acquisitions help innovation. Others quietly suppress it through underuse rather than design.

The OECD evidence is useful because it captures this middle ground. It suggests that acquired startups are often innovation-intensive before the transaction, but that their production of patents can decline afterward. At the same time, the acquirer does not necessarily increase  innovation. That does not prove that every acquisition is harmful. But it does show that the key danger is not limited to deliberately killing a company.

From a policy perspective, merger analysis should look beyond intent and market shares. Regulators should also ask whether the acquirer has a credible plan to integrate, fund, and develop the acquired technology. They should be attentive to post-acquisition innovation outcomes, especially in digital markets where value depends not only on current competition, but on whether innovation continues. That could mean better disclosure about integration plans, more careful scrutiny of multiple startup acquisitions by dominant firms, and a broader willingness to treat innovation harm as a real merger.

This approach would align merger review more closely with how innovation works. A startup’s most valuable contribution may not be the product it has already shipped, but the technical path it is still exploring. If that path is bought and then allowed to go stale, the market may lose exactly the kind of dynamism competition law is supposed to protect.

Marco Corradi is an assistant professor at ESSEC Business School in Paris and Singapore. This post is based on his recent paper, “Slack Acquisitions,” available here.