Concerns abound that ‘killer acquisitions’ by tech giants could deter startup entry and stifle innovation. This column analyses patent data before and after 1,200 acquisitions by eight technology giants between 1995 and 2020. On average, the ‘Big Eight’ firms appear to invest in the acquired technology rather than kill it. Acquired inventions become more influential, more widely cited, and more embedded in follow-on innovation by both the acquirer and competitors. However, longer-term changes to the competitive landscape need further investigation.
When the largest technology companies like Amazon, Google, and Microsoft go on acquisition sprees, buying up hundreds of startups, does this accelerate innovation or stifle future competition? Could the tech giants be purchasing innovative competitors simply to shelve them? The concern that certain deals might be ‘killer acquisitions’ has become a focal point in antitrust debates (Federico et al. 2017, 2018, Federico et al. 2020, Valletti 2025). One worry is that a ‘kill zone’ around the tech giants deters startup entry (Kamepalli et al. 2020). Evidence from pharmaceuticals shows that some acquisitions are indeed undertaken to shut down competing projects (Cunningham et al. 2021). Whether acquisitions by large technology firms have similar effects on innovation is much less clear.
In a new study (Ederer et al. 2026), we examine these questions by using patent data and LinkedIn profiles to analyse innovation before and after 1,200 acquisitions by eight technology giants between 1995 and 2020. Their findings provide a nuanced picture of the market for technology, which does not rule out the existence of digital killer acquisitions, but suggests that it is more common for these firms to acquire and invest.
Evidence from patenting
We begin by examining what happens in a technology field (defined using patent classifications) following the acquisition of a startup that is patenting in that field. We find that acquisition activity comes in waves, with large acquirers purchasing multiple startups in a field over a short period of time. Patenting tends to accelerate prior to the first acquisition in a field, and it continues to grow as long as acquisition activity continues in that field. Once acquisition activity in a field ends, patenting falls back. Figure 1 illustrates this pattern.
Figure 1 Patenting in a technology field before and after first acquisition
a) Common pre- and post-trend
b) Heterogeneous post-trend by no. of follow-on acq
Why does patenting accelerate before any deal takes place? We propose a stylised model in which beliefs about the commercial viability of technology drive both startup entry and acquisitions. When an area looks promising, startups enter and patent and incumbents buy. But when expectations sour, entry, patenting, and acquisition activity all fall together. The patterns in Figure 1 can therefore arise through selection, with acquisitions concentrating on technology areas that are already taking off, rather than through the effect of the deals themselves. To account for this, our analysis compares each acquisition field with fields on similar pre-acquisition patenting trajectories. Even relative to these matched fields, patenting continues to grow after an acquisition, with the growth concentrated in areas where further acquisitions follow.
As deals accumulate, the combined share of all patents in a technology area held by the ‘Big Eight’ acquirers often rises significantly. The combined effect of acquisitions and internal patenting raises that share from roughly 4% to around 17% after multiple acquisitions in the same area.
Evidence from citations and employees
The second part of our study focuses on the patents held by the acquired startups and uses forward citations as a proxy for cumulative innovation and economic significance of the acquired intellectual property. We find that acquired patents experience a significant increase of around 28% in their citation rate after an acquisition when compared to other patents in the same field (Figure 2).
Figure 2 Citation rates of patents held by acquired startups, after acquisition
In percentage terms, the largest citation increase comes from ‘self cites’ made by the acquiring big tech firm. The citation increase at the acquiring firm suggests that, on average, these firms invest in the acquired technology rather than killing it.
In absolute terms, most forward citations come from third parties, and the employment data suggest that labour mobility is an important part of the broader story. When inventors leave acquired startups for other firms, they carry knowledge with them. The data show that where retention of target inventors is low (i.e. more inventors departing for other employers), more citations come from firms hiring those inventors. Where retention is high, the acquiring firm accounts for more citations. This suggests that employee departures, often seen as a sign of failed integration, can actually serve as conduits for knowledge diffusion.
Evidence from inventors
The third part of our study follows the individual inventors employed by the acquired startups. Departures are quite common: 31% of employees and 23% of inventors leave within a year of the acquisition. The inventors who stay are not simply the stronger researchers to begin with. Before the acquisition, those who would later stay and those who would later leave were patenting at very similar rates.
After the acquisition, their paths diverge sharply. Inventors who remain with the acquiring firm go on to patent substantially more than those who leave (Figure 3). Together with the citation evidence above, this suggests that acquirers invest not only in the technologies they buy but also in the teams behind them, a pattern difficult to reconcile with the deliberate shelving of acquired innovation.
Figure 3 Patenting by inventors who stayed versus left after acquisition
Competition policy implications
These findings complicate simplistic narratives favouring blanket scepticism toward big tech acquisitions. If acquisitions routinely suppressed innovation, post-acquisition citations should fall, not rise. Yet, the data consistently show the opposite: acquired inventions become more influential, more widely cited, and more embedded in follow-on innovation by both the acquirer and competitors.
That said, we identify concentration in patent ownership as a potential concern that deserves further investigation. As acquisitions accumulate in a technology space, big tech’s collective patent share rises steeply. While the authors find little evidence that innovation volume falls, the ownership structure shifts. This may not suppress innovation in the short run and, in fact, our data show it doesn’t, but it suggests that there is potential for longer-term changes to the competitive landscape that make standalone commercialisation more difficult relative to acquisition.
Our findings therefore provide another rationale for emerging regulatory thinking that focuses on patterns of acquisitions rather than individual deals. The 2023 US Merger Guidelines, for example, explicitly recognise serial acquisitions as an area of antitrust concern.
Source : VOXeu
