Killer Acquisitions: Why Regulators Now Watch Big Tech’s Smallest Deals

Killer acquisitions and big-tech merger scrutiny concern the hardest question in merger control: how to police deals whose harm lies in a future that will never now happen. A large platform buys a small company with no revenue, modest users, but a trajectory — and the counterfactual world where that company became a rival dies quietly at closing. This guide covers the theory, the evidence, the rebuilt jurisdictional toolkit and the enforcement cases — part of the digital-markets pillar of our Competition & Antitrust hub.

What is the evidence that killer acquisitions are real?

The label came from pharmaceutical economics: the Cunningham, Ederer and Ma study found roughly 6% of pharma acquisitions targeted overlapping pipeline drugs whose development was then disproportionately discontinued — projects killed to protect the acquirer’s franchise. Digital-market evidence is structurally harder (no clinical-trial registry for products never built), but the pattern inputs are documented: hundreds of acquisitions by the major platforms over two decades, near-zero of which were blocked, several of which — Instagram, WhatsApp, YouTube, DoubleClick — became the pillars of today’s dominance debates.

The intellectual correction was as important as the evidence: merger review’s traditional question (“does this deal lessen competition that exists?”) systematically undervalues potential competition, because the probability-weighted future rival is worth more to consumers than its expected value suggests when concentration is already high. That reframing — protect the competitive option value, not just current overlap — now appears in agency guidelines, the Adobe/Figma abandonment, and the FTC’s retrospective Meta case seeking to unwind Instagram and WhatsApp a decade later.

How did jurisdictions close the threshold gap?

By adding non-turnover hooks. Deal-value thresholds: Germany (€400 million) and Austria (€200 million) notify high-price/low-revenue deals with local nexus; India followed (INR 20 billion with substantial Indian operations). Türkiye’s technology-undertaking exception disapplies the target-side local turnover threshold entirely for tech acquisitions touching the Turkish market — one of the world’s most aggressive designs, preserved in the 2026 threshold revision, as detailed in our thresholds guide.

The UK share-of-supply test needs no target revenue at all — the CMA reviewed Facebook/Giphy (and unwound it) on that basis. Designated-firm reporting: DMA Article 14 obliges gatekeepers to inform the Commission of all digital and data-related acquisitions regardless of size, and UK SMS firms carry mandatory reporting; the information gap that let a hundred quiet deals close unexamined is gone for the designated groups. And call-in powers spread across EU member states (Italy, Ireland, Denmark and others) after the Court of Justice’s Illumina judgment ended the practice of Article 22 referrals from states with no jurisdiction — restoring a below-threshold path to Brussels, now routed through national law.

What theories of harm do agencies apply to these deals?

Four, increasingly explicit. Loss of potential competition: the target would plausibly have entered the acquirer’s market (Adobe/Figma’s product-design trajectory; the mooted Visa/Plaid entry theory that ended that deal). Loss of dynamic competition: even absent entry, an independent target would have pushed innovation the merged entity can now pace. Ecosystem entrenchment: the deal adds data, users or capabilities that deepen the moat around the core platform (the CMA’s framework in mobile and the Commission’s in Google/Fitbit, cleared only with data-separation commitments).

Reverse killer acquisitions complete the set: the acquirer, not the target, stops building — buying instead of competing, with its own entry plans shelved (the concern raised around big-platform purchases of would-have-built capabilities). Evidence in all four is document-driven: the acquirer’s internal threat assessments, build-versus-buy analyses and premium justifications are the case. Deal teams should assume every valuation memo will be read by an agency asking one question: what exactly is being paid for, if not the elimination of a future competitor?

Are AI deals the new killer-acquisition battleground?

Yes — with a twist: the deals are structured to avoid being deals. Multi-billion cloud-and-equity partnerships (Microsoft/OpenAI, Amazon and Google with Anthropic), acqui-hires lifting founders and teams while licensing the technology (Microsoft/Inflection, and successors), and compute-for-rights arrangements all transfer competitive control without a classic acquisition of control. Agencies probed the perimeter: the CMA examined Microsoft/OpenAI and the Inflection hiring under its flexible jurisdiction (reviewing the latter as a merger and clearing it), the EU concluded partnership structures fell outside the EUMR as constituted, and inquiries into cloud-AI concentration continue on both sides of the Atlantic.

The policy question is whether control-based merger law can see influence exercised through compute dependency, exclusive licensing and board observation — or whether the quasi-merger structures require new hooks, as several authorities have proposed. Meanwhile the classic toolkit adapts: Türkiye’s tech exception and deal-value thresholds catch AI targets’ acquisitions where structured conventionally, and gatekeeper reporting surfaces the rest. The sector’s full competition picture — chips, compute, data, talent — is analysed in our AI and competition law guide.

How should acquirers and targets navigate the new landscape?

Acquirers: run the potential-competition screen on themselves before agencies do — what do internal documents say about the target’s trajectory and the build alternative? Structure timelines for call-in and referral risk (windows in some regimes run months after closing announcements); and for designated firms, treat the reporting duty as a strategic disclosure decision, since reported deals invite early engagement that can be shaped. Premium narratives need honest work: synergy stories that cannot explain the price feed the elimination inference.

Targets and their boards: preserve independence evidence (growth plans, entry roadmaps) with care — the same documents that justify valuation can kill clearance; negotiate break fees and interim covenants against the elongated, riskier review path; and consider the counterfactual seriously, because “we would have struggled alone” is now a litigated question with the company’s documents as exhibits. Both sides should read the cross-border filing playbook with one addition: in tech, the filing map includes jurisdictions with no thresholds at all — only judgment.

What does the enforcement scorecard actually show?

A transformed deterrence picture with few litigated wins. Formal below-threshold blocks remain rare; the policy’s force operates upstream — deals not attempted, structures redesigned, partnerships substituted for acquisitions. The visible record: Giphy unwound, iRobot and Figma abandoned under pressure, Within survived the FTC’s potential-competition challenge (a court loss that disciplined the theory), Fitbit and Activision cleared with heavy commitments, and the Meta retrospective case pending as the doctrine’s high-stakes test.

Read honestly, the scorecard counsels neither complacency nor panic: agencies lose potential-competition cases when trajectories are speculative (Within), and win leverage when documents show threat-elimination logic (Figma’s abandonment followed exactly such disclosure). The controlling variable is the evidentiary record the parties themselves created — which returns every planning conversation to document discipline, honest premium narratives and early counterfactual analysis. Deals with clean stories still close; deals whose own files tell the elimination story do not.

How should competition-conscious boards govern acquisition pipelines now?

With the regulatory lens embedded at pipeline stage, not signing. Concretely: standing counsel review of any target in an adjacent or nascent space before term sheets; document-hygiene training for corp-dev teams (threat language in sourcing memos is discoverable years later); a build-vs-buy record maintained honestly, since its absence reads as elimination logic; and, for designated or near-designated firms, a reporting-and-engagement strategy that treats early regulator contact as risk management rather than concession.

Boards should also stress-test the portfolio retrospectively: which past acquisitions would today’s framework challenge, what integration choices deepen or reduce unwinding exposure, and where partnership structures now under scrutiny need contractual contingency. The Meta case’s core lesson is temporal — merger risk no longer expires at closing — and governance that prices a decade of hindsight into today’s deals is the only durable response.

Does the doctrine chill beneficial acquisitions?

The serious counter-argument, and it deserves weight: acquisition is the dominant exit for venture-backed innovation, and pricing regulatory risk into exits raises the cost of capital for exactly the nascent firms the policy means to protect. Studies of post-tightening deal flow show reallocation more than reduction — more private-equity and cross-sector buyers, more partnerships, longer independent runways — but the welfare accounting remains genuinely contested.

Policymakers’ answer has been calibration rather than retreat: safe harbours below meaningful size, focus on the designated few, and effects-based review rather than per-se hostility. For practitioners the debate’s practical residue is argumentative — efficiency and exit-ecosystem evidence now belongs in every defence file, and authorities do engage with it when it is quantified rather than invoked.

Frequently Asked Questions