Big Tech’s AI Shopping Spree

Last year, SOMO published the Big Tech M&A Tracker, showing that Big Tech companies acquired at least 191 companies between January 1, 2019 and December 31, 2024. This update extends the picture to include acquisitions announced or completed in 2025. Notably, AI defined the battleground in 2025: Big Tech didn’t just buy AI startups, it poached top talent, licensed innovative technologies, and consolidated control over key parts of the AI stack.

Big Tech M&A rebounds, with AI at the centre

After a sharp dip in 2023, mergers and acquisitions activity among Big Tech firms staged a notable comeback. Average annual acquisitions had plunged from about 48 in 2019 to just 13 in 2023, edging up to 19 in 2024. This was likely the result of high interest rates and tougher merger enforcement in the United States under the Biden administration. However, 2025 data shows that the increase in 2024 was not an outlier but a starting point of shifting waves. M&A activity re-accelerated(opens in new window) across the global economy, and Big Tech followed suit. In 2025, Big Tech completed at least 25 acquisitions, more than a 30 per cent increase compared to the previous year.

While Big Tech pursued acquisitions across a range of sectors, none rivalled AI in scale. Fifteen of the 25 announced deals targeted companies operating in AI-related markets, including chips, data centres and cloud infrastructure. The signal is clear: Big Tech companies are positioning themselves not only as drivers of AI innovation but also as key actors in its commercial future, which could allow them to control and appropriate(opens in new window) the advances of AI innovation.

Sidestepping merger control

Notably, Big Tech isn’t always buying companies outright anymore; and that is no coincidence. Rather, they are increasingly turning to multi-billion-dollar “strategic” investments and alternative arrangements that can result in significant consequences without triggering traditional merger scrutiny. High-profile examples include Microsoft’s investment(opens in new window) of $13 billion in OpenAI, which gives it a roughly 27 per cent stake(opens in new window) in OpenAI’s for-profit arm; Meta’s acquisition of a 49 per cent stake in ScaleAI; and Google’s $3 billion investment(opens in new window) that reportedly(opens in new window) secured it a 14 per cent ownership stake in Anthropic.

In parallel, Big Tech firms have frequently opted to hire key personnel and license core technologies from target companies instead of acquiring them outright. The UK’s competition authority (CMA) has warned that(opens in new window) some of these partnerships and investments “may have been structured to seek to avoid [merger control rules].”

 On paper, these agreements can look less threatening than full acquisitions. But in practice, their impacts can be similar. Licensing agreements and aggressive talent recruitment can take the independence of the target firm away or diminish its capability to compete and innovate. Such practices could entrench Big Tech’s power while weakening potential future competitors. As these transactions are usually opaque, often shielded behind contractual terms, they are excluded from the updated Big Tech M&A Tracker unless they have been reported under the Digital Markets Act.

Nevertheless, in addition to the 6 other mergers extracted from other databases, the tracker includes all 19 acquisitions (excluding joint ventures) reported by Big Tech companies under the Digital Markets Act in 2025. These include acquisitions that do not involve the outright acquisition of target companies and therefore escape merger scrutiny by the European Commission. For example, Google claimed(opens in new window) that, in six out of seven acquisitions it reported under the Digital Markets Act in 2025, the Commission lacked jurisdiction to review the transactions under the EU Merger Regulation

These tactics raise serious concerns. They suggest that Big Tech firms can work around regulation and leverage alternative dealmaking to bypass EU merger control. As a result, these expose the underlying limits of the current enforcement frameworks when it comes to current forms of market consolidation.

Acquired…and then discontinued 

After the deals are done, there is often silence. Our previous analysis found that the websites of nearly 67 per cent of the companies acquired by Big Tech between 2019 and 2025 were not accessible anymore. In 2025 alone, this happened to seven companies. This lower number may simply reflect how recent some of these acquisitions happened.

More strikingly, five of the 25 target companies discontinued their services partly or entirely after being bought by Big Tech in 2025. This pattern of post-acquisition discontinuation and loss of autonomy suggests that the cumulative impact of acquisitions may exceed their individual outcomes. This highlights the need for ongoing post-acquisition monitoring to identify when such transactions risk weakening innovation and competitive dynamics in a diverse economy.

Four Big Tech acquisitions that stood out in 2025

Big Tech’s 2025 acquisition spree reveals a shift in tactics: from directly buying companies to acquiring key assets and innovation. The consequences extend beyond market competition to broader issues such as data privacy, governance and geopolitics. While each of the 25 acquisitions is interesting in its own way, four acquisitions stood out for how they illustrate these evolving tactics:

1. Google/Windsurf
   
   In simplified terms, Windsurf can be described as the ChatGPT for software engineers that received lots of attention across the tech world, helping developers write code. In April 2025, it was reported(opens in new window) that OpenAI was in talks to acquire Windsurf for $3 billion. But the deal collapsed. Reportedly(opens in new window) , OpenAI’s previous partnership agreement with Microsoft required the company to share intellectual property around technology as part of any acquisition, and Microsoft did not accept an exception for Windsurf.
   
   Two months later, Google hired(opens in new window) top executives and key talent from Windsurf and acquired a non-exclusive license to Windsurf’s technology, reportedly(opens in new window) for a staggering $2.4 billion fee. What remained of Windsurf was acquired by Cognition for an estimated(opens in new window) $250 million. If acquiring control over key elements of Windsurf could be obtained for $250 million, why would Google pay nearly ten times that amount for just a subset of its assets? This substantial price disparity raises broader questions about what constitutes an “acquisition” and how control over strategic assets in AI development may be transferred outside legal frameworks.
2. Amazon/Bee
   
   In July 2025, Amazon acquired(opens in new window) Bluush Inc., the company behind Bee, which is the maker of an AI-powered bracelet designed to listen to, record, and transcribe conversations occurring around the user, unless the device is turned off. This raises serious questions around privacy: devices that continuously capture ambient speech can collect data on bystanders who may well be unaware they are being recorded at all. Amazon’s possible plan to integrate(opens in new window) Bee’s capabilities into its virtual assistant Alexa further amplifies these concerns by combining pervasive data collection with one of the world’s largest consumer data platforms.
3. Meta/Manus
   
   Days before the close of 2025, Meta announced(opens in new window) the acquisition of the AI startup Butterfly Effect for a reported(opens in new window) price of over $2 billion. The company is best known for “Manus,” a so-called general AI agent(opens in new window) designed to complete complex tasks with minimal human oversight. But the strategic significance of the deal lies less in absorbing a singular product than in Meta’s ability to absorb AI talent.
   
   Butterfly Effect was founded(opens in new window) by Chinese entrepreneurs in Beijing in 2022 before relocating to Singapore, making the deal geopolitically sensitive. Meta’s acquisition attracted scrutiny from Chinese authorities, leading its Ministry of Commerce to open an investigation(opens in new window) into whether tech export or national security regulations(opens in new window) were violated. This case illustrates how Big Tech acquisitions can become flashpoints not only for market competition, but also for geopolitics and the global race for AI talent.
4. Google/Wiz
   
   With its $32 billion price tag(opens in new window) , Google’s acquisition of Wiz was the biggest Big Tech acquisition of 2025, and also the biggest(opens in new window) in Google’s M&A history, surpassing the acquisition of Motorola Mobility for $12.5 billion in 2012. Already in 2024, Google had reportedly(opens in new window) attempted to acquire Wiz – the deal failed, seemingly(opens in new window) due to antitrust concerns in the United States. The regulatory landscape appears to have shifted since.
   
   The price tag alone demands attention. Wiz claimed(opens in new window) in 2022 to be the “fastest-growing software company ever”, reaching $100 million in annual recurring revenue – from subscriptions and long-term contracts – in just 18 months. Its clients include(opens in new window) several major European corporations, including Siemens, BMW, and LVMH. Wiz’s rapid growth and the scale of the transaction have raised concerns(opens in new window) , including alleged data leakage issues voiced by a group of European cloud service providers(opens in new window) .
   
   Given the scale of the transaction and its potential impact on competition, the merger is subject to approval by the European Commission. If Google acquires Wiz, it would give an already powerful company even more control over the digital infrastructure that European businesses rely on. This acquisition could result in Google absorbing a fast-growing, multi-cloud platform into a single cloud ecosystem, thereby risking reduced choice and diversity. Beyond competition concerns, the deal also raises a geopolitical issue: it could further entrench a US-based tech giant in critical digital infrastructure in Europe. Given the importance of diversity and alternative supply channels for economic resilience, European regulators should actively prevent the emergence of additional dependencies, especially in critical sectors like cloud computing.
   
   It is deeply concerning that recent reports(opens in new window) suggest that Google’s proposed acquisition of Wiz may be cleared by the European Commission without an in-depth investigation. If confirmed, this would further illustrate how existing enforcement tools are failing to address the structural risks of consolidation at the top of the digital economy, while allowing the same incumbent companies to entrench their power further.

The big picture of 2025’s Big Tech mergers and acquisitions raises alarm bells. Aside from the joint ventures, of the 25 acquisitions in 2025, only one, Google’s takeover of Wiz, was notified to the European Commission for approval. Every other acquisition failed to become subject to EU merger control. This is very troubling. It shows the persistent inadequacy of current legal frameworks in addressing acquisitions involving innovative yet comparatively small companies. It also underscores the need for an effective and unified legal framework across the EU.