AI M&A This Year Has Been 🔥

What lessons can we glean from a banner H1 in deal activity

At the start of the year, one of our core predictions was that 2025 would see a surge in M&A activity across the generative AI landscape — and so far, that’s playing out faster than even we expected.

There were two types of M&A moves we anticipated:

1. Acquihires for tech and talent: With companies like OpenAI, Anthropic, Meta, Databricks, and Salesforce sitting on deep war chests and hungry for technical leverage, it’s no surprise they’re turning to M&A as a way to shortcut the challenge of recruiting top-tier AI talent. These players are already offering multi-million dollar compensation packages for top researchers and engineers — and in many cases, it’s more efficient to buy the whole team than hire them one by one. Just look at Meta’s reported $15B “investment” in Scale AI for 49% of the company, a nice way to poach Alexandr Wang and top talent without formally acquiring the business.

2. Strategic AI add-ons by later-stage/pre-IPO players: We also expected to see gen-enhanced, pre-IPO companies snapping up smaller gen-native startups to strengthen their AI narratives ahead of IPOs. These deals aren't just about features — they’re about positioning: showing the Street a credible, cohesive AI strategy.

So what’s actually happened?

---

The First Half of 2025: AI M&A By the Numbers

According to PitchBook, from January 1 to June 4, 2025, 365 M&A deals in the Artificial Intelligence and Generative Artificial Intelligence category that have been completed, representing over $10B in total deal value.

While large-scale M&A remains relatively limited due to elevated interest rates, regulatory pressure, and public market volatility — we have seen a number of smaller, strategic acquisitions gain momentum. With lower startup valuations, ongoing funding challenges, and a growing need for liquidity, conditions are ripe for early-stage M&A, and we expect that trend to continue through the rest of the year.

As shown in the above charts, despite the regulatory headwinds, 2025 has brought a noticeable pickup in M&A from Big Tech — with deals like Google’s $32B acquisition of Wiz signaling renewed appetite from the big tech players (compared against the lack of activity in 2024). We're also seeing more activity from mid- to late-stage, non-IPO companies, many of which are taking advantage of lower valuations and fewer regulatory constraints to grow inorganically.

Some of the key trends we saw among those deals:

• A wave of recent vintage startups getting scooped up. Many of these companies were founded in the last few years following the GenAI boom. We're seeing that a number of these AI startups are finding homes within larger platforms.

• The usual suspects are the main acquirers. OpenAI, Databricks, Google, Salesforce, and others are doing much of the buying. In many cases, they’re not just buying product — they’re buying velocity. These larger players need to build and ship at a faster pace than ever before with more pressure from the early-stage companies. There has not been as much activity from the pre-IPO companies as we expected, perhaps partially driven by the fact that they are also trying to manage their own cash burn and growth before acquiring and dealing with integrating.

• There’s action at both ends of the spectrum. From acquihires of early-stage teams to $1B+ deals like Windsurf and Weights & Biases, the appetite spans all stages. CoreWeave’s acquisition of Weights & Biases for example was a bold move to own the full AI development stack, combining high-performance compute with the go-to platform for model tracking, tuning, and evaluation, and signaling a shift toward more vertically integrated AI infrastructure.

A few notable deals:

• Large/Later-Stage Acquisitions: We are seeing the leading platforms racing to own critical layers of the stack — from data security (Wiz) to experimentation (Eppo), model ops (Weights & Biases), and software development workflows (Windsurf) — in a bid to power and differentiate the next generation of AI applications.

  • Windsurf → in talks to be acquired by OpenAI (reported at $3B)

  • Weights & Biases → acquired by CoreWeave for $1.7B

  • Eppo → acquired by Datadog for $220M

  • Wiz → acquired by Google for $32B

• Software Infrastructure / Dev Tools: There is a strong push to streamline and strengthen the developer workflow — from data infrastructure (Neon), to collaboration and testing (Grit), to capturing and leveraging engineering knowledge (Augmend) — as AI-native and software-first companies double down on tools that boost developer velocity and system observability.

  • Neon → acquired by Databricks

  • Grit → acquired by Honeycomb

  • Augmend → acquired by Datadog

• Data & Analytics Layer: These deals reflect a growing focus on enhancing the data layer — with acquirers targeting startups that improve data quality, labeling, infrastructure, and accessibility — all critical building blocks for making enterprise AI more reliable, efficient, and actionable.

  • Numbers Station → acquired by Alation

  • Seek → acquired by IBM

  • Refuel → acquired by Together AI

  • Arcus → acquired by Addepar

• Vertical / App Layer: There is a broader trend of incumbents wanting to enhance the AI application layer to bring more intelligence, usability, and domain-specific automation to core user workflows.

  • Moveworks (AI for ITSM) → acquired by ServiceNow for $2.9B

  • Galileo (AI UX) → acquired by Google

  • Moonhub (AI Recruiting) → acquired by Salesforce

What does this mean for the ecosystem?

This flurry of activity raises important questions about how the AI landscape and the startup journey is evolving. A few early hypotheses:

• Founders may feel more permission to “swing” — knowing that strong teams and product velocity can lead to a quick, clean outcome even if they don’t IPO.

• Conversely, some teams may sell earlier than they would have historically.
  With strategic acquirers offering compelling packages — and a tougher funding market outside the top 5% — we may see more companies opt for the “build to get bought” path.

• Talent is getting priced in earlier. In some ways, the market is functioning more like the NBA draft — with incumbents picking up promising teams before their breakout season. Just look at OpenAI’s acquisition of Jony Ive’s AI devices startup for $6.4B!

• Fewer standalone giants may emerge. There’s no shortage of talent or ambition — but with so many well-capitalized incumbents aggressively acquiring and building out their AI capabilities, it may be harder for companies to stay independent long enough to reach massive scale. Especially if they start to see that one of their competitors is emerging at a much faster scale.

---

Conclusion

We're only halfway through the year, but the pace of AI M&A is already reshaping the landscape. While some see this as a sign of consolidation or saturation, we see it as an indicator of just how critical AI has become to every major platform’s roadmap. This is a moment to be both opportunistic and strategic. Great outcomes are on the table, but navigating to them requires clarity.

We’re also keeping a close eye on the tech IPO window — if it reopens, it could significantly shift M&A dynamics. Much of this will hinge on how broader market conditions and macro trends unfold in the coming months!