Why build when you can buy? Inside the M&A frenzy fueling acqui-hires, infra plays, and IP grabs.
A decade ago, when AI wasn’t yet a buzzword, businesses struggled to raise money for the technology. Demis Hassabis, the CEO and co-founder of DeepMind — one of the pioneers in solving artificial intelligence — recently said in a podcast that he had to go a couple of years without paying himself because cash was so tight when they were just starting.
Today, we’re seeing an entirely different world.
Almost every company has some sort of AI play, and valuations have gone wild, whether in acquisitions or funding rounds. What DeepMind once raised for its entire seed round is now what some AI interns make in a year.
If we look at the numbers, the shift appears nothing short of dramatic. Global M&A alone in AI has exploded, clocking 454 deals worth $49.9 billion in 2024, with $55.3 billion already earmarked across 240 deals this year. Meanwhile, funding in Q2 2025 reached an impressive $47.3 billion across 1,403 deals.
Some of the biggest headline-makers in the category are either deep-pocketed big tech giants, such as Google and Meta, or those sitting right at the forefront of AI technology, i.e., players like OpenAI, xAI, and Anthropic.
Big tech players are largely acquiring or acqui-hiring young, niche startups with very promising AI products (not just foundation models) at high valuations.
For instance, Salesforce acquired Convergence, an AI startup that built an agent that beat OpenAI’s Operator, and brought in the team of Moonhub, known for building AI that vets and recruits talent. Similarly, Google spent $2.4B to license core tech from AI coding giant Windsurf, whereas Meta spent $14.3B to take a 49% stake in Scale AI and poach its CEO, Alexandr Wang, and acquired AI startups PlayAI and WaveForms.
On the other hand, frontier AI companies, which (mostly) have mature products and are way ahead in the race, are raising money at massive valuations to scale their products, expand infrastructure, and take on the incumbents, including the ones mentioned above.
The best examples here are OpenAI and Anthropic. The former is now reportedly raising money at a $500B valuation, while the latter is said to be in talks to raise a very oversubscribed round at a $170B valuation. These would make them some of the highest private valuations in venture history. Even Elon Musk’s xAI is in this race, with its last public valuation standing at $50B.
In addition to these, there are also some ‘would-be’ frontier AI startups, which are still in stealth with no public product or business, raising massive rounds and super-massive valuations purely based on their talent and tech promise. Examples include Mira Murati’s Thinking Machines Lab, valued at $12B, and Ilya Sutskever’s Safe Superintelligence Inc., valued at $32B.
What’s the motivation behind big tech’s AI splurge?
Many tech giants already operate multiple product lines and want to double down on AI to improve them. For them, the key reason behind their massive M&A splurge appears to be the need to have a powerful, already-proven IP and team — without spending time and resources on internal R&D and building an equally successful product from scratch.
“AI moves so quickly with things changing over weeks. Even if a company like Google can build a Windsurf again, an executive at Google has to take the bet that in 6 months of development, it will match the scale of the growing, faster-moving startup — and have to build up the brand too. That’s a tough bet for a big company to make. If you fail at that stage and Windsurf succeeds, there’s no way to come back instead of paying an even higher price,” Deedy Das, principal at Menlo Ventures and a close watcher of the AI domain, told Future Nexus.
Beyond the time and traction benefit, acquisitions also help teams with data and operational excellence. Per Das, the data that startups like Windsurf have collected could be useful to train future models of the buyers. Plus, their teams are capable of moving fast and have already developed an intuition for building in the space.
“It’s not like Google has bad engineers, so it’s not talent motivated (the Google-Windsurf deal), but they are likely not educated in the space and not moving as fast as a startup,” he pointed out while emphasizing each situation is different and sometimes, it is about talent too, like in the cases of Google-Character AI, Meta-Scale AI, and Salesforce-Moonhub.
Another big tech company that may soon join the list of players snapping up powerful niche startups is Apple. It has long lagged behind competitors in the AI space, with internal teams failing to deliver a revamped Siri experience. But now, CEO Tim Cook recently told analysts that the company is open to potential acquisitions to accelerate its roadmap.
As angel investor Sandeep Kondury told Future Nexus, many tech giants are even looking at acquisitions to own the AI narrative. Successful startups today have the potential to shape the AI language of the future — think of coining terms like Copilot and Agents. So, if a company looks at an agentic AI startup building an ‘AI employee’ or ‘AI marketer,’ they are not just eyeing the tech and talent; they also want the lingo leverage to define how users think about the product and win attention in a noisy market.
“Done right, M&A can boost a public company’s image, innovation credentials, and ultimately, its market cap. Done poorly, it signals a lack of foresight, and investors take notice. In a world where perception moves faster than execution, acquiring the narrative can be just as critical as acquiring the tech,” he noted.
Lastly, some of the big tech acquisitions are also being driven by the need to control the AI stack, leading to a consolidation. The best example here is Salesforce shelling out $8B for enterprise data company Informatica, strengthening the data management capabilities underpinning its CRM’s Agentforce offering. Another notable deal was Coreweave spending $9B to acquire Core Scientific to expand its GPU-heavy AI data centers.
“The most aggressive consolidation is happening at the infrastructure and orchestration layers, especially around data pipelines, model serving, and agent frameworks. Incumbents are racing to own the rails on which all AI applications will run. Whoever defines how agents are deployed, coordinated, and trusted will shape the next era of computing,” Kondury added.
What’s in it for startups?
As for startups on the other side, accepting acquisition offers from tech giants during peak growth periods affords them instant liquidity in the rapidly evolving AI market.
In Das’s words, the math is simple: If a $10M ARR business valued at $1B is approached by a reputable acquirer like Google, that valuation (100x revenue) already accounts for years of flawless growth. Using the classic 3-3-2-2-2 growth pattern, the startup would otherwise need to hit $30M, $60M, and then $120M ARR over the next three years just to justify today’s valuation in the public market. If they sell now, they get the money immediately and can invest it elsewhere, skipping the grind of those years.
He added that many founders take this path when they’ve “stopped believing in the business or lost the energy to keep going,” possibly due to massive infrastructure/capital spending or scalability challenges — both of which big buyers can easily help solve.
While acquisitions by big tech would provide liquidity to promising AI startups, those looking to stand their ground (except big names at the frontier or those with promising talent) should expect a capital ‘chill’ of sorts, with investors potentially eying breakouts in specific categories instead of generalists.
“Every industry, vertical, and function has been touched by AI. This kind of universal disruption is rare, and it signals that we’re now past the exploration phase. Investors will shift their lens from “Is this AI?” to “Is this the best version of AI in this market?” or even, “Does this startup own a market that wasn’t obvious before?” Kondury said.
On top of this, the case of central banks making cash scarce would mean assets will have to fight to attract capital.
“The startups must prove not just their tech, but their defensibility, narrative control, and ability to stay ahead of incumbents who are already repositioning around AI…So, yes, funding would tighten. But the next wave of AI capital will flow to those who create language, define space, and lead behavior, not just build tools.”