Last week, Google paid Windsurf $2.4 billion. Not to buy the company, but to license its technology and hire the founders plus their core team. Within days, most Windsurf employees found themselves stranded at a leaderless company until Cognition swooped in with an acquisition. The founders got Google jobs, investors got liquidity, and recent employees got nothing… until Cognition saved the day.
The reverse acquihire is the new normal, where Big Tech treats startups as a talent market, not a product or platform one. It’s a way to gobble up top talent without regulatory scrutiny, and technology without the messy business of actually acquiring companies. For recent rank-and-file employees, it can feel like betrayal.
What Is a Reverse Acquihire?
Traditional acquihires are straightforward: a company buys a startup primarily for its people, often shutting down the product afterward. Think Facebook’s acquisition of dozens of small startups in the 2010s, where the real value was engineering talent, not the apps they’d built.
A reverse acquihire flips this model. Instead of buying the company, Big Tech:
- Pays massive licensing fees to access the startup’s technology
- Hires the founders and key talent directly
- Leaves the corporate structure intact but gutted of leadership and vision
It’s regulatory arbitrage disguised as partnership. The acquirer gets talent and technology while avoiding M&A oversight, antitrust review, and the full cost of acquisition. Meanwhile, employees with stock options, especially unvested ones, often get nothing as no acquisition means no equity conversion, no payout, no upside.
The Numbers Don’t Lie
This isn’t anecdotal anymore. In 2024 alone, we’ve seen at least five major reverse acquihires worth over $8 billion in licensing fees, more than many traditional acquisition years. Compare this to the 47 AI startup acquisitions tracked by CB Insights in 2023, and a pattern emerges: Big Tech is choosing licensing over buying at an unprecedented rate.
The regulatory environment has made this shift almost inevitable. With increased antitrust scrutiny on major tech acquisitions (remember the FTC’s challenge to Microsoft’s Activision deal), licensing offers a cleaner path to the same strategic goals.
Case Studies
Character AI
- The Deal: Google licensed Character.AI’s models for $2.5–2.7B and rehired co-founders Noam Shazeer and Daniel De Freitas.
- The Outcome: Unlike other deals, Character AI used licensing proceeds to buy out investors and give remaining employees both one-time cash payouts and equity in the restructured company. Good on them! Employees who stayed now control the company through a cooperative structure. Strange outcome, but decent for employees.
Inflection AI
- The Deal: Microsoft paid $650M to license Inflection’s models and hired some key employees, including founders Mustafa Suleyman and Karén Simonyan. Employees likely received nothing given the $4B valuation hurdle from a previous round.
- The Outcome: Investors got liquidity. The company was effectively dissolved. Its core team now leads Microsoft’s consumer AI efforts.
Adept AI
- The Deal: Amazon hired Adept’s co-founders and several top team members for around $25M in licensing fees, which means employees nor investors truly made money on this because the last valuation was closer to $1B.
- The Outcome: Of 60+ employees, only about 20 remained. Sucks to be one of those 20. The remaining company pivoted to enterprise tools built on open-source models, a far cry from its original AGI ambitions. There’s no public indication that employees received meaningful payouts.
Windsurf
- The Deal: Google paid $2.4B for licensing rights to Windsurf’s AI coding platform and hired founders plus the core engineering team.
- The Outcome: The deal left most employees in limbo until Cognition acquired the remaining company within days. Late-stage investors saw returns, decent but not the venture-scale outcomes they typically target. Recent employees who joined in the final funding rounds reportedly received no payout from Google’s licensing deal, despite the company’s apparent “success.”
Scale AI
- The Deal: Meta took a ~49% stake at a $29B valuation with a $15B investment, while CEO Alexandr Wang moved to lead Meta’s new AI lab.
- The Outcome: This hybrid approach—part investment, part talent acquisition—caused Google to sever ties as Scale’s largest customer. It’s unclear how recent rank-and-file full-time employees with stock options were affected but I’ve heard through the grapevine that recent employees didn’t receive a payout. Who knows what’s going to happen to Scale without the leaders and founder. They apparently have cash but who knows how that’s structured.
Why Big Tech Prefers This Model
The economics are compelling for acquirers:
- Regulatory Advantages: No HSR filing, no antitrust review, no lengthy approval process. A licensing deal can close in weeks, not months.
- Cost Efficiency: Why pay $10B to acquire a company when you can get the technology for $2B and hire the top 20% of talent?
- Talent Arbitrage: In a market where top AI researchers command multimillion dollar pay packages, paying $2B to access and hire the best talent from multiple startups might not be a bad deal.
- Risk Mitigation: If the technology doesn’t pan out, there’s no massive write-down. The licensing fee is just an R&D expense.
- Competitive Moats: Absorbing the most promising startups’ leadership and technology prevents them from becoming future competitors.
The Anti-Competitive Implications
This trend represents a new form of market concentration that existing antitrust frameworks struggle to address. When Big Tech can effectively neutralize startup competition through licensing deals, the innovation ecosystem begins to break down:
- Startup Capture: The most promising companies never reach maturity as independent competitors.
- Talent Concentration: The best AI researchers increasingly flow toward the same five companies.
- Innovation Stagnation: When the best outcome for a startup is absorption into an incumbent, founders optimize for acquirability.
- Venture Misalignment: Some VCs may begin to prefer quick licensing exits over longer, riskier paths.
Economic Impact: The Numbers Behind the Trend
Based on available data from 2024 and 2025:
- The scale of reverse acquihire licensing deals now totals well over $5B, challenging the total value of standard AI startup acquisitions in recent years.
- ~300 employees moved from startups to Big Tech through reverse acquihires
- ~150 employees left behind at restructured companies
The math is stark: Big Tech is paying more for technology and talent while avoiding the costs and complications of full acquisitions.
What This Means for the Startup Ecosystem
The reverse acquihire trend is changing fundamental assumptions about startup risk and reward:
- For Founders: The path to a “successful exit” increasingly means returning to Big Tech with better titles and comp, but without building lasting companies. Importantly, all of these founders were rewarded for raising large amounts of capital at sky-high valuations, which is going to beget more epic fundraises.
- For Employees: Equity becomes even more lottery-like, with many outcomes yielding no payout despite a “win.”
- For Investors: Returns are still often very high for early stage investors but more predictable but flatter for late-stage. Definitely a new category of exit to underwrite against.
- For Innovation: The pipeline to independent competition is being dismantled.
Solutions: Fixing the Broken Social Contract
The current trajectory isn’t sustainable. But instead of cracking down on acquisitions, regulators should recognize that licensing + talent grab is often worse. Acquisitions at least come with transparency, equity conversion, and closure. A norm of founders bailing on their companies for big pay packages elsewhere is kind of sad.
Cultural & Market-Based Shifts:
- Employee Equity Awareness: Teams should understand that unless there’s a formal acquisition or IPO, options may never be worth anything.
- Founder Accountability: Founders should be expected to consider employee equity outcomes, not just investor returns, when exiting. Employees put their trust in you when they join your companies.
- Employee Participation: Companies could tie equity vesting to licensing events above certain thresholds or structure rev-share models for licensing proceeds similar to Character AI.
Industry Examples of Resistance
Not every company has embraced this model:
- Anthropic has taken strategic investments without giving up control or team.
- OpenAI‘s nonprofit structure complicates standard acquihires.
- Hugging Face has stayed independent while building a sustainable business.
These examples show that independence is still viable, but it requires intention (or structure in the case of OpenAI), capital discipline, and a long view.
The Path Forward
The reverse acquihire represents a fundamental shift in how innovation moves from startups to incumbents. While it offers Big Tech efficient access to talent and technology, it threatens the very incentives that make the startup ecosystem work.
The solution isn’t to ban these deals, they serve legitimate purposes and may be the best outcome for some struggling startups. But we need cultural and structural norms that preserve fairness:
- Recognize reverse acquihires as strategic acquisitions in spirit
- Reward teams, not just founders and investors
- Keep the startup dream worth chasing
Because when employees take risk and founders take all the upside, the startup contract is broken. And that’s not a system anyone should be proud to defend.
