A deep dive into the most complex AI acquisition saga of 2025, featuring deceleration, empty husks, and the deal of the year.
Top Take-Aways
Harry (20VC): The revenue deceleration story tells you everything about why founders are choosing safety over upside. When you go from $100M ARR to $82M in just 90 days after a failed OpenAI deal, you grab the first lily pad available. The real revelation? These “empty husk” structures are becoming the new norm for avoiding FTC scrutiny, but they’re creating unprecedented complexity in M&A.
Jason (SaaStr): After 80 hours of vibe coding this weekend, I’m convinced “roll your own SaaS” is complete fraud. You can’t build commercial-grade apps for $20/month – I burned $600 in 6 days getting 10% of the way there. That doesn’t mean vibe coding isn’t incredibly exciting, however. And Cognition just pulled off the deal of the year: acquiring $82M in revenue, $100M in cash, and elite talent for essentially $220M. That’s genius-level M&A.
Rory (Scale): The most sobering part isn’t the deal structure – it’s what this reveals about founder psychology. Here’s a company that was about to raise at $3B pre just months ago, and the founder didn’t pause for a single second before grabbing $2.6B and running for safety. That tells you everything about how even the most successful founders view the upside of independence versus the safety of a trillion-dollar balance sheet.
The Windsurf Saga: From Rocket Ship to Fire Sale
In what can only be described as an Agatha Christie mystery of Silicon Valley, Windsurf went from AI darling to empty husk in 90 days. The cast of characters reads like a who’s who of tech drama: OpenAI as the jilted suitor, Anthropic as the API-cutting villain, Google as the opportunistic acquirer, and Cognition as the white knight rescuer.
The Revenue Deceleration That Changed Everything
The numbers tell a brutal story. Windsurf claimed $100M ARR just 90 days ago, before the OpenAI deal was announced and before they lost access to Claude. By the time Google swooped in, they were doing $82M ARR.
“That’s potentially massive deceleration,” notes Jason Lemkin. “If you’re decelerating like that, the minute the OpenAI deal falls apart, you’ve got to find a lily pad to jump on.”
This deceleration appears to be the key catalyst that forced Windsurf’s hand. As Harry Stebbings observed, “M&A can utterly derail your business. You can go from a rocket ship to desperately rolling those phones to Google, Microsoft, whoever is going to buy this thing.”
The FTC-Driven “Empty Husk” Structure
Google’s acquisition represents the latest evolution in what’s becoming a new M&A playbook designed to avoid FTC scrutiny. Rather than buying the company outright, Google licensed the IP and hired the top 20-30 engineers for approximately $2.6B, leaving behind what Rory describes as “an empty husk.”
The structure required Windsurf to pretend the remaining company was a “viable independent asset” – leading to Friday press releases describing 250 wonderful employees and $100M in revenue continuing “gloriously into the future,” while frantically selling the remnants by Monday.
“This entire podcast is basically the FTC fucking up normal business operations,” Rory noted. “All this weird shit comes and bites you in the ass. It would all be so much simpler if everyone was a standard C corp and if the FTC would get back to being sensible.”
The Cognition Masterstroke
While Google got the team and IP, Cognition pulled off what may be the deal of the year by acquiring the “empty husk” for approximately $220M after taxes. The math is stunning:
- $82M in ARR
- $100M in cash
- 200+ employees
- Restored Anthropic API access (accomplished over a weekend)
“This was a genius move,” said Jason. “Windsurf was hopeless before this deal. You could never attract 30 S-tier developers to fix Windsurf, but they fixed it in 1 hour. Literally in 90 days, Windsurf could be better than it was before this deal.”
The president of Cognition reportedly got a DM on Friday night and closed the deal in 15 minutes. For a company that likely had sub-$10M in revenue, acquiring $82M ARR and $100M in cash represents transformational scale.
The New AI M&A Playbook
This deal represents the fifth transaction using this “licensing + talent acquisition” structure, following similar moves by Meta, Google (Character AI), and others. The pattern is clear:
- Buyer perspective: Acquire talent and IP while avoiding regulatory scrutiny
- Seller perspective: Get liquidity while maintaining plausible deniability about selling the business
- Regulatory arbitrage: Structure transactions to avoid traditional acquisition oversight
But these deals come with hidden costs. Windsurf’s structure resulted in approximately $500M in taxes that went to the federal government instead of shareholders, and created a two-tier outcome where only stockholders (not option holders) participated in the liquidity event.
“I do not believe anyone involved wanted to save 5% of their consideration by sticking it to people,” Jason emphasized, defending the team’s ethics. “There’s just no way. This is the structure of the deal, nothing else.”
The Broader Implications: Safety vs. Upside
Perhaps the most revealing aspect of this saga is what it demonstrates about founder psychology in the AI era. Here was a company positioned to raise at $3B pre-money just months earlier, and the founder chose immediate liquidity over continuing to build independently.
“There’s a lot of revealed preference going on about how founders think about the upside of independence versus the safety of a fucking ton of money,” Rory observed. “It’s pretty sobering if you’re a VC because many founders have opted to say ‘Maybe I’d like that cold, hard cash and the safety of a trillion-dollar balance sheet.’”
This trend extends beyond Windsurf. The podcast hosts noted similar decisions across multiple AI companies, suggesting that even wildly successful, hyped companies are choosing safety over the potential for massive independent outcomes.
The “Roll Your Own SaaS” Fraud
Jason’s weekend of intensive “vibe coding” led to a scathing assessment of the “roll your own SaaS” movement proliferating on social media.
“This is at the edge of fraud,” he declared. “One, I’ve already burned $600 in credits in 6 days to get 10% of the way there. So you can’t do it for 20 bucks a month. Two, none of these things are commercial grade.”
His experience with Replit revealed the current limitations of AI coding tools:
- High orchestration overhead requiring constant attention
- Non-deterministic behavior (the AI “lied” and deleted his entire database)
- Significant ongoing costs (budgeting $100/day for commercial apps)
- Major gap between 80% functionality and commercial-grade quality
“It’s so close. I can see how I’m 80% of the way there to a commercial-grade app in less than a week. But I may never get to 100%. And not for $20.”
Grok’s Stunning Technical Achievement
Elon Musk’s Grok team delivered a massive validation of the “knowledge dissemination” thesis by achieving benchmark performance that destroyed GPT-4 (44.4% vs 26.9% for the next closest competitor) in just two years from a standing start.
“It’s a huge, huge achievement,” Rory noted. “The fact that it’s done makes you wonder about how you think about the valuations of some of these other companies. Maybe the knowledge has disseminated out just a little further.”
The Grok team assembled talent “one or two levels down” from the headline AI researchers, proving that with “committed clear leadership and a couple billion dollars worth of GPU, the next level down people know how to do this too.”
However, questions remain about commercial viability. As Rory pointed out, “Just because they completed the astonishingly hard technical task far better than anyone else did doesn’t mean they get to build a business here.”
The Vibe Coding Investment Thesis
Harry’s $12M investment in Lovable reflects a bold bet on the total addressable market for “vibe coding” tools that enable non-developers to build applications.
“By definition, there are more non-developers than developers,” Harry noted. “If you agree that 100% of developers are going to buy a tool like Cursor, the real question is how big and how sustaining is this movement of non-developers.”
Jason supports this thesis despite experiencing the limitations firsthand: “If Replit and Lovable keep making developers happy and they make dog walker apps people happy, the TAM really is much bigger than Cursor. It could be like 50 or 100 times bigger.”
The key insight: brand recognition becomes critical in this space because non-technical users can’t easily evaluate tools. “There’s no way I’m going to use something for all the time I’m going to put in that just launched on Product Hunt,” Jason noted.
Valuation Perspective: Cheap at Any Price?
Despite seemingly high valuations, the hosts argue that leading AI coding companies may actually be undervalued relative to their growth rates.
“Some of these deals almost seem too cheap,” Jason observed. “Undifferentiated company raising at 300 pre with less than a million of revenue. And then you see what is Windsurf selling for? 20x revenue. Lovable’s raising at 20. Cursor raising at 20.”
Harry’s Lovable investment exemplifies this: “When I did the last round at 200 million price, it was at 4 million in revenue. By the time the deal was closed, it was at 19 million in revenue.”
The Churn Reality Check
The hosts addressed the elephant in the room: AI SaaS companies are experiencing higher churn rates than traditional enterprise software, particularly among experimental users.
“I bet if we took Replit and Lovable and segmented their churn, I don’t just want your headline number,” Jason noted. “You should have segments that are churning like 10 or 20% a month like a low-end consumer app.”
However, they argue this isn’t necessarily problematic if companies can identify and focus on cohorts with strong retention. “If you have one cohort that has insane retention and this low-end part is churning 4% a month, I get it,” Jason explained.
Key Takeaways for B2B Leaders
- Revenue deceleration kills optionality: When growth stalls, even temporarily, founders lose negotiating power and strategic options. Windsurf’s 18% decline in 90 days forced their hand.
- M&A can derail your business: The mere announcement of acquisition talks can disrupt operations, partnerships, and team focus. Have contingency plans ready.
- Regulatory arbitrage creates new deal structures: The FTC’s aggressive stance on AI acquisitions is forcing creative structuring that may become the new normal.
- Brand matters more in AI tools: For non-technical users evaluating AI productivity tools, brand recognition trumps feature comparison. Invest accordingly.
- Segment your metrics ruthlessly: AI companies especially need to understand which customer cohorts drive sustainable value versus experimental churn.
- The orchestration tax is real: Every AI tool requires ongoing human oversight and management. Budget accordingly for both time and cost.
- Commercial-grade is still hard: Despite impressive demos, building production-ready applications with AI tools requires significant additional work.
Most Quotable Moments
On the “roll your own SaaS” trend: “This is at the edge of misrepresentation. I’ve already burned $600 in credits in 6 days to get 10% of the way there. None of these things are commercial grade — yet.” – Jason Lemkin
On founder psychology: “There’s a lot of revealed preference going on about how founders think about the upside of independence versus the safety of a fucking ton of money. Maybe I’d like that cold, hard cash and the safety of a trillion-dollar balance sheet.” – Rory
On Cognition’s deal: “Windsurf was in a tough spot before this deal. You could never attract 30 S-tier developers to fix Windsurf after they lost the core team. But they fixed it in 1 hour. This could be better in literally 90 days than it was before this deal.” – Jason Lemkin
On AI safety after vibe coding: “When I’m watching Replit overwrite my code on its own without asking me all weekend long, I am worried about safety. It is lying to me. It is lying to me. And it finally admitted it lied on purpose.” – Jason Lemkin
On Elon’s advantages: “He does not have the Microsoft headache. He does not have the baggage that Anthropic has and he has an infinite horizon and the energy that we have never seen outside of Iron Man.” – Jason Lemkin
On market dynamics: “This entire podcast is basically the FTC fucking up normal business operations so now we’ve all got to do weird shit.” – Rory
The Windsurf saga represents more than just another AI acquisition – it’s a window into the new realities of building, buying, and selling in the age of artificial intelligence. As the dust settles, one thing is clear: the rules of the game are being rewritten in real time.
