Steven Sinofsky & Balaji Srinivasan on the Future of AI, Tech, & the Global World Order
Steven Sinofsky and Balaji Srinivasan analyze the collision of regulation, capital, and innovation, exploring how an 'anti-tech assault' from regulators is reshaping M&A. They introduce the 'acquifire'—a new deal structure born from pressure—and debate the fundamental conflict between the permissionless 'network' and the regulatory 'state', projecting how this battle will define the future of AI.
In a dynamic discussion, Steven Sinofsky and Balaji Srinivasan deconstruct the tightening grip of regulation on the tech industry and its profound impact on capital markets, M&A strategy, and the future of AI innovation.
The New Reality: An “Anti-Tech Assault”
The conversation opens by framing the current environment as an “all-out anti-tech assault.” For years, capital markets have grown increasingly hostile. The Sarbanes-Oxley Act (Sarbox) diminished the viability of IPOs, forcing companies to stay private longer. Now, a wave of aggressive antitrust actions from the FTC and DOJ is closing the M&A exit ramp.
Balaji Srinivasan points to the blocked Figma-Adobe, JetBlue-Spirit, and iRobot-Amazon deals as evidence of a chilling effect, where companies are left to “silently die” without viable exits. He argues this isn’t isolated to M&A but is part of a broader offensive that includes limiting AI compute (flops), SEC “lawfare” against crypto, and a zero-sum mentality from Washington D.C. where any success in the tech world must be co-opted or controlled by the state.
The Fundamental Conflict: Network vs. State
A core framework for the discussion is the inherent conflict between the “network” (the internet, tech) and the “state” (government, regulators). The tech industry, particularly software, grew into a global economic force largely without government permission or oversight. As Srinivasan notes, “You need a license to cut hair… you didn’t need a license to own a computer, the most powerful device ever created.”
This permissionless innovation created a parallel sphere of influence. As tech platforms scaled, they began to perform regulatory functions, taking “market share” from the state:
- Uber/Lyft became the de facto regulators of taxis with real-time tracking and rating systems.
- YouTube/Facebook became the de facto regulators of speech and content distribution.
This erosion of state power prompted the “empire to strike back.” Regulators are now applying industrial-era antitrust laws, designed for a world of tangible assets and resource constraints (like railroads and coal), to the intangible, non-zero-sum world of software. Steven Sinofsky highlights the absurdity of trying to calculate a market share metric like the Herfindahl-Hirschman Index (HHI) for software, where markets are fluid and hard to define. What is the market for word processors? Is it just Microsoft Word, or any text field on any device?
Srinivasan adds that this is compounded by a lack of numerical literacy among regulators, who often fail to grasp the scale of the industry they oversee, treating terms like “million” and “billion” as interchangeable synonyms for “a big number.” This leads to policies based on “a mask for animus” rather than sound judgment.
The Power Law and Revisionist History of M&A
Sinofsky argues that corporate M&A, much like venture capital, follows a power law. The vast majority of acquisitions are “provably a net destroyer of value.” The few that succeed, however, are transformative.
The discussion revisits the acquisitions of YouTube by Google and Instagram by Facebook. At the time, these were widely seen as reckless gambles:
- YouTube was dismissed as a copyright-infringing morass of “kitten videos.”
- Instagram was a pre-revenue app with 13 employees, and Mark Zuckerberg’s $1 billion purchase was viewed as an act of hubris right before Facebook’s IPO.
In hindsight, regulators and critics “retcon” these high-risk, visionary bets into obvious, anti-competitive moves by “evil monopolists.” This ignores the immense risk taken by the acquirers. As Srinivasan puts it, “Everybody wants a piece of the reward. No one wants a piece of the risk.” By blocking M&A, regulators stifle the very ecosystem that creates competition. A successful acquisition signals a failure of the big company to build the product itself and injects capital and excitement back into the startup ecosystem, fertilizing the ground for the next wave of competitors.
Innovating Under Pressure: The Rise of the “Acquifire”
The intense regulatory climate has forced companies to innovate their deal structures to avoid antitrust scrutiny. This has led to the emergence of what Srinivasan calls the “acquifire,” a structure seen in recent deals involving Scale, Character, and Windsurf.
- Traditional Acquisition: The acquirer buys the entire company, and proceeds are distributed via the cap table.
- Acquihire: The team is hired, but the company and its IP are often shut down. The founders and team get status (a prestigious new job) but investors may not get their money back.
- Acquifire: A large company hires a key group of employees (typically elite AI engineers) for a significant sum. The original company is left as a shell, but with a large amount of cash in its bank account, intended to be paid out to investors. The team gets the money, but not the status of a full acquisition.
The Windsurf-Google-Cognition saga serves as a case study for the unintended consequences of this structure. Google “acquifired” Windsurf’s core engineering team, leaving behind a large sum of cash and the remaining 200 employees (mostly in sales). Because the deal was not a formal acquisition, communication was constrained. The remaining team, feeling they had been denied the status of a Google acquisition, rationally negotiated a second deal to be acquired by Cognition, thereby regaining status. This highlights how regulatory interference creates messy, non-transparent outcomes where the last person with a face is blamed, while the faceless regulator avoids accountability.
AI: The Next Platform Shift and Looming Threats
The speakers agree that AI represents the next major platform shift. The current frenzy is focused on building the essential tooling for this new era, making teams with this expertise incredibly valuable and justifying the high price tags of acquifires.
However, they see serious risks to continued US leadership in AI:
- Copyright Assault: Legacy media and creators, threatened by AI, are launching “desperate lawsuits” to kill the technology.
- Energy Constraints: The massive energy requirements for data centers will soon hit physical limits.
- Geopolitical Competition: China is aggressively releasing powerful, open-weight models (e.g., Kimmy, DeepSeek) to commoditize the US’s current advantage, much like Google Docs commoditized Microsoft Office.
- Regulatory Overreach: The combination of copyright attacks, potential bans on using foreign models, and general anti-tech sentiment could lead to AI innovation being stifled in the US, forcing it to decentralize globally, similar to what happened with crypto.
A Proactive Path Forward: Antitrust for the Government
The conversation concludes not with a plea for better regulation, but with a call for a proactive strategy to circumvent it. Srinivasan argues the tech industry must stop playing defense. The path forward is to build jurisdictional competition.
“Assume the US federal government was a monopoly, and how do we build competition to that? How do we build jurisdictional competition? How do we build choice?”
The proposed strategy involves:
- Drafting ideal, pro-innovation model legislation for key sectors like AI.
- Identifying friendly jurisdictions—states or small countries eager for investment.
- Building a “sales team” to present these legislative packages, backed by commitments of capital and company-building from a coalition of tech leaders if the laws are passed.
The goal is to create zones of innovation, or “startup societies,” where building can happen at the “speed of physics, not permits.” This shifts the dynamic from reacting to regulators to creating a market of jurisdictions where the best regulatory frameworks win.