JULY UPDATE: Acqui-Hiring is Heating Up
Following the initial release of our post, “Securing AI Talent in AI Startups by Using IP Moats,” on July 1, 2025, developments in the AI marketplace have demonstrated that the pattern of acqui-hiring is accelerating. Now more than ever, AI startups must look to IP for protection in the dynamic AI marketplace for talent. Recent events provide striking validation of this post’s central thesis about acqui-hire vulnerability and the critical importance of IP moats.
On July 11, 2025, Google agreed to pay approximately $2.4 billion to license the technology of AI coding startup Windsurf and hire CEO Varun Mohan along with a few key employees. [1] This deal materialized after OpenAI’s planned $3 billion acquisition of Windsurf fell through. Mirroring Google’s Character.AI arrangement last fall, Google avoided taking a direct stake in Windsurf, allowing the tech giant to extract the top AI talent without formally acquiring the company, its intellectual property rights, or its technology. By following this strategy, Google was able to gut another AI startup of key talent and IP without paying full value. In the case of Windsurf, hundreds of recent hires were left empty-handed and feeling abandoned by their leaders. After unexpectedly learning of the acqui-hire, the recent hires walked out of a company meeting, understanding that their futures were as uncertain as what was left of the company.[2] Luckily for the remaining Windsurf employees, a rival startup company, Cognition, has set out to bring them on—but at fire-sale prices, certainly not the exit they were hoping for when they joined Windsurf.
On July 16, 2025, Meta similarly poured $14.3 billion into Scale AI, a San Francisco based data labeling company. Meta gained a 49% stake in the company, a release to the company’s IP rights and claims, and a few top employees including founder Alexandr Wang to lead Meta’s new superintelligence team.[3] After Scale AI announced the deal with Meta, both Open AI and Google terminated their contracts with Scale AI, leaving the company’s future uncertain and forcing it to layoff around 200 full-time employees, representing 14% of its staff. [4] This sequence illustrates the harsh reality of the destructive value proposition for both remaining investors and employees of companies that become acqui-hire targets, stripped of their key talent. Both cases demonstrate how promising AI startups increasingly find themselves reduced to talent extraction pipelines rather than being valued as complete, cutting-edge technological enterprises. The harmful effects these scenarios create for remaining employees and investors alike may ultimately discourage members of both groups from continuing to participate in the AI marketplace.
In this blog post, we provide a strategy that may help to preserve the full value of the AI startups and protect the startups from becoming acqui-hire targets in the first place. Critically, both Scale AI and Windsurf lacked blocking IP rights that may have forced would-be acquirers to engage with the full value of the company rather than focusing primarily on team acquisition and negotiating non-exclusive licenses to the IP. If these companies owned critical blocking rights, none of these key employees would have been able to recreate the same businesses at Google or Meta, despite having skill sets relevant for implementing them. Without having key IP blocking rights, neither Scale AI nor Windsurf had the leverage to negotiate a traditional acquisition. Instead, Alphabet (Google’s parent company) and Meta were able to successfully execute end-runs around a traditional acquisition and avoid regulatory scrutiny by buying less than the entire companies (and only access to weak IP rights, if any), to focus on top talent grabs.
Additionally, IP savvy investors should not only insist that an AI startup obtain blocking IP rights but should also require the startup to include a vital protective provision in its Certificate of Incorporation. This provision should prevent the company from effectively liquidating itself through a transaction that sells less than a controlling stake while transferring away licenses, releases, or ownership of the company’s key blocking IP rights. For example, such a provision may include terms requiring the consent or approval of a class of shareholders prior to the transaction being effective for the company to sell, license, transfer, assign, pledge, or otherwise dispose of any blocking intellectual property rights that are core to the business’ value, whether in a single transaction or in a series of related transactions.
GenerativeIQ helps AI startups identify and build the blocking IP rights that allow them to be successful and avoid destructive patterns plaguing the marketplace. With a unique multi-stage approach developed after years of experience obtaining IP rights and a critical understanding of marketplace developments, GenerativeIQ invests, advises, and protects.
DISCLAIMER: The opinions expressed in this blog post are not provided as legal advice.
By Bob Steinberg and Jacob Levine
Bob Steinberg is the Founder of Generative IQ® LLC, a venture fund that provides capital to early-stage IP rich AI start-ups. He has been protecting and litigating IP rights, working with technology entities and entrepreneurs navigating the IP landscape and monetizing blocking rights for over 30 years.
Jacob Levine is a J.D. Candidate 2027 at Harvard Law School where he serves as a Project Director for the Harvard Law Entrepreneurship Project. He graduated from American University in 2022 in the Global Scholars Program with dual degrees in Computer Science and International Studies. Jacob enjoys playing basketball and jazz music, woodworking, weight training, and reading. He loves to go on backpacking excursions with family and friends and attend concerts.
[1]https://www.cnbc.com/2025/07/11/google-windsurf-ceo-varun-mohan-latest-ai-talent-deal-.html.
[2]https://www.wsj.com/tech/ai/meta-ai-recruiting-mark-zuckerberg-sam-altman-140d5861?st=BxobTU&reflink=article_gmail_share.
[3]https://www.forbes.com/sites/janakirammsv/2025/06/23/meta-invests-14-billion-in-scale-ai-to-strengthen-model-training/.
[4]https://www.bloomberg.com/news/articles/2025-07-16/scale-ai-to-cut-14-of-staff-following-meta-investment.
