[ad_1]
The AI industry is in the middle of a cutthroat talent war, with tech giants and startups competing fiercely for top researchers, engineers, and executives. This has led to aggressive poaching, multimillion-dollar compensation packages, and growing regulatory scrutiny.
According to a recent report by SignalFire, Anthropic has become a major destination for talent departing OpenAI and DeepMind. Engineers are eight times more likely to exit OpenAI for Anthropic and nearly eleven times more likely to leave DeepMind for the same.
The report also analysed retention rates across top AI labs, with Anthropic emerging as the leader. Nearly 80% of people who joined Anthropic two years ago still worked there, which is relatively rare in AI, where job-hopping is common. DeepMind isn’t far behind at 78%. OpenAI’s retention is lower at 67%, followed by Meta’s 64%.
Meanwhile, Anthropic is reportedly preparing to launch its first employee share buyback programme. As per The Information, the company intends to repurchase shares from both current and former employees at a $61.5 billion valuation. This matches its recent $3.5 billion Series E funding round. The move would give employees a chance to cash out some equity while helping Anthropic retain top talent.
Such employee liquidity initiatives are becoming increasingly common in the AI sector. By allowing early share sales, startups try to keep their teams motivated without relying solely on high compensation packages.
AIM spoke with Deedy Das from Menlo Ventures about what drives talent to switch companies. He said, “The biggest factor is total comp (with equity). No one wants to talk about that, but that’s the main reason people switch. 7-figure (a salary of one million dollars or more per year) offers are very common for specialised skills.”
It’s not just about the numbers. Some of the biggest names in AI are making the jump. Sholto Douglas joined Anthropic in March after previously working at Google DeepMind. He had recently appeared on a podcast with Dwarkesh Patel and predicted that by 2027–28, AI will likely automate nearly every white-collar job.
Niki Parmar, a former Google AI researcher and co-author of the groundbreaking ‘Attention Is All You Need’ paper, recently joined Anthropic.
Earlier this year, Neil Houlsby, a former research scientist and manager at Google DeepMind, was brought on to lead Anthropic’s new Zurich office. Pavel Izmailov, who worked on reasoning and safety at OpenAI, joined Anthropic after reportedly being dismissed from OpenAI.
Jan Leike, who previously co-led OpenAI’s Superalignment team, also moved to Anthropic in May last year, citing concerns that OpenAI was prioritising “shiny products” over building a strong safety culture.
Das stated that there is a shortage of talent because the knowledge required to train models at scale is not really taught in schools. It is primarily possessed by the few individuals who have experience in it. “It costs money to train models, and most people in university don’t even have access to the hardware to do it. Most of the academic models are small.”
Big Tech Goes All-In to Retain Talent
Big tech companies are doing everything they can to retain employees and prevent them from leaving. A recent report reveals that Google DeepMind is enforcing non-compete agreements that bar some UK-based staff from joining rival firms for up to a year after they leave.
The restrictions vary, depending on employees’ roles and seniority levels. While six-month non-competes are common for individual contributors, some senior researchers have been subject to agreements lasting as long as 12 months. In certain cases, employees have been placed on extended garden leave, where they continue receiving pay but are no longer actively working at DeepMind.
Moreover, in a bid to retain top talent, DeepMind is shelling out annual compensation packages of up to $20 million, introducing unscheduled equity awards, and shortening stock vesting timelines from four years to three.
Google has recently been able to make impactful hires.
Tim Brooks, who co-led the team behind OpenAI’s video model Sora, left the company in October 2024. He’s now at Google DeepMind, working on video generation and world simulation.
Logan Kilpatrick, who was in charge of developer relations at OpenAI, has also moved to Google. He now leads product for Google AI Studio and the Gemini Developer API at DeepMind. Since his arrival, Google has seen a major boost, especially with the success of its Gemini 2.5 models.
Meanwhile, Microsoft is allowing managers to request retention bonuses for employees they believe are too valuable to lose. According to a recent report, the process involves filling out forms that include questions such as, “What harm is done if the employee leaves Microsoft?”
The document instructs managers, “In the context of AI transformation as a key priority, please indicate if this individual is critical AI talent and share the risk to the AI initiative(s) if the talent is not retained.”
OpenAI Mafia
The global AI talent pool is limited, with only a few thousand researchers capable of developing advanced models such as ChatGPT or Gemini. This competition extends beyond big tech, as research labs and startups are also aggressively recruiting.
One of the most striking departures was Mira Murati, who left her position as OpenAI’s chief technology officer in September 2024 to start her own AI venture. She didn’t go alone; she brought around 20 OpenAI employees with her and later expanded the team to 60 by hiring from OpenAI and other AI labs.
At the same time, a few leading OpenAI researchers interested in joining former chief scientist Ilya Sutskever’s new firm, SSI (Safe Superintelligence), were reportedly offered retention bonuses of $2 million, along with equity increases exceeding $20 million, if they remained at OpenAI, according to Reuters. Interestingly, some of them only needed to commit to staying for a year to receive the full bonus. Neither SSI nor OpenAI has commented on the matter.
Meanwhile, Reuters reported that other OpenAI researchers considering offers from Eleven Labs were given retention bonuses of at least $1 million to stay put. High-level researchers at OpenAI are said to regularly receive compensation packages exceeding $10 million annually.
However, that doesn’t mean OpenAI isn’t hiring from rivals. In December last year, the company brought in three AI researchers—Lucas Beyer, Alexander Kolesnikov and Xiaohua Zhai—from Google DeepMind, all of whom specialise in computer vision and machine learning.
Most recently, Tim Zaman from Google DeepMind joined OpenAI to work on frontier clusters. Since early 2024, at least 44 former Google employees, half of them engineers, have joined OpenAI, with a total of 85 making the switch over the past 18 months.
The company also brought in Karan Singhal, who now leads health research at OpenAI. He led the team that recently launched HealthBench. Before that, he worked on medical foundational models at Google.
The Tactics of Poaching
Big-tech giants like Microsoft, Google, and Amazon are using different tactics to attract top AI talent. One standard method is pseudo-acquisition, where they hire key employees and license technology from AI startups without actually buying the companies.
For instance, Google brought in Noam Shazeer along with much of the Character.AI team. Microsoft hired several people from Inflection, including its founder, Mustafa Suleyman.
Amazon did something similar in a deal with Adept, a generative AI startup. It hired most of the team and licensed the company’s technology without a full acquisition.
The practice has drawn antitrust scrutiny, with Senators Ron Wyden and Elizabeth Warren urging the FTC to investigate whether Big Tech is bypassing regulatory oversight by acquiring talent and technology from startups through informal deals.
The US Justice Department is currently investigating whether Google’s agreement with Character.AI violates antitrust laws and whether the deal gives the tech giant an unfair advantage through access to the startup’s technology.
The AI talent war shows no signs of cooling. As companies race to build the next generation of AI models, the demand for specialised talent will only grow. Startups will continue to dangle equity and autonomy, while Big Tech counters with unmatched resources.
[ad_2]
Source link
