Faced with the poaching offensives of giants like Meta, which often cost hundreds of millions of dollars, significant differences have emerged in the strategies for responding to the AI talent war in Silicon Valley.
The Wall Street Journal reportedRefers toOpenAI has taken a radical step by announcing the elimination of the "vesting clamp" for new employees, meaning they will be granted stock options upon joining. However, Microsoft's AI CEO, Mustafa Suleyman, stated in an interview that he refuses to join this arms race of "sky-high salaries" and bluntly stated that Meta's money-spending strategy is flawed.
OpenAI: Breaking industry norms, employees can hold shares on their first day.
In the traditional Silicon Valley startup scene, standard equity incentive plans typically have a one-year "restriction period." Employees must work for a full year before receiving their first share of stock, which is seen as a "golden handcuff" that binds talent. However, to retain top talent in a highly competitive market, OpenAI is breaking this rule.
Starting in April of this year, OpenAI shortened the restriction period from one year to six months. However, application lead Fidji Simo later announced the removal of this restriction, meaning new employees will have equity interests upon joining the company and will not have to worry about being laid off and losing everything due to project failure.
According to sources familiar with the matter, besides encouraging employees to take risks, a more practical reason for this move is to be more competitive in the talent war against Google and Meta. Therefore, OpenAI's stock-based compensation expenses this year are expected to reach $60 billion, almost half of its projected revenue.
xAI had no choice but to follow suit?
Elon Musk's xAI reportedly shortened its lockout period in late summer, but the reason is somewhat embarrassing. Reports indicate that xAI faces a high-pressure culture (Elon Musk claims that being awake means working) and image controversies (such as Grok's anti-Semitic remarks and the scantily clad virtual avatar Ani), causing hiring managers to frequently encounter obstacles when trying to recruit, forcing them to relax their criteria to increase the acceptance rate.
Microsoft: We don't play that game.
Compared to the talent wars between OpenAI and xAI, Microsoft, the biggest backer of OpenAI, seems quite unfazed, with its AI head, Mustafa Suleyman, appearing completely at ease.
During the interviewRegarding rumors that Meta offered a signing bonus of up to $1 million or even $2.5 million to hire top researchers, Mustafa Suleyman stated bluntly: "I don't think anyone can (or should) offer comparable terms."
Mustafa Suleyman further criticized Meta CEO Mark Zuckerberg's "mass hiring" strategy, arguing that it was not the right approach. Suleyman emphasized that he continues his hiring philosophy from his time at DeepMind, remaining "very selective" in his selection of talent and employing a "progressive" hiring strategy.
Rather than spending money to build a massive army, Mustafa Suleyman clearly prefers to assemble a lean and efficient elite team with a high degree of cultural compatibility, and unqualified candidates will be decisively eliminated.
Analysis: Is it a "thirst for talent" or a "bubble warning sign"?
In my opinion, these two news items reveal two completely different survival logics within the AI industry:
OpenAI's "Speed War": As a startup that still needs to justify its high valuation (despite already being a behemoth), OpenAI doesn't have the same deep-rooted confidence as Microsoft. Faced with Meta's financial resources and Google's computing power, they must use the most aggressive compensation structure to minimize resistance to talent mobility, even if it makes investors panic about the speed at which they burn through cash.
Microsoft's "positional warfare": Mustafa Suleyman's remarks reflect Microsoft's stability as a long-established giant (and perhaps also an excuse for cost control). It doesn't lack computing power or application scenarios; what it lacks are "elite troops" capable of implementing its technology.
The removal of stock option restrictions signifies that the AI talent market has completely shifted to a "seller's market." However, this also raises concerns: with employees receiving stock options upon joining and the cost of switching jobs being extremely low, will these top minds become "mercenaries," frequently moving between various giants? This may not be a good thing for AGI research, which requires long-term investment.
