Silicon Valley captivated by epic battle for AI talent
Cutthroat recruitment efforts to land the smartest minds in Silicon Valley have hit a feverish new peak in recent days.
Hundreds of employees at one of Silicon Valley’s hottest AI start-ups gathered in their offices last Friday expecting a celebratory announcement.
For months, OpenAI had been talking to Windsurf about buying it for $US3 billion, and now it seemed like the rank-and-file were finally getting confirmation that the deal was about to become official. Windsurf’s marketing team even began filming the all-hands meeting for promotional material.
Instead, they learned that Windsurf’s chief executive, Varun Mohan, had left the company to join Google, taking with him a small group of artificial-intelligence researchers and engineers. After hearing the news, some of the staff began to cry.
By Monday morning, another twist in the saga would bring those Windsurf employees back to the same room for a second announcement: The rest of their company would be acquired by a rival AI start-up.
In other words, it was just a typical weekend in Silicon Valley. The war among some of the richest companies on the planet for talent is playing out in an unprecedented frenzy of talent raids, secret deals and betrayals, leading brainy AI researchers whose minds have never been so highly valued to become as rich as NBA players and Hollywood stars.
The most powerful CEOs in tech are dangling pay packages worth more than $US300 million to their most prized recruits – and even that kind of money isn’t always enough to win them over.
Every feverish new development in the all-out brawl for talent has captivated Silicon Valley. The company at the centre of the action is Meta, which is in the midst of one of the most astonishing recruiting blitzes of all time. Led by Mark Zuckerberg, who is personally assembling his AI dream team, the company has poached the leaders of promising start-ups, stunning their investors and employees.
It has also given so-called exploding offers that expire within days to potential hires so it’s harder for Meta’s rivals to negotiate effective counters.
The playbook has left the industry to wonder whether the social, mission-first contract that once united founders and employees is unravelling.
Meanwhile, some executives are bemoaning the erosion of what was once a bedrock principle of Silicon Valley: Be a missionary, not a mercenary.
Missionaries vs. mercenaries
OpenAI CEO Sam Altman described the battle in those terms in a Slack message he sent to researchers in late June when Zuckerberg was attempting to raid his company.
“I am proud of how mission-oriented our industry is as a whole,” Altman wrote in a message seen by The Wall Street Journal. “Of course there will always be some mercenaries. Missionaries will beat mercenaries.”
That idea is widely credited to John Doerr, one of the titanic figures in venture capital. He is now the chairman of the famed VC firm Kleiner Perkins, which was on the board of Windsurf. For decades, he told generations of entrepreneurs who came to Silicon Valley with dreams of starting a company to embrace their inner missionaries. In mercenary cultures, “the central goal is a lust for making money,” Doerr once said.
But at missionary companies, “there’s a lust not only for making money, but for making meaning.”
Meta rejects the characterisation that its newly hired employees were simply chasing money. Zuckerberg has said that his company’s allure isn’t the compensation packages but its access to the massive amounts of computing power that researchers need to make breakthroughs.
“Meta Superintelligence Labs will have industry-leading levels of compute and by far the greatest compute per researcher,” he wrote in a Threads post this past week.
But the market for talent has been turned upside down by his willingness to throw colossal sums of money at wonky AI researchers – a decision partly inspired by a conversation with an influential figure at the company he’s targeted the most in his recruitment spree.
In the spring, Zuckerberg approached OpenAI’s chief research officer, Mark Chen, for a casual catch-up and ended up asking him for advice on how to improve his company’s generative-AI organisation. Given how much money Meta was already spending on hardware and computing power to train AI – more than 100 times what it was spending on humans – Chen suggested that Zuckerberg might want to invest more in talent, according to people familiar with the conversation.
Zuckerberg asked Chen if he would consider joining Meta – and what it would take to bring him aboard.
A couple hundred million dollars? A billion?
Chen demurred, saying he was happy at OpenAI. But the conversation helped plant the seeds of an idea.
Zuckerberg got to work creating a list of the world’s best AI researchers. He fired off emails, texts and WhatsApp messages to potential hires, gauging their interest and inviting some to meet him at his homes in Lake Tahoe and Palo Alto, Calif.
Zuckerberg settled on Alexandr Wang to lead the new lab. To get him, Meta paid $14 billion for a stake in Scale AI, the data-labelling start-up founded by the 28-year-old entrepreneur.
Wang was an idol to many young start-up founders, who saw him as a model for how to take risks and win big. He dropped out of the Massachusetts Institute of Technology as a freshman and moved to California to build Scale, becoming one of the world’s youngest self-made billionaires once the start-up’s valuation skyrocketed. Scale began to grow even faster as the AI boom took off, when it hired contractors from all over the world to tag data for tech companies training AI models.
On a Friday morning in mid-June, Wang told employees gathered in the plant-filled atrium of Scale’s cavernous San Francisco headquarters that he was leaving the company. He walked down the building’s staircase to thunderous applause from his staff.
Wang teared up when he recounted how he started the company at age 19. Some employees also cried and stayed afterwards to take pictures with their former CEO. “It was like the end of a Disney movie,” one of them said.
In the days after Wang’s departure, OpenAI and Google ended their contracts with Scale, a blow to its revenue. This week, Scale cut 14 per cent of its remaining staff. A spokesman for the company said it remains strong and well-funded.
‘Someone has broken into our home’
For his new AI division focused on superintelligence, or AI that is smarter than humans, Zuckerberg has poached from Anthropic, Google DeepMind and Apple, and pulled at least a dozen employees from OpenAI.
But many have turned him down. He’s offered more than 10 of OpenAI’s researchers eye-watering pay packages of $300 million over four years, including $US100 million the first year, according to people familiar with the matter. And after months of recruiting, Meta still doesn’t have a chief scientist for its lab.
“I’ve lost track of how many people from here they’ve tried to get,” Altman told his staff in a Slack message.
Still, the pressure has rattled OpenAI. Chen, whose conversation with Zuckerberg had helped spur Meta’s recruiting push, reassured OpenAI employees in late June that the company’s leadership was responding to the aggressive recruitment by adjusting compensation packages and “scoping out creative ways to recognise and reward our top talent,” he wrote in a Slack message seen by the Journal.
He also equated Meta’s talent raid to a burglary.
“I feel a visceral feeling right now – as if someone has broken into our home and stolen something,” Chen wrote. “please trust that we haven’t been sitting idly by.”
Chen shared another OpenAI research leader’s message to employees: “if they pressure you, or make ridiculous exploding offers just tell them to back off, its not nice to pressure people in potentially the most important decision.” OpenAI employees reacted to Chen’s note with heart emojis – and emojis of Chen’s face.
Zuckerberg and Altman hadn’t spoken since the battle for talent began until the CEOs met last week at the Allen & Co. conference in Sun Valley, Idaho, according to a person familiar with the matter.
Buzziest hires
Among the Meta Superintelligence Lab’s buzziest hires are Nat Friedman and Daniel Gross, AI investors who became popular among start-up founders after they secured a cache of Nvidia chips for their portfolio companies during the chip shortage two years ago.
Friedman, the former CEO of the coding platform GitHub, was hired to be one of the AI lab’s leaders alongside Wang.
Gross was the CEO of AI company Safe Superintelligence, or SSI for short, which he founded last year with former OpenAI chief scientist Ilya Sutskever, a luminary in the field who left OpenAI last year after voting with other board members to fire Altman in November 2023.
Zuckerberg received an introduction to Sutskever from Gross. The two had known each other ever since Zuckerberg considered buying Gross’s search-engine start-up more than a decade ago. In late January, the three of them got together for lunch at Zuckerberg’s home in Palo Alto. Shortly afterwards, the Meta CEO said he was interested in buying SSI.
Sutskever immediately rebuffed the offer and told him that SSI was not for sale, according to people familiar with the matter. A few weeks later, Meta invested in an SSI funding round that valued the start-up at $US32 billion.
Gross told close associates he disagreed with Sutskever’s decision not to sell SSI. In May, he received his own offer to join Meta alongside Friedman. Sutskever was blindsided when he learned that his CEO and co-founder was defecting to a rival.
When the news of Gross’s likely move began to leak out in the middle of June, waves of surprise rippled through the text chains of founders, investors and prominent tech CEOs. They were stunned that he would abandon his start-up so quickly for a competitor.
“I have incredible empathy for Daniel. This was an unimaginably hard decision,” said Shaun Maguire, a Sequoia partner who helped lead the firm’s investment in SSI.
Earlier this month, in a post on X announcing Gross’s departure, Sutskever made it clear that his company is not for sale, no matter how many billions of dollars might be on the table.
To hire Gross and Friedman, Meta had to do more than cut large checks. It also had to help unwind NFDG, their venture firm, and offered to buy out up to 49% of the shares owned by investors in its funds.
Until recently, the firm’s website offered a simple description of its mission: Nat Friedman and Daniel Gross invest in start-ups together.
Then the language on the website was tweaked: Nat Friedman and Daniel Gross invested in start-ups together.
Windsurf’s second wave
Many of the most lucrative offers have been extended to researchers who were already working for leading AI labs, but the latest battlegrounds in the talent wars are the industry’s most promising start-ups, such as Windsurf.
The most recent drama in Silicon Valley began months ago, when OpenAI agreed to buy Windsurf for $3 billion. But the deal process dragged on and eventually broke down when Microsoft, OpenAI’s largest investor, objected to some of the terms. After the exclusivity window for negotiations with OpenAI expired, Windsurf turned to a backup plan.
Mohan, the Windsurf CEO, quickly struck a $US2.4 billion deal with Google. Google got access to the start-up’s technology and hired away some of its key employees.
These so-called acquihire deals have become a common strategy for Silicon Valley giants to evade antitrust scrutiny, but they leave the remaining employees working for companies that are shells of their former selves. In the case of Windsurf, hundreds of recent hires were left empty-handed and feeling abandoned by their leaders. When they walked out of the meeting last Friday afternoon, their futures were as uncertain as the company’s.
“I want to acknowledge that this is extremely shocking news and it will take at least 24 hours to digest it,” Windsurf’s new CEO, Jeff Wang, said during the emotional all-hands. “It might take all weekend.”
But before the weekend began, Wang received an email at 5:30pm from Scott Wu, the CEO of coding start-up Cognition, with the subject line: “chat?”
Wu, a competitive programmer who once won three gold medals at a global coding Olympics for teenagers, had launched a viral AI coding tool called Devin last year.
Wu and Wang chatted and struck a deal on Saturday afternoon for Cognition to buy the rest of Windsurf.
The two spent the rest of the weekend hunkered down with their colleagues in Windsurf’s offices hammering out the deal. Wang told employees on Monday morning that all of them would get paid out as part of the deal, regardless of whether their equity had vested.
This time, he received a standing ovation.
Dow Jones Newswires
More Coverage
Add your comment to this story
Travelling abroad? How to avoid surprise phone charges and save money on roaming
A daily pass from your carrier is convenient, but a prepaid, downloadable eSIM can be a lot cheaper.
Trump sues News Corp over Epstein birthday letter article
The defamation lawsuit was filed in a Florida federal court
To join the conversation, please log in. Don't have an account? Register
Join the conversation, you are commenting as Logout