The High-Tech Giants aren't the only ones fighting the talent of top AI.
From tens of thousands of dollars to tens of millions of dollars worth of poaching packages, the talent war “has a waterfall effect on comps” across early stage startups, says Christina Cordoba, former VC and COO of problem tracking company Linear.
The pay and stock packages are climbing, and startups are promoting the unique impact and ownership of working in smaller locations, recruiters told Business Insider.
According to Shawn Thorne, managing director of executive search firm True Search, the AI director of Series A company can direct the background and time spent at Top Research Labs between $300,000 and $400,000.
This has no prior management experience. Before the AI boom, he said the salary for Vice President of Engineering could be close to $200,000 to $250,000.
“We have two sets of rules,” Thorn said. “There's a price you need to pay for every talent, and a price you need to pay for AI talent.”
It's not just the basic salary that's exploding. Fairness is a “big factor” as startups compete with the “opportunity costs” of top researchers or applied engineers.
The top engineer stock package rose from 2% to 5%, Thorne compared to a portion of the previous early stage non-managed employee percentage.
Other negotiation chips include enabling AI researchers or engineers to win co-founder titles, computing access, and time for research outside of research, Thorn said.
“We've seen top candidates win with remote flexibility, front-road equity, big sign-on bonuses, tougher cycles of compensation refreshes and friends on teams,” said Nathan Fisher, investor and co-founder of recruiting services singles prouts. “It's not just an offer, it's just how fast and personal the process goes.”
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
The fight for AI talent has become unprecedented. In June, Meta spent $14.3 billion on what was considered a acquisition of Scale AI investment and Scale AI CEO Alexandr Wang, and Google snapped Windsurf's top talent for $2.4 billion this month.
Openai CEO Sam Altman previously said Meta had tried to poach talent with a $100 million signature bonus.
Startups can provide ownership and impact
Freenetic Job-Hopping is becoming a new normal for AI, Cordoba said, contributing to a “mercists' mindset” and a higher bar for employee retention.
Startups have advantages. They can offer the opportunity to be part of something transformative and the potential financial stairwell that comes with it. They can also provide a sense of freedom, as opposed to being a limited range of machines gears. (Large tech CEOs like Meta CEO Mark Zuckerberg are well aware of these dynamics and likewise beg top candidates to provide less reporting and more AI calculation power.)
Fisher said he doesn't know of any early-stage startups that can compete with the $10 million reward package, but he said, “You can win by pitching sharper issues, faster cycles, and you can win by providing top AI engineers with the ability to own a product with strong advantages.”
The bigger war on talent has not affected all startups, including those that build existing AI models. Shie Gabbai, the chief operating officer of AI travel planning firm Layla, said that while almost all of the company's funding goes to employment engineers, the salary has not been affected.
Gold Star AI candidate
AI startups face growing gaps between Goldstar researchers (the type selling billions of startups) and rank and file employees who may not have the same pedigree.
According to Mark Bai, managing director of True Search, the latter could potentially earn more salaries along what is typical in Silicon Valley.
As AI is increasingly replaced with early engineering jobs, tech companies will diversify their salaries among a small number of people.
Goldstar candidates are usually doctoral degrees in well-known computer science programs focusing on machine learning, such as Stanford, Berkeley, MIT, Waterloo and Carnegie Mellon, and five years later, they landed on top AI research teams at major companies, Thorn and Bye said.
“Almost every client that Mark and I talk about wants to poach someone from open or fair in the meta,” Thorn said, referring to Meta's basic AI research group.
Fisher added that Xai is often flagged. Many top researchers know each other too. It's a great benefit when building a team.
“More than certain skills, companies want people who live and breathe AI,” Fisher said.
He said this includes people who will ship side projects, stay close to model updates, write code, talk at meetings, have a social media presence and write academic papers.
AI WARS promotes new work strength
The war of talent is a key symptom of “the level of urgency of technology that has been lacking over the years,” Cordoba said.
Every time a new AI model is launched, a new wave of companies will be created and die, Bai said. The strength is revealed by the expectations of a more strict return office and the establishment of “founder mode.”
“They are introducing some of this 996 culture,” Bai said. He mentioned the extensive technical work culture that works in China from 9am to 9pm six days a week.
Bai added that some companies work seven days a week. “There's a literal office with bunk beds.”
The high pressure atmosphere is synonymous with startup culture, but Thorne said the pendulum looked back in a big way after the sub-pandemic.
“This is the most intense I've ever experienced from a client over the past decade,” Bye said.
