The tech giant pays millions of dollars for talent

The arms race for artificial intelligence is intensifying, and as the tech giants rush to get to the top, they are hanging millions of dollars in front of a small talent pool of specialists who have come to be known as the AI ​​talent war.

Large tech companies such as Meta, Microsoft and Google are competing for top AI researchers to bolster their artificial intelligence sector and dominate the multi-billion-dollar market.

Meta CEO Mark Zuckerberg recently hired expensive employment to bolster the company’s new AI Superintelligence Labs. This included AI co-founder Alexander Wang as part of a $14 billion investment in startups.

Meanwhile, Openai CEO Sam Altman recently said that CEO Meta had tried to seduce top-open talent with a $100 million signature bonus and an even higher reward package.

Google is also a player of The Talent War and seduces Varun Mohan, co-founder and CEO of Artificial Intelligence, who codes Startup Windsurf, and has joined Google Deepmind in a $2.4 billion deal. Meanwhile, Microsoft AI is quietly hiring Google Deepmind employees.

“There was fierce competition for talent in the field of software engineering 15 years ago, but as artificial intelligence became more and more capable, researchers and engineers specialized in this field were relatively stable,” Alexandru Voica, Head of Corporate Affairs and Policy for AI Video Platforms, told CNBC Make It.

“There’s this supply and demand situation because demand is rising, but supply is relatively constant, and as a result, there’s (wage) inflation,” added Voica, a former meta employee and now a consultant at Mohamed bin Zayed University.

Voica said the multi-million dollar compensation package is a phenomenon the industry has never seen before.

This is what lies behind the AI ​​Talent War:

The specialist’s inflated salary is working with the billion-dollar price tag for building AI models to work with the technology behind your favorite AI products like ChatGpt.

There are many different types of AI companies. There are also things like Synthesia, Cohere, Replika, and Lovable. Others such as Openai, Anthropic, Google, Meta, and others build and train large-scale language models.

“There are only a handful of companies that can afford to build these types of models,” says Voica. “It’s very capital-intensive. There are not many companies that have billions of dollars to spend on building models, and as a result, these companies are a way to get closer to this.

Anthropic CEO Dario Amodei told Time Magazine in 2024 that he expects training costs for frontier AI models to be $1 billion that year.

Stanford University’s AI Institute recently produced a report showing the estimated costs of building selected AI models between 2019 and 2024. For example, Openai’s GPT-4 would cost $79 million to build in 2023, while Google’s Gemini 1.0 Ultra was $192 million. Meta’s Llama 3.1-405B cost $170 million to build in 2024.

“Companies that build products pay to use these existing models to build on them. So capital expenditures are low and there’s not much pressure to burn money,” Voica said. “The spaces where things are very hot in terms of pay are companies that are building models.”

Machine learning engineers are AI experts who can build and train these large-scale language models, and their demand is high on both sides of the Atlantic, says Benritvinov, associate director at Robert Walters, Technology Recruitment Company.

“There’s no doubt that demand will increase significantly, especially in terms of both AI-centric analytics and machine learning. That’s why we’re deploying large language models and people who are deploying more sophisticated deployments of either GPT-backed, more advanced AI-driven technologies or solutions,” explained Litvinoff.

This includes the “slim talent pool” of experienced professionals who have worked in the industry for years, he said. Also said AI research scientist who has completed his PhD from the top five or six universities in the world and has been given up by the high-tech giants upon graduation.

Zuckerberg is reportedly offering $250 million to 24-year-old AI genius Matt Deitke, who dropped out of the University of Washington’s PhD program.

Meta directed CNBC to Zuckerberg’s comments on information that Facebook founders said there was an “absolute premium” for top talent.

“Many details reported are not accurate in themselves. But it’s a very hot market. So, you know, there are a few researchers who are in demand in all of the different labs, and there are the best researchers,” Zuckerberg told the Technology Publications.

“The amount of money spent on recruiting people is actually still very little when talking about the overall investment and super intelligence.”

Litvinoff estimated that machine learning and principal engineers are currently earning six-figure salaries in the London market, ranging from £140,000 to £300,000 in more senior roles, on average.

In the US, the average salary for a machine learning engineer was $175,000, actually reached nearly $300,000 in the high-end.

As the tech giant continues to plague the best minds with AI with the temptation of mammoth salary, startups risk being left behind.

“Some of these startups trying to compete in this building model space are so expensive to build models that it’s difficult to see how to move forward, but I don’t know if the companies buying those models can afford to pay a price that covers the cost of building models,” Voica said.

Mark Miller, founder and CEO of InsureVision.ai, recently told Startups Magazine that this talent war also creates a “big opportunity gap” in the traditional industry.

“The whole industry, including insurance, healthcare and logistics, cannot compete on salary. Innovation is necessary, but you don’t have access to talent,” Miller said. “The current situation is absolutely unsustainable. We can’t accumulate all the talent in one industry, but others are withering.”

Voica said AI experts must make their choice. Some people take on the higher pay and bureaucracy of big tech, while others tend to lean towards startups where staff have lower pay but have more ownership and impact.

“In a large company, you’re essentially a cog of a machine, but at a startup you can have a lot of influence. You can have a lot of influence through your work and you can feel the impact,” says Voica.

However, until the AI ​​model build prices drop, AI talented salaries may remain.

“Unless companies have to spend billions of dollars building models, they’ll spend tens of millions, or even hundreds of millions, of hired engineers to build those models,” Voica added.

“All of a sudden tomorrow will reduce the cost of building these models by 10 times, so will the salary I expect.”