Meta CEO Mark Zuckerberg offered a bonus signing $ 100 million to top Openai employees.
David Paul Morris | Bloomberg | Getty images
Artificial Intelligence Arms Race is heating up, and scuffles to top the top as tech veterans, they are hanging millions of dollars in front of a small talent pool of experts known as AI Talent War.
It is looking like big tech firms Meta, MicrosoftAnd Google Compete for top AI researchers in an attempt to enhance your artificial intelligence divisions and dominate the market of multibibilians.
Meta CEO Mark Zuckerberg recently started the company’s new AI Superintendent laboratories and started in competition. This included poaching Scale AI co-founder Alexander Wang As part of an investment of $ 14 billion in startups.
Meanwhile, Openai’s CEO Sam Altman recently said that the CEO of Meta had tried to woo the top openi talent. $ 100 million signed bonus And even high compensation package.
If I am going to spend a billion dollars to make (AI) model, then an engineer has a relatively low investment for an engineer.
Alexandru Voca
Head of Corporate Affairs and Policy
Google talent is also a player in war, breathtaking Varun MohanArtificial Intelligence Coding Startup Windsurf’s co-founder and CEO, Google to join the deal of $ 2.4 billion in Deepmind. Meanwhile, Microsoft AI has quietly hired two dozen Google deepmind employee,
“In software engineering space, 15 years ago there was an intensive competition for talent, but as artificial intelligence had become more and more capable, specific researchers and engineers in the region have been relatively stable,” Alexandru Voca, the chief of corporate affairs and policy at the ALC video platform synthesia, described CNBC.
“You have this supply and demand situation, where the demand has now skyrocketed, but the supply has been relatively stable, and as a result, (wage) inflation is,” Voca, a former meta employee and currently an advisor to Mohammed bin Zaid University Artificial Intelligence.
Voca said that the package of compensation of multi-military dollars is an incident that the industry has never seen before. “
What is AI talent behind war here:
AI models cost billions
Plans for experts come hand in hand with the price tag of the Arab-Dollar of the AI model to build the AI model-the technique behind your favorite AI products.
There are different types of AI companies. Some, such as synthesia, cohere, replication, and cute, manufacture of products; Other, including openi, anthropic, Google and Meta, make and train large language models.
“There are only a handful of companies that can manufacture those types of models,” Voka said. “This is very capital-intensive. You need to spend billions of dollars, and a lot of companies do not have billions of dollars to spend a model. And as a result, those companies, the way they arrive on it: ‘If I am going to spend a billion dollars to make a model, $ 10 million is expected low for an engineer.”
Anthropic CEO Dario Amodi told Time magazine in 2024 that he expected Training Frontier AI Model Cost To be $ 1 billion that year.
AI Institute of Stanford University recently created a report shown Select the estimated cost of construction AI model Between 2019 and 2024. Openai’s GPT-4 costs $ 79 million to manufacture in 2023, for example, while Google’s Gemini was 1.0 Ultra $ 192 million. Meta’s Lama 3.1–405B cost $ 170 million in 2024.
“Companies that manufacture products, pay these existing models to use and to manufacture on top of them, so capital expenditure is low and does not have that much pressure to burn money,” Voca said. “The place where things are very hot in terms of salary are companies that are building models.”
AI experts are in demand
The average salary for machine learning engineers in the US is $ 175,000 in 2025, in fact according to data.
Pixelonestocker | Moment | Getty images
Machine learning engineers are AI professional who can manufacture and train these large language models – and demand for them is high on both sides of Atlantic, Ben Litvinoff, Associate Director of Technology Recruitment Company Robert Wallers, said.
Litvinoff said, “There is definitely a huge increase in demand in relation to both AI-focused analytics and machine learning, so people working with big language models and people deploy more advanced GPT-supported or more advanced AI-operated technologies or solutions.”
This includes a “slim talent pool” of experienced experts, who have worked in the industry for years, as well as AI research scientists who have completed their PhDs in the top five or six universities in the world and are flapping by technical veterans when graduated.
It leads to mega pay packets, allegedly with Zuckerberg Offer $ 250 million A 24 -year -old AI Genius Matte Deiteke, who was out of a computer science doctorate at Washington University.
Meta directed CNBC to information Zuckerberg’s comments, where the Facebook founder stated that there is a “full premium” for top talent.
“A lot of nuances that have been stated are not themselves are not accurate. But it is a very hot market. I mean, as you know, and a small number of researchers, who are the best, which are in demand by all different laboratories,” Zuckerberg told Tech Prakashan.
“The amount that is being spent to recruit people is still much lower than the overall investment and when you talk about super intelligence.”
Litvinoff estimated that, in the London market, machine learning engineers and principal engineers are currently earning six-hearted salaries from £ 140,000 to £ 300,000 for more senior roles.
In the US, the average salary for a machine learning engineer is $ 175,000, which has reached about $ 300,000 at high end, In fact,
Startups and traditional industries are left behind
Since Tech veterans continued the best mind in AI with the greed of mammoth pay, there is a risk that the startups leave behind.
“Some of these startups that are trying to compete at this location of the building model are difficult for them to see the way forward, as they are stuck at this place: are very expensive to make models, but companies that are buying those models, I do not know if they can pay prices that cover the cost of manufacturing models,” the Voyal said.
Mark Miller, Founder and CEO of InsureVision.Ai, Recently told to Startups magazine This talent war was also creating a “large -scale opportunity difference” in traditional industries.
Miller said, “The entire industry like insurance, healthcare and logistics cannot compete on salary. They need innovation, but cannot use talent,” Miller said. “The current situation is absolutely unstable. You may not have an industry that accumulate all talents, while others are withering.”
Voca said that AI professionals would have to make an option. While some will take the high salary and bureaucracy of the Big Tech, others will bend towards the startup, where the salary is low, but the employees have more ownership and effects.
“In a large company, you are essentially a coong in a machine, while in a startup, you can have a lot of effect. You can make a lot of effect through your work, and you feel that effect,” Voca said.
As long as the price of the construction of the AI model is not low, however, the AI talent is likely to remain high salary.
“As long as companies have to spend billions of dollars to make models, they will spend millions or millions of people to hire engineers to manufacture those models.”
“If suddenly tomorrow, the cost of making those models decreases by 10 times, the salary I expect will also be reduced.”
