How Meta Is Winning the AI Talent War

Everywhere you hear that AI is taking jobs. Artificial intelligence growing at lightning speed. Amazon doing another round of layoffs. Mark Zuckerberg's year of efficiency. Back in focus this morning after another round of sweeping layoffs and job cuts. But one group in particular has been the most affected. Tech workers. Let's zero in on the tech sector for a second here. That industry seen sharp decelerations in hiring in recent years. However, at the same time, big tech CEOs are offering multi-million dollar packages OpenAI's CEO Sam Altman addressing Mark Zuckerberg's Meta platforms. He's trying to hire away OpenAI's workers. Apple, for example, has lost several of its best developers due to the extraordinary salaries offered at Meta. Meta now notching another name in its expensive hunt for AI talent, poaching a top Apple executive for reportedly tens of millions of dollars. So if AI is taking jobs, how are developers earning millions? Over the past three years, the public narrative has been dominated by headlines about massive layoffs at tech companies and promises of automation thanks to AI. At the same time, in the hallways of those same companies and in financial media, seemingly contradictory stories are emerging. AI engineers and scientists receiving multi-million dollar offers. Offers even higher than those made to traditional senior executives. Entire teams are also being poached with compensation packages that, in some cases, exceed $100 million in multi-year deals. Without exception, absolutely every tech company is fiercely fighting for the best AI talent. Google's AI unit is now the latest target in the talent wars for top researchers. Meta is a clear example of this. The company has been investing more than $20 billion annually in artificial intelligence infrastructure. That spending includes data centers, chips, and above all, the hiring of experts capable of creating competitive models against OpenAI or Google DeepMind. One of the most notable cases was Ruoming Peng, a senior Apple engineer. In 2025, Meta signed him with a contract that, according to Reuters, could reach $200 million over several years. Peng led key projects at Apple, and his departure was seen as a sign of the AI talent war. Peng's hiring was not an isolated case. According to Business Insider, Meta offered annual packages of $10 to $20 million to several AI researchers. Meta poaching now eight, as far as we know, top researchers from OpenAI. Let's assume, based on reports, eight packages at $100 million each. That's nearly a billion dollars in comp. By comparison, the average salary for a software engineer in the U.S. in 2024 was $124,000 per year, according to data from the Bureau of Labor Statistics. In other words, some AI engineers are earning more than 100 times what an average developer makes. This phenomenon has an economic logic. Training a large language model can cost between $50 and $100 million, depending on scale and GPU usage. Hiring an engineer who can optimize that process and reduce costs can mean savings of tens of millions. That is why, from the perspective of a company investing billions, Peng's $10 million salary seems small in comparison. The contrast with layoffs remains striking. The largest tech companies have laid off around 25,000 employees so far this year. Mounting layoffs in the tech industry have many concerned about the changing landscape. Experts are weighing in about the fear of artificial intelligence replacing people. Most of these layoffs are concentrated in support, marketing, human resources, and mid-level development roles. These are positions that are easier to replace with automated tools. Generative AI, for example, can handle writing tasks, software testing, or basic data management. In that same context, the demand for AI experts skyrocketed. LinkedIn data shows that AI job postings grew 33% in 2024, while overall tech hiring fell 21% in the same period. This gap explains why some are losing jobs while others are multiplying their salaries. The paradox runs even deeper because the highest-paid engineers are precisely the ones working on projects designed to reduce the need for more workers. Tools like GitHub Copilot, powered by AI, already generate more than 40% of the code in popular software projects, according to Microsoft. This means that part of the work once done by junior developers is now carried out by automated systems. At Google, an internal report revealed that the use of AI in programming tasks increased productivity by 35% among teams that adopted it at the beginning of 2023. Generative AI, it's the buzzword of 2023, and it's safe to say artificial intelligence has become a growing force for businesses. At the same time, this reduces the need to hire more staff for repetitive functions. In other words, the highest-paid engineers are designing the systems that could eventually replace them. Today, they are indispensable, but in the medium term, they risk seeing their own creations diminish the value of their work. According to experts, the job market is being segmented into three levels. At the base, workers replaceable by AI or automated processes. In the middle tier, professionals facing wage stagnation and greater competition. And at the top, an elite of AI engineers negotiating million-dollar contracts. This dynamic is reflected in salaries. In 2024, a software engineer at Meta earned on average $210,000 per year. But the exceptional packages for AI researchers exceeded $5 million annually, not counting stock options and long-term bonuses. If you're really in this race and the leaders in the race are sometimes weeks ahead of the guy behind them, you're going to pay up for the top talent. This concentration also creates inequalities within companies. While Meta offered tens of millions to a single engineer, it maintained hiring freezes in other areas and demanded cost cuts from entire departments. This generates internal tensions and a perception of unfairness. Investment in artificial intelligence is also growing at an accelerated pace. McKinsey estimates that AI adoption could add up to $4.4 trillion to the global economy every year. In 2024 alone, AI startups raised more than $50 billion in venture capital investment, a record figure. Hardware spending is another key indicator. NVIDIA reported revenues of $22 billion in just the first quarter of 2024, a 427% increase compared to the previous year, thanks to demand for chips to train AI models. NVIDIA, the AI titan, largely seen as the most valuable company in the world, putting out a better-than-expected earnings report. This shows the scale of money moving around AI and partly justifies paying millions to those who know how to leverage it. But the social dilemma remains. If companies spend billions on AI to cut labor costs, the result is higher productivity and fewer available jobs. In other words, the benefit is concentrated among shareholders and a small elite of engineers, while the majority face increasing precariousness. Some experts believe this trend will lead to even sharper polarization. Economists warn that AI could reproduce the inequality already created by automation in the manufacturing industry. A few benefit enormously, while most see their wages stagnate or their jobs disappear. AI is rapidly reshaping the workplace. More companies are turning to it as a substitute for hiring or retraining workers, and that's raising urgent questions about the future of the job market. The reality is clear. While hundreds of thousands of employees lose their jobs, a handful of specialized developers are earning more than ever. But the question is how sustainable this will be over time. High-profile cases, such as Meta's signing of Ruomingpeng with a multi-million dollar package, illustrate the premium the market places on AI expertise and leadership. However, that same technology that elevates the value of certain individuals also has the potential to reduce the need for human labor in other areas. And in the long run, it could even replace functions currently carried out by those same specialists.