You thought AR and VR were expensive? Wait 'til you see what Meta wants to spend on AI this year

The next round of Meta layoffs began this week, with an estimated 5% staff reduction count topping out around 4,000 people. Many employees received notification earlier this week, giving them an hour to assemble any last-minute correspondence before access was cut off.

Hot on the heels of the layoff is news that Meta is rapidly rehiring many of these positions, per Business Insider, pivoting them to AI development. Meta's latest earnings show it's still mainly an ad-driven company, with over 3.3 billion people being served ads in some capacity across the company's many platforms. AI, it's said, will be the next major way to deliver ads in a more profound, personalized way.

But the amount of money Meta plans to spend on AI development in 2025 is staggering. Zuckerberg told investors the company plans to spend upwards of $65 billion on AI this year, citing it as a key component of the company's future revenue strategy.

To put that in perspective, Meta has spent roughly $60 billion on its Reality Labs division since 2019. Now, Meta wants to spend another $5 billion more than that on AI in one-fifth of the time. It's a huge shift in a company that has been spending billions on products and R&D over years, not months.

While Reality Labs has cost Meta lots of money over the years, the group has been turning healthy revenue numbers since 2020, which could outpace spending if Meta cuts down on R&D costs. Thus far, Reality Labs has offset its 5-year costs by pulling in nearly $10 billion in revenue. The revenue guidance for Reality Labs is straightforwar,d as the company mainly profits from the sales of devices, services, and apps sold on the Meta Horizon digital marketplace.

But the way Meta makes money on Meta AI isn't nearly as clear or straightforward. Meta has been injecting AI into every product and service it has, from products like Ray-Ban Meta Smart Glasses and the Meta Quest 3 to apps like Instagram and WhatsApp. Zuckerberg told investors last April that Meta AI will primarily make the company money through ad revenue.

As Stratechery notes, this is done through enhanced machine learning techniques — using AI to better understand Meta's customer base and serve even more personalized ads to it — as well as generative visuals targeted toward specific customers. In other words, Meta AI is being used as an ad delivery vehicle, not a product in the traditional sense.

But Meta's strategy for getting AI into its products feels disconnected. Thousands of Meta employees are now jobless after receiving positive performance reviews, made even worse by the company publicly calling them "low performers." Everyone from Instagram product designers to fledgling data engineers and even longer-term product counselors are part of the layoffs, and many cite having received no warning ahead of yesterday's notice.

It's true that Zuckerberg did give some advanced notice of upcoming layoffs, but the CEO said that the company would be laying off "low performers" during January 29th's earnings call — not employees that just received positive performance reviews. The surprise is deepened further by the company's earning reports, which showed a record $48 billion in revenue, a 21% increase over the previous year.

Per internal memos obtained by Business Insider, managers were to identify between 12% and 15% of their employees to be labeled as "low performers." This comes after the previous round of layoffs in October 2024, cutting each of the company's departments by another 7-10% to meet the goals outlined by Zuckerberg. Employees who spoke with Business Insider noted that a culture of anxiety and fear had swept through the company.

To make matters worse, Zuckerberg said that the company would be "backfilling these roles this year," meaning folks laid off this week will likely have their roles replaced by the year's end. It screams short-term thinking without considering long-term repercussions, something the tech industry has been guilty of as it began mass layoffs across the sector a few years ago.

It's said that Meta's strategy now is to blitz-hire AI engineers to ensure the company is a leader in the AI space. Our current understanding of the company's strategy is that the backfilled roles will pivot to AI development instead of their previous scope, ensuring that Meta's upcoming hires are laser-focused on the company's latest goal.

Meta's CFO, Susan Li, directly addressed the hiring process in the recent earnings call. "In 2025, we expect headcount growth will continue to be primarily driven by technical roles across our priority initiatives within infrastructure, monetization, Reality Labs, generative AI, as well as regulation and compliance. We anticipate headcount growth in our business functions will remain relatively limited."

This leaves even more uncertainty in the wake of several rounds of layoffs during 2023 and 2024 despite Meta labeling only 2023 as the "year of efficiency." Now, even good employees with a track record of excellence are being replaced to further a single, specific goal.

While Meta's strategy seems dubious from many angles, the company deserves credit for its generous severance packages. Impacted folks were given "16 weeks of pay and an additional two weeks for every year of service," according to two folks who spoke with Business Insider. A similar amount of severance pay was offered to laid-off workers during several other rounds of lay-offs, according to people I had previously spoken with.

And Meta's moves are certainly helping stock owners profit from a heightened valuation. Meta stock is up 200 points (roughly 40%) since October when the latest string of layoffs began. If anything, the company's strategy is resonating well with investors who believe the company to be on the right track.

But how much of this will come back to bite Meta in the proverbial butt? Ousting employees with positive performance reviews feels underhanded unless decisions are being made for financial reasons. Given how well Meta has been doing and how much money the company continues to rake in, this is certainly not the case.

Bleeding talent is a huge problem across the tech sector and certainly not siloed to Meta, but one wonders why the company can't simply hire more AI engineers and keep the rest of the employees onboard during the process. After all, if profits only continue to go up, isn't it madness to get rid of the people who made that possible? It certainly seems so to me.