Investing.com -- Meta (NASDAQ: META) is considering major layoffs that could affect more than a fifth of its workforce as the company ramps up spending on artificial intelligence infrastructure, Reuters reported over the weekend, citing people familiar with the matter.

The company’s shares rose in premarket trading by more than 3% by 05:18 ET.

The potential cuts have not been finalized and no timeline has been set, the report said. Top executives have recently signaled their plans to senior leaders and asked them to begin preparing how to pare back, as Meta looks to offset costly AI infrastructure bets and prepare for greater efficiency from AI-assisted workers.

If the company ultimately proceeds with cuts of around 20%, it would mark Meta’s most significant layoffs since a restructuring in late 2022 and early 2023 that it dubbed the “year of efficiency.” Meta employed nearly 79,000 people at the end of last year.

The discussions come as CEO Mark Zuckerberg pushes the company to compete more aggressively in generative AI. Meta has offered large pay packages, some worth hundreds of millions of dollars over four years, to recruit top AI researchers for a new superintelligence team.

The company has said it plans to invest $600 billion to build data centers by 2028 to support its AI ambitions. Meta has also been expanding its AI efforts through deals and acquisitions, including the purchase of Moltbook, a social networking platform built for AI agents, and plans to spend at least $2 billion to acquire Chinese AI startup Manus, Reuters previously reported.

Analysts say the reported layoffs reflect the growing cost pressures tied to the AI arms race but also highlight the potential productivity gains from the technology.

"We think the report underscores both the higher costs of AI infrastructure but also cost benefits to R&D heavy companies from coding and other efficiencies," analyst Justin Post said in a note.

He estimates that a reduction of roughly one-fifth of Meta’s workforce could generate about $7 billion to $8 billion in annual savings, assuming average employee costs of around $500,000.

"Based on cost commentary in the article, we do not expect Meta to materially lower its FY26 expense guide of $162-$169bn, though we view the report as suggesting cost discipline at Meta vs outlook," Post added.

Separately, JPMorgan analyst Doug Anmuth offered a similar view, estimating that a 20% headcount reduction could produce roughly $5 billion to $6 billion in cost savings, assuming costs of about $300,000 to $400,000 per employee.

However, he added that such savings would still represent a relatively small offset compared with Meta’s rapidly rising expense base as the company accelerates investment in AI infrastructure and related depreciation.

"But still, if the $6B were added to our 2027 profit and tax-affected, it would result in ~$2 in incremental GAAP EPS above our current $31.50 projection," Anmuth continued.

Jefferies analyst Brent Thill said the "reported ~20% headcount reduction would reinforce that AI is beginning to deliver real productivity gains at scale, while helping offset a significant AI capex ramp."

"The takeaway is not just better Meta margins, but a broader read- through for tech/software as investors reassess the link between headcount, growth, & profitability," he added.