Metas AI Gamble Pays Off

Heavy spending plan fails to dent investor confidence after strong quarter

Meta Platforms delivered record quarterly revenue and stronger-than-expected earnings, sending its shares sharply higher despite outlining an aggressive expansion in artificial intelligence infrastructure that could see capital expenditure almost double.

The social media group said revenues rose 24 per cent year on year in the fourth quarter to $59.9bn, ahead of market expectations of $58.4bn. Net income climbed 9 per cent to $22.8bn, also beating forecasts of $21bn. Investors responded swiftly, pushing shares in Meta Platforms up around 8 per cent in pre-market trading.

The results provided reassurance after months of scrutiny over chief executive Mark Zuckerberg’s decision to commit tens of billions of dollars to building advanced AI systems.

Expanding AI ambitions

Meta forecast capital expenditure of between $115bn and $135bn in 2026, significantly above analysts’ estimates of roughly $110bn. By comparison, the company spent $72bn in 2025. Total expenses this year are projected to range from $162bn to $169bn, driven largely by infrastructure and talent investment tied to AI.

Zuckerberg told investors the company was entering what he described as a “major AI acceleration”, adding that he expected the momentum to build further through 2026.

The scale of the investment underlines Meta’s determination to compete at the frontier of artificial intelligence, an arena currently dominated by rivals including OpenAI and Google. Over the past year, Meta has intensified efforts to develop what Zuckerberg has termed “personal superintelligence”, recruiting leading researchers and expanding its global network of data centres.

Wall Street had previously reacted nervously to the company’s infrastructure build-out. Following its October earnings report, shares fell more than 11 per cent in a single session, erasing nearly $208bn from its market capitalisation after investors digested plans for higher data centre spending.

This time, the strength of operating performance appears to have shifted the mood.

AI lifts core advertising business

Meta’s core advertising operations, spanning Facebook, Instagram and WhatsApp, continue to fund its technological push. Zuckerberg credited “AI-driven performance gains” and robust holiday demand for the quarter’s growth.

Artificial intelligence is already embedded in Meta’s recommendation systems and advertising tools. Zuckerberg said today’s feed and ad targeting systems are “primitive compared to what will be possible soon”, arguing that increasingly sophisticated large language models will allow the company to deliver more personalised content and higher conversion rates for advertisers.

The strategy is clear. Rather than relying solely on speculative long-term AI bets, Meta is applying machine learning advances directly to its existing social platforms, where incremental improvements in engagement translate quickly into revenue gains.

Revenue for the current quarter is forecast at between $53.5bn and $56.5bn, again ahead of expectations, reinforcing confidence that the advertising rebound remains intact.

New models and monetisation

The company’s next generation of large language models is expected to be released in stages over the coming months. Meta’s previous Llama 4 model, launched last April, failed to meet some internal and external expectations. Zuckerberg struck a more measured tone this time, describing forthcoming models as “good” and promising steady progress at the frontier over the course of the year.

Internally, Meta has been working on a new model, codenamed Avocado, designed to compete more directly with the latest offerings from Google. The objective is not only technical parity but commercial impact, with new subscription, advertising and commerce opportunities built around AI-driven services.

Meta has also strengthened its position in AI agents through the acquisition of Manus, an AI agent platform that already counts a substantial number of business subscribers. Zuckerberg said the technology would be integrated into Meta’s broader product suite. Chinese authorities are reviewing the transaction over potential export control issues, though the company expects to proceed.

Infrastructure on an unprecedented scale

Earlier this month Meta unveiled Meta Compute, an initiative aimed at constructing what Zuckerberg described as “hundreds of gigawatts” of AI data centre capacity over the coming decades.

The scale is extraordinary. A single gigawatt of data centre capacity can cost tens of billions of dollars and requires electricity output comparable to that of a nuclear reactor. Building at such magnitude will demand vast capital, long-term power agreements and increasingly efficient hardware.

Zuckerberg said Meta is investing heavily in developing its own chips to reduce dependence on third-party suppliers. He also expressed confidence that the cost per gigawatt would decline over time as the company optimises its technology and supply chain.

At the same time, Meta is deploying AI internally to enhance productivity, including coding tools intended to improve engineers’ output and shorten development cycles.

Strategic reshaping

While spending on AI surges, Meta is scaling back elsewhere. The company has pared back its lossmaking metaverse ambitions, cutting around 1,500 jobs earlier this year in its Reality Labs division and closing several virtual reality gaming studios.

The pivot reflects a broader shift from immersive virtual worlds towards practical AI-driven hardware. Sales of Meta’s Ray-Ban smart glasses more than tripled in 2025, Zuckerberg said, suggesting that AI-enabled wearables may offer a clearer commercial pathway than virtual reality environments.

Leadership changes reinforce this strategic turn. Meta recently appointed former Goldman Sachs executive Dina Powell McCormick as president and vice-chair. A former senior official in President Donald Trump’s first administration, she has been tasked with developing partnerships with governments to finance and deploy large-scale data centres worldwide.

Balancing risk and reward

The question for investors is whether Meta can sustain its earnings momentum while absorbing the heaviest capital spending cycle in its history.

For now, the numbers are supportive. Strong cash generation from advertising is underwriting the AI build-out, and revenue guidance indicates continued resilience in the core business.

Yet risks remain. Building vast computing capacity ahead of clear demand carries execution and financial hazards. Regulatory scrutiny of AI technologies continues to intensify globally. Competition from other model developers remains fierce.

Meta’s latest quarter suggests that, at least in the near term, markets are prepared to grant Zuckerberg the latitude he seeks. Record sales have bought time. Whether the company’s ambitious infrastructure programme ultimately produces durable competitive advantage will determine if this AI acceleration proves transformative or simply expensive.