^{Photo: NDTV Profit}

Investors Give Meta the Go Ahead to Spend Big on AI

Meta CEO Mark Zuckerberg is preparing to dramatically increase the company’s artificial intelligence spending in 2026, and this time Wall Street appears firmly on board.

After years of investor unease over Meta’s aggressive tech investments, the company’s latest earnings report reassured markets that its core business remains strong enough to bankroll an ambitious AI push. Meta beat expectations on both revenue and profit in the fourth quarter, prompting its shares to jump as much as 10% in after hours trading.

The company also revealed that AI related capital expenditures for the coming year will land between $115 billion and $135 billion, nearly double what Meta spent on capex last year. The funds will primarily go toward expanding data centers, acquiring advanced chips, and building the infrastructure needed to train large scale AI models.

Zuckerberg made it clear that Meta is not slowing down.

“As we plan for the future, we will continue to invest very significantly in infrastructure to train leading models and deliver personal super intelligence to billions of people and businesses around the world,” he told analysts during the earnings call.

Ad Business Delivers the Cash to Fuel AI Ambitions

Meta’s renewed credibility with investors rests on one key pillar: advertising.

The company posted 24% year over year revenue growth in the fourth quarter, driven largely by strength across Facebook, Instagram, and WhatsApp advertising. Improvements in AI powered ad targeting and recommendation systems have boosted engagement and conversion rates, helping Meta generate tens of billions of dollars in quarterly revenue and billions more in free cash flow.

That financial performance has given Zuckerberg room to maneuver. Although Meta stock lagged broader markets last year, the latest results showed that its ad engine remains highly profitable, allowing the company to fund massive infrastructure projects without jeopardizing its balance sheet.

Online advertising still represents the overwhelming majority of Meta’s revenue. As long as that machine keeps producing cash at scale, investors appear willing to tolerate heavy spending on longer term AI bets.

Meta Is Still Starving for Computing Power

Chief Financial Officer Susan Li told analysts that Meta remains “capacity constrained,” meaning the company simply does not have enough computing power to support both its rapidly improving ad systems and its expanding AI research efforts.

“Our teams have done a great job ramping up our infrastructure through the course of 2025, but demands for compute resources across the company have increased even faster than our supply,” Li said.

This shortage is a major reason behind Meta’s ballooning capex plans. The company is building massive new data centers and acquiring large volumes of advanced GPUs and custom silicon to train increasingly complex models.

Industry estimates suggest that Meta now ranks among the world’s largest buyers of AI accelerators, alongside hyperscalers such as Microsoft, Google, and Amazon. Much of the upcoming $115 billion to $135 billion spending range will be allocated to these efforts.

A Big Swing on Talent and Core AI Models

Beyond hardware, Zuckerberg has also been investing aggressively in talent and foundational AI technology.

One of Meta’s most notable moves last year was its $14.3 billion investment in Scale AI, a deal that brought founder Alexandr Wang and several senior engineers and researchers into Meta. Wang now leads Meta’s newly formed AI unit, which is working on a next generation frontier model code named Avocado, designed to succeed the company’s Llama family of open models.

Zuckerberg said Meta expects its first wave of new models to demonstrate meaningful progress quickly.

“I expect our first models will be good but, more importantly, will show the rapid trajectory that we’re on,” he said. “And then I expect us to steadily push the frontier over the course of the year as we continue to release new models.”

The company has already integrated AI features across its apps, from content recommendations to creative tools for advertisers. However, large scale, standalone AI products capable of generating meaningful new revenue streams have yet to fully materialize.

When pressed on what users should expect in 2026, Zuckerberg emphasized breadth over a single flagship launch.

“We’re not just launching one thing, and we’re building a lot of things,” he said.

Why Meta Insists on Owning Its Own AI Stack

Zuckerberg also addressed why Meta is determined to develop its own powerful foundation models rather than relying entirely on third party platforms.

He described Meta as a “deep technology company” that cannot afford to be dependent on external ecosystems.

If Meta does not control its own models, Zuckerberg warned, it risks being limited by what others allow it to build. Owning the core technology gives Meta the ability to shape how AI is embedded across social platforms, messaging apps, advertising tools, and future consumer products.

That philosophy mirrors strategies at other tech giants, where vertical integration across chips, data centers, and models is increasingly viewed as a competitive necessity.

Wall Street’s Patience Hinges on Ads and Execution

For now, investors appear willing to give Zuckerberg wide latitude, largely because Meta’s advertising business continues to exceed expectations and generate enormous quarterly cash flows.

The bigger question is timing. While Meta is pouring unprecedented sums into AI infrastructure and research, it remains unclear when these investments will translate into major new revenue lines beyond advertising.

Still, with strong earnings momentum, expanding margins, and a dominant position in global digital advertising, Wall Street seems prepared to let Zuckerberg pursue his vision of “personal super intelligence.”

As long as Meta’s core platforms keep delivering growth and profitability, the company’s multibillion dollar AI gamble is likely to retain market support, even if tangible payoffs take years to fully emerge.

‍