Meta's ambitious spending plans for 2026 have taken the tech world by surprise. The company revealed on Wednesday that its capital expenditures could reach as high as $145 billion this year, a dramatic leap from $72 billion in 2025. The announcement came as part of Meta's quarterly earnings report, which also showed a 33% revenue increase—the fastest since 2021. Despite that positive metric, Meta's shares fell more than 7% in after-hours trading, reflecting investor anxiety over the massive outlay.
Why the Spending Is So High
The primary driver behind Meta's skyrocketing costs is the global memory chip shortage, exacerbated by the AI boom. CEO Mark Zuckerberg explained that higher component costs, especially for memory, account for most of the spending increase. The AI industry's insatiable demand for data centers has constrained the supply of high-bandwidth memory chips, driving up prices and creating a crisis that affects not only cloud providers but also consumer electronics like laptops and smartphones.
Meta is not alone in this spending spree. Rivals like Google, Amazon, and Microsoft also reported earnings on the same day, each committing tens of billions to AI infrastructure. However, Meta's $145 billion figure is the most aggressive, representing roughly double its prior-year cap-ex. The company is racing to build out its AI capabilities after falling behind in the generative AI race, particularly to Google, which has surged ahead with products like Gemini.
Zuckerberg's AI Turnaround Bet
Zuckerberg acknowledged the situation roughly 10 months ago, when he admitted Meta had been left behind. He launched a major catch-up effort, committing billions to research and development and poaching top talent. A key move was recruiting Scale AI's founder Alexandr Wang to lead Meta Superintelligence Labs, a new division dedicated to pushing the frontiers of AI.
The first fruit of this investment is Muse Spark, a proprietary AI model that Meta plans to open-source. The model was unveiled earlier this month and represents a step forward for the company's AI strategy. During the earnings call, Zuckerberg emphasized that the release demonstrates Meta's ability to build a leading lab. He also promised that users will see two new AI agents later this year: one for personal use and one for business applications. Early testing of business AIs has already shown a tenfold increase in weekly conversations since the start of 2026.
Lessons from the Metaverse Flop
Many analysts are cautious about Meta's AI bet, given the company's recent history with the Metaverse. Reality Labs, the division responsible for virtual and augmented reality, posted an operating loss of $4 billion in the latest quarter on just $402 million in sales. Since its inception, Reality Labs has accumulated over $80 billion in losses. While some experts believe AI has a clearer path to monetization than the Metaverse, the scale of Meta's spending remains a concern.
Zuckerberg, however, is confident that AI will generate returns through improved recommendation systems, hyper-personalized feeds, and advertising tools. Meta is already using AI to translate and dub videos for over half a billion weekly users on Facebook and Instagram. The company's recommendation system is being overhauled to leverage the new model, aiming to boost engagement and advertiser value.
AI-Driven Internal Changes
Ironically, as Meta invests heavily in AI, it is also cutting jobs. The company is laying off 10% of its workforce and offering voluntary buyouts to 7% of U.S. staff. While executives declined to directly link the layoffs to automation, CFO Susan Li noted that a "leaner operating model" will help offset the substantial investments. This trend reflects a broader Silicon Valley shift, where CEOs increasingly see AI as a tool to reduce headcount. A recent survey found that 99% of CEOs expect AI-driven layoffs within two years.
Meta's internal use of AI extends beyond video translation. The company is also incorporating AI into its ad systems, with early tests showing improved performance. Zuckerberg stated that the increasing returns from AI investments on engagement and advertiser value are becoming clear. He suggested that the scale of Meta's platforms—billions of users across Facebook, Instagram, and WhatsApp—makes AI-driven personalization particularly powerful.
The path forward for Meta is fraught with risk. Spending $145 billion on AI infrastructure while still licking wounds from the Metaverse is a gamble of historic proportions. However, the company's strong revenue growth and user base provide a cushion. If Zuckerberg's bet pays off, Meta could regain its position as a leader in artificial intelligence. If not, the losses could dwarf those of Reality Labs. For now, all eyes are on the two AI agents and whether they can deliver the transformation that Meta so desperately needs.
Source: Gizmodo News