Wednesday was a landmark day for the tech industry, with Meta, Google, Amazon, and Microsoft all reporting earnings simultaneously in the afternoon. Among the four, Meta emerged as the clear loser, its shares tumbling more than 7% even though revenue surged 33% in the past quarter—the company's fastest growth since 2021. The likely culprit: Meta raised its already staggering spending expectations for the year, announcing that 2026 capital expenditures would be at least $10 billion higher than previously estimated, potentially reaching as much as $145 billion.
During the earnings call, CEO Mark Zuckerberg defended the massive outlay, stating, "I have confidence in this investment." He attributed the bulk of the increase to "higher component costs, particularly memory pricing." The AI boom has triggered an unprecedented data center buildout across the industry, tightening the global supply of memory chips and driving prices upward. This memory crisis has not only affected Meta and its AI competitors but also pushed up the prices of consumer electronics like laptops and smartphones.
To put the $145 billion figure in perspective, Meta's capital expenditure in 2025 was $72 billion. Doubling down in a single year signals an all-in bet on AI—a bet that comes after the company's previous high-stakes venture into the Metaverse, spearheaded by Reality Labs, has largely floundered. In the same earnings report, Reality Labs posted an operating loss of more than $4 billion on just $402 million in revenue, adding to the division's cumulative losses exceeding $80 billion over the past six years. Yet Zuckerberg remains undeterred, framing AI as the next frontier where Meta cannot afford to lag behind.
The AI Catch-Up Effort
Meta has been playing catch-up in the AI race, with rivals like Google and Microsoft racing ahead. Roughly ten months ago, Zuckerberg acknowledged the situation and launched an aggressive turnaround plan. That plan involved pouring billions into research and development, poaching top talent from across the industry, and establishing Meta Superintelligence Labs—a new division led by Scale AI's founder Alexandr Wang. The division aims to develop leading-edge AI models that can compete with OpenAI's GPT, Google's Gemini, and other frontier systems.
Earlier this month, Meta debuted its first major output from this lab: Muse Spark, a proprietary AI model that the company plans to eventually open-source. Industry analysts view Muse Spark as a positive step, but they caution that Meta still has much work to do before it can claim a successful catch-up. "This was the first release from Meta Superintelligence Labs, and it shows that our work is on track to build a leading lab," Zuckerberg assured investors. "Now that we have a strong model, we can develop more novel products as well."
Among those novel products will be two AI agents: one for personal use and one for business use. Zuckerberg revealed that the business agent is already in early testing, with weekly conversations growing tenfold since the start of 2025. "We're already seeing promising engagement," he said. The consumer-focused agent is expected to arrive later this year, likely integrated into Meta's existing platforms like Facebook, Instagram, and WhatsApp.
Internal AI Adoption and Workforce Changes
AI is also transforming Meta from within. The company is laying off 10% of its workforce and offering voluntary buyouts to 7% of U.S. staff, following a trend that has swept Silicon Valley where automation is increasingly replacing human roles. While executives declined to directly link the layoffs to AI automation, CFO Susan Li noted that a "leaner operating model" would help "offset the substantial investments we're making." The cuts are expected to save billions annually, helping to fund the AI expansion.
One area where AI is already yielding tangible results is content localization. Meta's AI-powered translation and dubbing tools now serve over half a billion users weekly on Facebook and Instagram, enabling videos to reach a global audience without human translators. The company is also integrating the new Muse Spark model into its core ad business and recommendation algorithms, aiming to hyper-personalize feeds for users. "Since our recommendation systems are operating at such large scale, we'll phase in this new research and technology over time," Zuckerberg explained. "But the trend over the last few years seems clear: we are seeing an increasing return on the amount that we can improve engagement for people and value for advertisers."
The broader context for Meta's spending spree is the global competition for AI dominance. Rivals like Google have spent similar sums on data centers and chip supply, though Meta's $145 billion figure stands out as the highest single-year capital expenditure ever projected by a major tech company. Analysts have expressed concern that the company may be overinvesting, especially given its checkered history with the Metaverse. However, supporters argue that AI is a foundational technology that could reshape Meta's entire business model, reducing dependence on advertising revenue and opening new streams from AI-powered services.
Memory chip prices have become a critical input cost. The AI boom has driven demand for high-bandwidth memory (HBM) used in GPU clusters, straining supply chains. Major memory manufacturers like Samsung and SK Hynix have prioritized HBM production, squeezing supply for standard DRAM and NAND flash, which in turn has raised prices for consumer goods. Meta's increased spending partially reflects these inflated prices, and the company expects them to persist through 2026.
In the earnings call, Zuckerberg struck a defiant tone, emphasizing that the AI investment is a long-term play. He pointed to early successes in ad relevance and user engagement as proof that AI can boost the company's core business. Meta's revenue growth of 33%—the highest in four years—was driven in part by AI-enhanced ad targeting, which commands higher prices from advertisers. If the AI bet pays off, Meta could emerge as a dominant force in both consumer AI and enterprise AI tools.
But the road ahead is fraught with risks. The company's shares fell despite strong earnings because investors are wary of the ballooning expense. Meta's net income for the quarter was healthy, but the specter of $145 billion in capital spending—more than the GDP of some small countries—raises questions about profitability and return on investment. Zuckerberg has promised that the spending will moderate after 2026, but he offered no specifics.
Meanwhile, the launch of the two AI agents later this year will be a key test. If they can attract millions of users and demonstrate clear utility, Meta could justify its enormous outlay. If they fall flat, the company may face renewed pressure from shareholders. The memory crisis shows no signs of abating, and competitors continue to pour money into AI. For now, Zuckerberg is betting the company on artificial intelligence, hoping that this time, his vision will pay off.
Source: Gizmodo News