On a Wednesday that saw four of the largest technology companies report earnings simultaneously, Meta Platforms Inc. stood out for all the wrong reasons. Despite reporting a 33% year-over-year revenue increase—its fastest since 2021—the company’s stock dropped more than 7% in after-hours trading. The culprit was a staggering upward revision of capital expenditure expectations for 2026, now projected to reach at least $145 billion, fueled by the escalating costs of artificial intelligence infrastructure.
The new forecast represents an increase of at least $10 billion from previous guidance and more than double the $72 billion Meta spent in 2025. CEO Mark Zuckerberg attributed the bulk of the increase to “higher component costs, particularly memory pricing.” The AI boom has triggered an unprecedented global buildout of data centers, straining the memory chip supply chain and sending prices soaring. This “memory crisis” has not only affected hyperscalers like Meta but also trickled down to consumer electronics, driving up the cost of laptops, smartphones, and other devices.
The Context of Meta’s AI Spending Spree
Meta’s aggressive spending comes at a time when the company is playing catch-up in the AI race. For years, rivals like Google have dominated with large language models and generative AI tools, while Meta’s earlier bets—most notably on the metaverse—have yet to pay off. The Reality Labs division, which oversees Meta’s virtual and augmented reality efforts, posted an operating loss of more than $4 billion in the first quarter of 2026 on just $402 million in revenue. Since its inception, Reality Labs has accumulated losses exceeding $80 billion.
Zuckerberg acknowledged the competitive pressure roughly 10 months ago, announcing a major pivot toward AI. The centerpiece of that strategy is the formation of Meta Superintelligence Labs, a new R&D division led by Alexandr Wang, founder of Scale AI. Wang was poached from one of the most prominent AI infrastructure companies to oversee Meta’s push to build cutting-edge models and applications.
“This was the first release from Meta Superintelligence Labs, and it shows that our work is on track to build a leading lab,” Zuckerberg told investors during the earnings call, referring to the company’s new AI model, Muse Spark. The model is proprietary but Meta plans to open-source it in the future, a move that aligns with its tradition of releasing major AI research to the broader community.
Two AI Agents on the Horizon
Zuckerberg also previewed two new AI agent products—one aimed at personal use and another for business applications. While details remain scarce, he noted that early versions of business AIs are already being tested and that weekly conversations with these agents have grown tenfold since the beginning of 2026. The agents are part of Meta’s broader effort to monetize AI and regain ground lost to competitors.
The company is also integrating AI deeply into its core platforms. CFO Susan Li revealed that more than 500 million weekly users on Facebook and Instagram now consume videos that have been translated and dubbed by Meta’s AI. Additionally, the recommendation algorithm—responsible for surfacing content to billions of users—is being re-engineered to leverage the new model and deliver hyper-personalized feeds. “Since our recommendation systems are operating at such large scale, we’ll phase in this new research and technology over time,” Zuckerberg said. “But the trend over the last few years seems clear that we are seeing an increasing return on the amount that we can improve engagement for people and value for advertisers.”
The Memory Crisis and Its Ripple Effects
Meta’s soaring capital expenditure is not an isolated phenomenon. The AI-driven demand for high-bandwidth memory (HBM) and advanced logic chips has created a global shortage. Memory manufacturers like Samsung, SK Hynix, and Micron have ramped up production, but prices have risen sharply due to supply constraints. This has forced companies like Meta to pay significantly more for the chips necessary to train and deploy large AI models.
Analysts have noted that the memory crisis could persist for several more quarters, as new fabrication capacity takes time to come online. In the meantime, hyperscalers and enterprises face a difficult choice: either pass on higher costs to consumers or absorb them into already stretched budgets. For Meta, the decision has been to invest heavily now in hopes of reaping long-term AI dividends.
Some investors remain skeptical, pointing to the metaverse debacle as a cautionary tale. However, experts argue that AI has more proven revenue potential than virtual worlds. “The difference is that AI is already generating tangible value in advertising, content personalization, and internal efficiency,” said a tech industry analyst. “Meta is late to the party, but it has the resources to catch up if it spends wisely.”
Internal Restructuring and Layoffs
To help offset the massive AI investments, Meta is also cutting costs elsewhere. The company announced it would lay off 10% of its workforce and offer voluntary buyouts to an additional 7% of its U.S. staff. This follows a broader Silicon Valley trend of automating roles traditionally performed by humans, though executives declined to confirm whether the layoffs were directly tied to AI automation. Li described the moves as part of a “leaner operating model” that would help “offset the substantial investments we’re making.”
Internally, AI is already taking over many functions. Meta has deployed AI tools for coding, content moderation, and customer support. The company’s open-source large language models have also been adopted by outside developers, creating an ecosystem that could eventually feed back into Meta’s own products. Zuckerberg emphasized that the company’s long-term vision remains centered on building AI that can assist people in both their personal and professional lives.
Despite the immediate market reaction, Meta’s revenue growth indicates that its core advertising business remains strong. The company is betting that the new AI capabilities will further boost advertiser returns and user engagement, justifying the historic spending levels. As the global memory crisis evolves and new competitors emerge, Meta’s ability to execute on its AI strategy will determine whether this $145 billion gamble pays off or becomes a costly footnote in its history.
Source: Gizmodo News