Meta’s Nuclear Ambition: Powering the AI Arms Race

Meta is forging an unconventional path in the AI infrastructure race, turning to nuclear power to fuel its next-generation data centers. This strategic pivot comes amid escalating regulatory scrutiny, particularly from China, which is investigating the company's recent acquisition of an AI startup. These developments position Meta at the heart of a new global competition encompassing compute power, technological sovereignty, and regulatory control.

A significant challenge has surfaced on the regulatory front. Chinese authorities have initiated a review of Meta's approximately $2 billion acquisition of the AI startup Manus. The probe centers on potential violations of export controls for sensitive technology.

Although headquartered in Singapore, Manus has Chinese origins. The company developed a "general purpose" AI agent capable of autonomously handling complex, multi-step tasks. Its commercial success was rapid, achieving an annual recurring revenue exceeding $100 million within just eight months of its product launch.

Market analysts interpret China's move as a warning to firms with Chinese ties that relocate operations to Singapore to circumvent domestic regulations—a practice often termed "Singapore washing." This involves a formal shift to a more liberal jurisdiction while core technology, talent, and value creation often remain linked to China. According to experts like Wendy Chang from the Mercator Institute for China Studies, Beijing aims to stem the permanent outflow of critical AI technology and skilled personnel through foreign acquisitions, especially by U.S. corporations.

Securing Megawatts for the "Prometheus" Cluster

The centerpiece of Meta's energy strategy is its planned "Prometheus" supercluster, a high-performance data center in Ohio. To secure a long-term, stable power supply for this energy-intensive project, Meta has entered agreements with three nuclear power providers: Vistra, TerraPower, and Oklo.

Collectively, these deals are projected to unlock roughly 6.6 gigawatts of capacity by 2035—an amount exceeding the current total power consumption of the entire state of New Hampshire. The specifics of the partnerships are:

• Vistra: Meta is providing funding to extend the operational life and increase output from nuclear plants in Ohio and Pennsylvania.
• TerraPower: The company is backing two projects slated to deliver power from 2032, with options agreed for up to six additional projects by 2035.
• Oklo: A campus featuring advanced nuclear technology is planned for Pike County, Ohio, by 2030, intended to supply Meta's computing infrastructure.

The market response was notably positive for Meta's partners: Vistra shares posted a double-digit gain following the announcement, while Oklo saw a high single-digit percentage increase. Notably, OpenAI CEO Sam Altman holds a roughly 4.3% stake in Oklo, a position valued at approximately $650 million. For Meta, these agreements signal to investors a commitment to securing stable, predictable, and politically viable energy sources to advance its AI ambitions independently of volatile power markets.

Should investors sell immediately? Or is it worth buying Meta?

Massive Investments Divide Wall Street

Beyond these headlines, Meta's substantial capital expenditure plans continue to be a focal point for investors. The stock currently trades about 7.8% below its 52-week high of $708, closing Friday at $653.06 after a gain of just over 1% from the previous session.

The catalyst for a recent valuation discount was the October 2025 earnings report. Meta announced it would increase investments to $70-$72 billion for 2025, with management projecting "significantly stronger" expenditure growth for 2026. Market interpretation suggests investments could surpass the $100 billion threshold.

For context, Meta generated nearly $110 billion in operating cash flow over the past twelve months. Effectively, all of this incoming cash is being channeled into AI infrastructure and data centers. This aggressive strategy continues to split opinion on Wall Street:

• Proponents argue it could secure a massive, enduring lead in the AI infrastructure race.
• Critics highlight concentration risks and the danger that such enormous outlays may not translate quickly enough into sustainable revenue and profit growth.

This divide is reflected in the wide range of price targets, which span from $685 to $1,117, underscoring the uncertainty regarding the return on these mega-investments.

Valuation Metrics and the Path Ahead

Current valuation multiples place Meta in an elevated, though not extreme, growth segment. Based on trailing twelve-month figures, the price-to-earnings (P/E) ratio stands at just under 29, with a forward P/E of approximately 22. The company commands a market capitalization of about $1.65 trillion and offers a relatively low dividend yield of around 0.3%.

From a technical analysis perspective, the share price has stabilized above its key moving averages following a recent recovery. It trades notably above both the 50-day and 200-day averages, yet remains perceptibly below its annual peak. A Relative Strength Index (RSI) reading of 47 indicates a market that is neither overbought nor oversold, suggesting investors are carefully weighing the news flow rather than reacting excessively.

The critical question is whether Meta can convert its billions in AI and infrastructure spending into measurable growth within a reasonable timeframe. The Q4 2025 results, scheduled for release on January 28, 2026, are anticipated to provide the first comprehensive look at how these AI investments are impacting its advertising business and new product areas. This data will be pivotal in assessing the justification for the current valuation framework.