Meta’s recent stock surge of over 7% on Tuesday was driven by a Bloomberg report indicating the company’s plans to launch a venture to sell surplus AI computing power. This strategic move positions Meta in direct competition with established giants like AWS, Microsoft Azure, and Google Cloud. However, Mark Douglas, the CEO of connected-TV advertising platform MNTN, suggests that the emphasis on Meta’s developments may overlook larger economic challenges facing the AI infrastructure sector.
## Economic Viability of Data Centers
Douglas critiques the feasibility of building AI infrastructure in the United States, asserting that investors might be underestimating the long-term implications. “Data center capacity in the United States may not age well,” he warned, pointing out the high costs associated with such constructions. Many communities resist having data centers, leading to potential issues that investors should consider. He predicts that, within two years, these facilities may lose their attractiveness as viable investments.
## Competitive Landscape Shifting
In Douglas’s perspective, the competition is emerging from an unexpected source: massive, sovereign wealth-backed data centers in the Gulf region. He highlighted that the Kingdom of Saudi Arabia is developing extensive data center capabilities at significantly lower operational prices. “Why extract oil just to ship it when you can utilize it to power large-scale data centers?” he queried, underscoring the strong backing of the Saudi Public Investment Fund, one of the largest sovereign wealth funds globally. For investors, he cautions against backing firms focused on hyperscale data centers in the U.S. under these circumstances.
Douglas also mentioned that certain data centers in the Gulf are structured as legal extraterritorial zones, serving as “data embassies.” This allows multinational companies to comply with stringent data residency requirements without technically relocating their data abroad. “While everyone rushes to build data centers in less desirable communities in the U.S., this alternative remains largely unexamined,” he added.
## Meta’s Cloud Ambitions
Against this backdrop, Douglas expressed skepticism about Meta’s initiative to sell surplus AI computing capacity similarly to AWS and Google Cloud. “The move seems more about garnering attention than a genuine market fit,” he noted. Despite this notion, he acknowledged the strategy mirrors that of SpaceX and xAI, which have successfully leveraged their data center capabilities for a select group of clients but suggests that this model may not align with Meta’s traditional business structure.
Highlighting the limited number of companies that can effectively utilize raw data center space, he questioned whether Meta would attract sufficient demand. “Transitioning from billions of social media users to catering to a handful of data center clients seems misaligned,” he remarked, reinforcing his doubts about the strategic direction.
## Challenges of Diversification
Douglas emphasized that despite the stock market’s immediate positive response to such announcements, transitioning into new industries is fraught with difficulties. “There’s a misconception that just because a successful company like Meta decides to explore a new sector, it is guaranteed to thrive,” he explained. This assumption underestimates the complexities of building a new product line or service, especially when past ventures have not always succeeded.
He pointed out that the talent necessary for innovation often diverges from larger corporations due to their focus on entrepreneurial ventures. This disconnect may further complicate Meta’s ability to successfully transition into the AI infrastructure domain.
## Meta’s AI Strategy and Advertising Future
While skeptical about the cloud venture, Douglas remains optimistic regarding Meta’s overarching AI strategy, particularly its open-source models. He believes these advancements will enhance Meta’s advertising efforts, even amid investor concerns regarding ROI from AI investments. “Meta has the resources, ambition, and potential for growth in AI, particularly when utilizing vast datasets for model training,” he stated.
He also challenged the notion that improved ad-targeting technologies adversely affect consumers. “Efficiency in marketing can ultimately lead to lower prices for consumers,” he argued. In discussing whether Meta’s move toward cloud solutions aligns it more closely with Amazon or Microsoft, he believes the analogy mischaracterizes the situation. “Meta is set to become the world’s largest advertising entity, and any shift towards AI infrastructure should be viewed as a tactical detour rather than a foundational change.”