The artificial intelligence (AI) infrastructure market is transforming significantly and a recent development in this field is from an unlikely alliance of two industry giants. Meta Platforms and world’s largest asset manager BlackRock announced their partnership in $14 billion investment in a massive data center campus in El Paso, Texas. The partnership represents a stark change in the way that tech firms are funding their grand plan for AI, given that capital requirements have reached such proportions that even the tech giants are finding they need external capital partners to help with the cost.
As soon as I started hearing about this partnership, I couldn’t not consider how the economics of AI infrastructure has changed over the last few years. Once it was largely financed by internal cash flows and conventional debt markets, it is now a highly capital-intensive business that sees companies enter into strategic deals with real estate and infrastructure partners – institutional investors. This new reality will be demonstrated by the Meta-BlackRock venture, which is expected to require a significant amount of computing power to build on the gigawatt scale that any single company could not afford to finance with its own balance sheet.
The El Paso campus is already in construction, and will offer one gigawatt of compute capacity as soon as it goes into operation in 2028. In comparison, the amount of computing power needed is necessary for training and running the intricate AI models that Meta has been developing, such as the underlying technology behind the Meta AI app, their image-to-video advertising solutions and the smart glasses products. This is only a single component of Meta’s massive data center project, which the company intends to spend a whopping $600 billion on building out by 2028.
The financial arrangements of this partnership illustrate the advanced thinking within the realm of AI infrastructure financing. Meta will hold a 20 per cent stake in the venture and have the final say, with BlackRock-managed funds holding 80 per cent. The debt aspect of this arrangement is especially intriguing because $12.5 billion of BlackRock’s investment will be paid for with debt, using the stability and creditworthiness of both partners. Moreover, the venture will also make a distribution of $1 billion to Meta, which seems to be meant to even out the share ownership and reward the company for its investment.
Meta doesn’t have nothing to lose in this partnership. The company will provide land and construction assets currently in progress, valued at about $2.3 billion. BlackRock, on the other hand, is giving a significant cash injection, in the region of $4.9 billion. The combination of assets and cash puts the project on a good footing, and it is equally beneficial to both in terms of risk and reward.
This project is a huge undertaking. Data centers are the physical manifestation of the digital economy, and gigawatt scale centers such as that in El Paso are the new benchmark for AI computing. To help illustrate, Meta is constructing several of these massive facilities in the U.S., such as the plant in rural Louisiana, which the company has said will reach five gigawatts in compute power; a figure potentially worth more than $50 billion to invest. The numbers are indicative of how much computational power is required for these AI systems, which are incredibly powerful and resource-intensive.
The advantage for the El Paso location is near the Texas-New Mexico border. The area offers reliable access to power, advantageous tax provisions, and access to existing fiber optic networks which can handle the vast amounts of data AI operations require. In addition, land is available and the permitting process is relatively streamlined in Texas, which is an attractive place for such large scale development.
The deal between Meta and BlackRock is an interesting development in the financing of technology infrastructure. Traditionally, tech firms have funded expansion by using their own cash flow or conventional business loans. But, as AI investment becomes unprecedented large, companies are getting more creative in financing plans. BlackRock’s involvement will not only provide funding but also bring expertise in managing the vast and complex scale of data centre investments, a match for Meta’s lofty plans.
This is not an insignificant move as investors are closely watching companies’ AI investments. Shares of Meta have fallen about 10 percent so far this year as investors examine the costs of the company’s AI expansion. The cost of implementing the infrastructure is large, but the benefit of the future of AI technologies is substantial. This collaboration allows Meta to modulate that equation by spreading its investment risk over BlackRock, but with the firm still holding a significant stake in the company.
Along with this, the data center campus will contribute to Meta’s mission of building personal superintelligence, something the company has been talking about as the next step in AI evolution. The infrastructure being built in El Paso will help to launch new streams of revenue from a range of AI enabled products and services that could compensate for the high up-front investment costs over time. Its launch in 2028 will come at a time when numerous experts are anticipating that AI will be further embedded in daily life and business practices.
Projects of this magnitude pose opportunities and challenges to the El Paso region from a community point of view. Building and running this huge plant will bring jobs, boost local economic growth and even draw more technology investments to this area. As the project progresses, however, there will be questions regarding the power consumption, water usage, and environmental impact. The scale of these data centers raises a critical issue of sustainability and resource allocation, given their massive power consumption and extensive cooling systems.The size of these data centers comes with its own set of challenges, particularly in terms of sustainability and resource allocation, as they consume vast quantities of electricity and rely on extensive cooling systems.
