• The Buyside Journal
  • Posts
  • Macquarie Commits up to $5 Billion to Applied Digital’s AI Data Center Ambitions

Macquarie Commits up to $5 Billion to Applied Digital’s AI Data Center Ambitions

Major HPC Funding Accelerates Applied Digital’s Shift from Cryptomining to Advanced AI Infrastructure

The Buyside Journal is your ultimate private markets resource. We deliver sophisticated and meticulously curated financial news to keep you informed and ahead of the curve.

"“The data center is the new computing unit.”

-Jensen Huang

The fear of an imminent “AI bubble” has not stopped large institutional investors from pouring billions into high-performance computing (HPC) data centers in the United States. The latest example is Macquarie, one of Australia’s leading financial services firms, which has committed as much as $5 billion to accelerate Applied Digital’s expansion in HPC facilities that cater to advanced artificial intelligence (AI) and machine learning (ML) workloads.¹ This partnership speaks to the ongoing race among data center operators and infrastructure investors to secure the capacity required for next-generation AI systems.

A Major Vote of Confidence

Macquarie’s asset management division will first invest $900 million to fund Applied Digital’s Ellendale High Performance Computing (HPC) campus in North Dakota¹. This 400-megawatt (MW) facility is central to Applied Digital’s strategic pivot away from cryptomining toward HPC hosting and AI cloud services.² Beyond the immediate capital infusion, Macquarie has secured a right of first refusal on an additional $4.1 billion for Applied Digital’s future HPC data center projects over the next 30 months.²

In exchange, Macquarie gains a 15 percent stake in Applied Digital’s HPC business while the company retains the remaining 85 percent.¹ According to Wes Cummins, Chairman and CEO of Applied Digital, this arrangement bolsters the firm’s ability to develop up to 2 gigawatts of HPC capacity across multiple sites, including Ellendale.² By recouping $300 million of equity previously committed to the campus, Applied Digital has also improved its financial flexibility for subsequent expansion plans.²

From Cryptomining to High-Performance Computing

Applied Digital (formerly Applied Blockchain) originally built its name in cryptomining, hosting hardware for external Bitcoin and cryptocurrency clients.¹ However, like several other North American data center operators, it realized that AI and HPC workloads offer a more stable and lucrative business model than crypto alone. The Ellendale campus reflects this pivot, comprising multiple buildings designed specifically for GPU-intensive computing and advanced cooling requirements to handle the heat from high-density racks.¹

In addition to Ellendale, Applied Digital operates other sites in North Dakota.² Its HPC business includes hosting and cloud-based services that enable enterprise AI training and inference tasks.² The shift toward HPC puts the firm in direct competition with other GPU-focused operators that have also gained significant venture and private equity backing, especially amid the global boom in generative AI, natural language processing, and computer vision technologies.²

Why the HPC Market Attracts Large Capital

  1. Rising AI/ML Demand¹: Modern AI applications—particularly large language models and advanced ML algorithms—consume massive amounts of compute power. Traditional data centers often lack the specialized GPU capacity and power density to efficiently support these workloads. Operators equipped for HPC can command premium leasing rates and multiyear commitments.

  2. Attractive Returns: Market estimates suggest that investing in a GPU-focused data center can yield substantial revenues in just a few years.³ Some HPC operators have leveraged GPU inventories as collateral to secure debt financing, indicating both the high demand for GPU compute and investor confidence in the resale market for these specialized assets.¹

  3. Growing Regulatory Pressures on Traditional Banks¹: As banks face stricter capital requirements, the private credit and direct-lending sectors have become more active in financing large-scale infrastructure. Partnerships between banks, private equity, and data center operators are increasingly common, providing a pathway to deploy significant capital into AI-adjacent projects at scale.

  4. Diversification for Investors¹: For firms like Macquarie, HPC data centers serve as “pick-and-shovel” plays in the AI gold rush. Rather than betting on a specific AI application, they invest in the underlying infrastructure that every AI-driven enterprise will need.

Macquarie’s Broader Data Center Portfolio³

Macquarie is no stranger to the digital infrastructure sector. Beyond this deal with Applied Digital, the firm has invested in major data center platforms such as Aligned Data Centers and Virtus Data Centres. Globally, it continues to acquire and develop hyperscale sites, leveraging its deep capital reserves and experience to capitalize on the growing appetite for colocation and HPC services.

Recent moves by Blackstone, Brookfield, and other significant private equity players to acquire data center assets worth tens of billions underscore that HPC facilities are now among the most sought-after real estate in the digital economy. Macquarie’s $5 billion commitment to Applied Digital merely confirms that HPC data centers remain a centerpiece of the AI revolution.

Challenges and Considerations²,³

Despite the optimism, high-performance data centers also face potential pitfalls:

  • Overcapacity Risk: If multiple operators build at breakneck speed, an oversupply of GPU compute could temporarily drive down utilization rates and leasing prices.

  • Rapid Technological Evolution: Next-generation GPU architectures and AI accelerators appear in quick succession, prompting ongoing capital expenditures to remain competitive.

  • Energy Costs and Sustainability: HPC sites draw substantial power. Operators must navigate utility pricing, potential regulation around carbon footprints, and local incentives for renewable energy.

  • Supply Chain Bottlenecks: The AI hardware supply chain can be strained, especially for in-demand GPUs. Delays in obtaining equipment can slow down project completion.

Nonetheless, as AI continues to integrate into nearly every sector—from finance and healthcare to manufacturing and retail—the need for advanced compute is expected to remain robust.

Investment Outlook

For institutional investors, the Macquarie–Applied Digital partnership exemplifies the broader enthusiasm for HPC data centers. Amid soaring demand for AI training and inference capacity, specialized infrastructure assets are likely to sustain high growth and compelling returns, provided they can quickly bring facilities online and secure long-term client contracts.

By diversifying into HPC hosting, operators like Applied Digital are betting that AI/ML workloads will not only offset the volatility of cryptomining but also generate more predictable, higher-margin revenues. From Macquarie’s perspective, owning a significant piece of the physical backbone of AI—the data centers themselves—may prove more resilient than investing directly in AI software startups, which can be susceptible to rapid shifts in market sentiment.

In short, capital deployment into HPC capacity reflects the strategic view that powering AI is akin to selling the picks and shovels during a gold rush. While concerns about an AI bubble remain in some corners of the market, many major investors see no sign of demand slowing for GPU-packed data centers, believing the real opportunity has only begun to emerge.

Conclusion

By infusing up to $5 billion into Applied Digital’s HPC data center projects, Macquarie reinforces its commitment to driving the infrastructure that underpins AI growth in the United States. This colossal deal highlights the accelerating trend among institutional investors to anchor themselves in the AI ecosystem, focusing on the real estate and hardware side of the digital economy.

For Applied Digital, the alliance marks a significant leap forward. With Macquarie’s deep pockets and expertise in infrastructure, the company can swiftly expand into new markets, capture more enterprise AI workloads, and mature as a key player in the American HPC sector. As the race for GPU compute intensifies, large-scale, liquid-cooled, and power-dense data centers will likely remain at the forefront of investment strategies—solidifying the role of HPC infrastructure as the undisputed backbone of AI innovation.

Follow Us:

1. The Register, “Fears of an AI bubble have yet to scare off venture capitalists and private equity firms,” 2025.
2. DataCenterDynamics, “Macquarie invests up to $5 billion in Applied Digital’s AI data center business,” 2025.
3. PitchBook, “Macquarie boosts its bet on data centers amid AI boom,” 2025.

The content herein is solely for informational purposes and should not be viewed as investment or any other advice or a current or past recommendation, or an offer to sell or the solicitation to buy securities or adopt any investment strategy. Certain of this material has been generated by an artificial intelligence language model, ChatGPT, which has been prompted to provide topical finance-related articles. The articles herein may not reflect the most current news, events, or developments. While we strive for accuracy, there may be limitations, inaccuracies, or biases present, and The Buyside Journal (including, for the avoidance of doubt, its affiliates) assumes no liability for the content herein and does not guarantee the accuracy, adequacy or completeness of such information (and does not undertake any duty to correct or update such information). Readers are encouraged to independently verify the information herein and consult with professionals for specific advice or information. Predictions, opinions, and other information contained herein are subject to change continually and without notice of any kind and to the extent accurate initially may no longer be true after the date indicated. Forward-looking statements are subject to numerous assumptions, risks and uncertainties, which change over time. Actual results could differ materially from those anticipated in forward-looking statements.