The Neocloud Sustainability Paradox: Can New AI Operators Deliver Where Big Tech Failed?
Neocloud companies are emerging as a counterweight to traditional hyperscalers, promising to prioritize sustainability from the ground up rather than making lofty climate pledges they struggle to keep. As AI infrastructure demand explodes globally, a new generation of cloud providers is taking a different approach to the power and cooling challenges that have derailed earlier climate commitments from Amazon, Microsoft, and Google.
Why Did Big Tech's Climate Goals Fall Apart?
Back in 2020, the major cloud providers made sweeping climate commitments. Amazon pledged net-zero carbon emissions by 2040, Microsoft promised to be carbon negative by 2030, and Google vowed to operate on carbon-free energy around the clock by 2030. But when the AI boom arrived in 2022, those goals collided with an insatiable hunger for computing power. Data center developers and cloud companies began competing fiercely for every available power resource, and sustainability took a backseat to growth.
Today, with the AI data center buildout in full swing and a US government more focused on preserving the fossil fuel industry than protecting the planet, there is significant doubt about whether these climate goals will be achieved. Amazon shareholders have even called for more transparency about how the company is meeting its sustainability targets while expanding its capacity by 3.9 gigawatts in 2025 alone, though the company's board has urged shareholders to reject this line of questioning.
What Are Neoclouds, and Why Do They Matter?
The launch of OpenAI's ChatGPT catalyzed a wave of new cloud companies focused specifically on offering AI and GPU-based compute capacity. These "neoclouds" include companies like CoreWeave, Crusoe, Lambda, Nebius, and Nscale, and they are growing at a rapid pace. According to recent research from Synergy, neocloud revenue is set to exceed $25 billion for the year, with these companies taking up and bringing online large amounts of capacity and forecasting growth in the gigawatts.
What distinguishes neoclouds from traditional hyperscalers is their willingness to challenge the "announce first, achieve later" approach to climate goals. Nidhi Chappell, president of AI infrastructure at Nscale and formerly Microsoft's head of Azure AI and HPC infrastructure, explained the company's philosophy directly: "We will achieve the goals before we publish the goals. We aren't a big corporation, we don't go after lofty goals and then try to push them out." Instead, Nscale focuses on having "sustainability in mind" from the start, though Chappell acknowledged that as the company scales, it will need to balance sustainable options with practical growth.
How Are Neoclouds Leveraging Geography for Sustainability?
One of the most effective strategies neoclouds have adopted is locating data centers in the Nordic region, where renewable energy is abundant and cooling is naturally efficient. The Nordic countries, including Norway, Sweden, Denmark, Finland, and Iceland, generate more than 90 percent of their power from renewable and low-carbon sources. This geographic advantage has become a major draw for neocloud operators.
Nscale operates multiple data centers in the Nordic region, with its first AI cloud offering launching in May 2024 from a facility in Glomfjord, Norway, powered by hydroelectric energy and using natural cooling. CoreWeave operates a site in Sweden, NexGen Cloud has sites in Norway and Sweden, and Nebius has multiple Nordic-based sites. According to CBRE research from November 2025, 57 percent of European neocloud capacity is based in the Nordics, driven both by the availability of cheaper power and difficulties building in top-tier markets like the United States.
Beyond just renewable power, Nordic locations enable innovative sustainability initiatives that traditional data centers struggle to replicate. Nebius, which was spun out of Russian tech giant Yandex, operates a data center in Mäntsälä, Finland, that reused more than 50 gigawatt-hours of heat in 2024, enough to heat 2,500 homes during that period. The facility covered 72 percent of the heating demand for the local district heating network that winter, and in return, the district saw electricity bills decrease by 10 percent in 2025 due to the heat from Nebius' data center.
"These types of circularity solutions are embedded in our design, and they impact the PUE of sites and how that number behaves significantly," explained Daria Mukhortova, head of sustainability at Nebius.
Daria Mukhortova, Head of Sustainability at Nebius
Nebius' cooling systems are all designed with the potential for heat exchange, though deployment depends on local demand. In Finland, district heating infrastructure exists to capture that waste heat, but in parts of the United States, such systems are rare. Mukhortova noted that heat reuse conversations happen between data center teams and local partners, but only when there is actual demand for the recovered heat.
What Challenges Do Neoclouds Face in the United States?
While neoclouds based in Europe can access sustainable power relatively easily, the situation in the United States is far more complicated. The hunger for AI is a global phenomenon, and in the US, demand is described as "gnawing and insatiable." TensorWave, a neocloud focused exclusively on AMD hardware, operates sites in Pennsylvania, Virginia, and Arizona, while developing a location in Chile. The Pennsylvania project, being developed by TecFusions, is planned to be mostly powered by natural gas.
Darrick Horton, CEO of TensorWave, articulated the core challenge facing all neoclouds in the US market: "The challenge in this space is access to power primarily. We think, for the foreseeable future, there will be a severe shortage of large-scale access to power." Despite this constraint, Horton noted that the power shortage is actually driving adoption of sustainable energy technologies, with increased discussion of on-site generation, microgrids, solar power, and nuclear energy.
Darrick Horton, CEO of TensorWave
However, Horton also acknowledged a sobering reality: "There is no such thing as seeking out a site with certain specifications anymore; this is very much a seller's market." Neoclouds, despite their sustainability commitments, often cannot afford to be selective about power sources when capacity is scarce.
How to Evaluate Neocloud Sustainability Claims?
- Track Record Over Promises: Look for companies that publish annual sustainability reports with operational data, like Nebius does, rather than relying solely on future pledges. Actual performance metrics are more reliable than aspirational goals.
- Geographic Strategy: Assess whether a neocloud is deliberately locating in regions with abundant renewable energy, such as the Nordic countries, or if it is simply building wherever power is available, regardless of source.
- Heat Reuse and Circular Design: Evaluate whether cooling systems are designed from the ground up to enable heat recovery and local community benefits, which reduces overall power consumption and creates economic incentives for sustainability.
- Transparency on Trade-offs: Be skeptical of companies that claim perfect sustainability while scaling rapidly. Honest operators acknowledge the balance between sustainable options and practical growth constraints.
What Does This Mean for the Future of AI Infrastructure?
The neocloud market is far more dynamic and complex than the traditional cloud provider landscape, but as these newcomers begin providing a significant chunk of global data center capacity, the question of how this new era of cloud computing will operate sustainably remains partially unanswered. The contrast between neoclouds and traditional hyperscalers is striking: where Amazon, Microsoft, and Google made sweeping climate pledges in 2020 and then struggled to meet them as AI demand exploded, neoclouds are taking a more cautious, achievable approach.
Cory Hawkvelt, chief product and technology officer at NexGen Cloud, captured the pragmatic philosophy driving neocloud sustainability: "One, it's generally the right thing to do, and we have always believed in that, but two, it's also usually the cheapest outcome, so it's sort of a happy coincidence." This alignment of ethics and economics may prove more durable than the idealistic pledges of the previous decade.
Meanwhile, Meta's Hyperion AI campus in Louisiana is expanding to 5 gigawatts with a $50 billion investment, transforming from a large data center into a regional power anchor that will reshape grid planning and transmission expansion across the region. This scale of infrastructure development underscores the urgency of solving the sustainability challenge, whether through neocloud innovation or traditional hyperscaler adaptation.