ALPS Pioneers Sustainable AI Data Centers with 'Energy-to-Compute' Strategy and Asset Reuse

ALPS, an Italian scale-up specializing in data center design and operation, has appointed Brian Turnbow as Technical Director for AI, signaling a strategic shift towards more sustainable AI and High-Performance Computing (HPC) infrastructure. The company is championing an "energy-to-compute" approach, which fundamentally rethinks data center placement and construction. Instead of building new facilities in power-constrained areas, ALPS focuses on locating data centers where energy, particularly renewable sources, is already abundant. This includes redeveloping disused industrial sites, leveraging existing grid connections and buildings to minimize new construction and land use. This development is significant because the burgeoning demand for AI and HPC is placing immense pressure on existing energy grids and natural resources. Data centers are projected to consume around 950 TWh of electricity by 2030, roughly 3% of global electricity demand, with AI-focused data centers seeing a 50% rise in consumption in 2025 alone. The traditional model of data center expansion often involves significant new construction and strain on local power infrastructure. ALPS's strategy directly confronts these challenges by integrating data centers with energy systems and prioritizing asset reuse. This approach offers a tangible solution to the paradox of increasing compute demand and the urgent need for environmental sustainability in the digital sector. The move by ALPS aligns with a broader, well-established trend in green cloud and DevOps, where the industry is increasingly scrutinizing the environmental footprint of digital infrastructure. Hyperscalers like Google and Microsoft are already investing heavily in renewable energy and optimizing data center efficiency through AI-driven cooling and other innovations. However, the sheer scale of AI growth means that even with these efforts, emissions from the digital sector continue to rise. The focus is shifting beyond just purchasing renewable energy credits to more fundamental changes in infrastructure design and deployment, including considerations of embodied carbon in construction and water usage. The IEA's projections highlight that while efficiency is improving, it's not keeping pace with demand, making innovative siting and resource utilization critical. In practice, this means practitioners should begin to evaluate their AI and HPC infrastructure needs not just on computational power and cost, but also on the environmental impact of their chosen solutions. Organizations should inquire about their cloud providers' energy sourcing strategies, particularly regarding hourly matched clean power and the reuse of existing assets. For those building their own infrastructure, ALPS's model suggests exploring locations with existing renewable energy capacity and considering the redevelopment of brownfield sites to reduce embodied carbon. This approach can lead to more resilient and sustainable compute capacity, mitigating risks associated with grid instability and increasing regulatory pressure on environmental reporting. The long-term implications point towards a future where the physical location and design of data centers are intrinsically linked to their environmental performance, demanding a more holistic and sustainable approach to digital infrastructure planning.