Issue #10: Nscale's Vertically Integrated Profile

The compute market is oversupplied, and margins are under pressure.

The rapid build-out of GPU infrastructure has left providers battling for profitability. Hyperscalers survive by layering value-added services, but NeoClouds face a tougher road. Without comprehensive solutions or scale advantages, many are scrambling to compete on price alone.

What’s the solution?

Vertically integrate.

By owning every layer of the stack—from renewable energy-powered data centres to cutting-edge GPU services—vertically integrated models are designed to thrive in a margin-constrained market. They don’t just weather the storm; they prove there’s another way to scale.

One company emerged from stealth in 2024 doing exactly this, and with a $155m Series A, 1.3GW of greenfield pipeline, and a planned $2.5b UK investment, the model clearly has legs.

This is Nscale, and here’s how they’re doing it.

The GPU Audio Companion Issue #10

Want the GPU breakdown without the reading? The Audio Companion does it for you—but only if you’re subscribed. Fix that here.

/

Company Background

Nscale launched in 2024, but its roots lie in Arkon Energy, a renewable energy-focused data centre operator that began as a Bitcoin mining business.

Founded in 2021, Arkon built a reputation for leveraging renewable power to fuel high-performance computing. By 2023, Arkon had expanded its operations to include 177MW of capacity across facilities in Norway, Australia, and the US, setting the stage for Nscale’s debut.

The decision to spin out Nscale was strategic, aimed at addressing the growing demand for purpose-built AI infrastructure.

Leveraging Arkon’s expertise, Nscale emerged with a vertically integrated model that combines renewable energy, cutting-edge GPU clusters, and a proprietary AI stack.

The Glomfjord Facility: A Flagship for Sustainable AI

Nscale’s cornerstone is its 60MW facility in Glomfjord, Norway.

Powered entirely by hydroelectric energy and cooled using Arctic air and adiabatic liquid systems, the site embodies Nscale’s sustainability-first philosophy. And it’s also one of the first facilities globally to deploy the AMD’s MI300X.

Beyond performance, the facility integrates with the local community, repurposing waste heat for infrastructure projects like heating public swimming pools, serving as a blueprint for sustainable AI infrastructure.

Expanding Horizon

Since its launch, Nscale has accelerated its global growth with a $155M Series A funding round led by Sandton Capital Partners.

This capital is fueling a 1.3GW pipeline of greenfield data centres, with 120MW set for development in 2025 across Europe and North America. The company has also bolstered its leadership team, hiring top executives like Karl Havard (CCO) and Ron Huisman (CFO) to steer its ambitious expansion.

By owning every layer of its stack - data centres, GPU hardware, and software - Nscale delivers tailored AI infrastructure faster, more efficiently, and at a lower cost than traditional providers. Their vertically integrated approach positions them to meet the soaring demand for AI infrastructure while setting new standards for sustainability and performance.

Executive Team

Nscale’s leadership team spans AI, cloud, and sustainable infrastructure:

• Josh Payne (Founder & CEO): Driving the company’s vision with experience in renewable energy and M&A.

• Sam Palmisano (Board Advisor): Former IBM CEO, offering strategic insight for global expansion.

• Karl Havard (CCO): Ex-AWS and Google, leading commercial strategy and customer success.

• Alex Sharp (COO): Data center veteran, previously at Iron Mountain, overseeing operational scale.

• David Power (CTO): HPC and AI expert, co-founder of vScaler, driving tech innovation.

• Ron Huisman (CFO): Financial strategist with a background at AtlasEdge Data Centres.

• Daniel Bathurst (CPO): Leading product development across AI and HPC solutions.

The Edge

Nscale’s differentiation lies in three core pillars:

1. Vertical Integration

Nscale owns every layer of the stack, giving them unparalleled control over cost, performance, and sustainability:

• Custom-Built Data Centres: Explicitly designed for GPU clusters and AI workloads.

• Optimised GPU Deployments: Tailored solutions for training, fine-tuning, and inference.

• Cost Advantage: Customers benefit from direct access to infrastructure without hyperscaler markups.

2. AI-Specific Solutions

Their infrastructure is purpose-built for the generative AI lifecycle:

• Training Clusters: Managed Kubernetes clusters and SLURM schedulers to optimise large-scale model training.

• Inference Services: Auto-scaling GPU nodes deliver high-performance real-time AI inference.

• AI Marketplace: Pre-configured tools like PyTorch and TensorFlow streamline deployment.

• Fine-Tuning Environments: Dedicated resources for customising large language models.

3. Sustainability at Scale

Sustainability isn’t just a buzzword for Nscale—it’s a design principle:

• Renewable Energy: Their data centres run on hydropower, reducing carbon footprints.

• Closed-Loop Cooling: Liquid cooling systems minimise environmental impact.

• Global Expansion: 1.3 GW of greenfield data centre capacity is in development across Europe and North America.

Recent Moves

Nscale’s recent milestones reflect its rapid momentum and ambitious vision:

• $2.5 Billion UK Investment: Developing modular and fixed AI data centres, including a 50MW flagship facility in Loughton (scalable to 90MW).

• $155M Series A Funding: Backed by Sandton Capital Partners to accelerate global expansion.

• Singtel Partnership: Unlocking GPU capacity across Europe and Southeast Asia, leveraging a unique GPUaaS model.

• Public Cloud Launch: A flexible AI platform for training and inference, set to go live in Q1 2025.

• Top500 Supercomputer Recognition: The Glomfjord Svartisen Cluster ranks among the world’s most powerful.

• Kontena Acquisition: Adding expertise in modular data centre design to expand deployment options.

• Open Innovation AI Collaboration: Deploying 30,000 GPUs across the MENA region.

What’s Next?

Nscale is entering 2025 with significant momentum and a roadmap that’s as ambitious as it is strategic.

The Loughton facility, scheduled to go live in 2026, will set new benchmarks for AI infrastructure in the UK. Hosting up to 45,000 NVIDIA GB200 GPUs, the site will drive innovation for enterprises and startups alike while creating 750 new jobs over three years. Modular data centers slated for construction in Q3 2025 will further extend Nscale’s footprint.

Globally, Nscale’s 1.3GW greenfield pipeline positions it as a leader in scalable, sustainable AI infrastructure. By leveraging partnerships with AMD, NVIDIA, and regional players like Singtel, Nscale is ensuring its infrastructure can meet the demands of the fast-moving AI market.

But the real test lies in the Global South.

By extending their vertically integrated model to regions underserved by traditional hyperscalers, Nscale could offer a scalable, sustainable alternative for nations navigating the challenges of digital transformation.

The big question? Can they replicate their model at scale while staying true to their principles of sustainability and cost efficiency?

Nscale isn’t just building AI infrastructure.

They’re building a blueprint for the future of compute.

Keep The GPU Sharp and Independent

Good analysis isn’t free. And bad analysis? That’s usually paid for.

I want The GPU to stay sharp, independent, and free from corporate fluff. That means digging deeper, asking harder questions, and breaking down the world of GPU compute without a filter.

If you’ve found value in The GPU, consider upgrading to a paid subscription or supporting it below:

☕ Buy Me A Coffee → https://buymeacoffee.com/bbaldieri

It helps keep this newsletter unfiltered and worth reading.