Next-generation data centres,
powered by sustainable energy.
WE BUILD CLEAN, GREEN
SUPERCOMPUTING
MACHINES
At Scale42* we build next-generation data centres - clean, green supercomputing machines - powered by 100% sustainable energy.
LOW-COST CLEAN ENERGY
+
ACCELERATED COMPUTING
=
NEXT-GENERATION DATA CENTRES
We scour the globe for low-cost sources of clean energy - from stranded hydropower to solar, biogas or wind located away from centres of demand or awaiting grid connection - wherever there's a sustainable source of electricity, we can turn it into high performance compute.
We connect clean power to the latest accelerated computing hardware. We've specialised in ASIC-accelerated compute since at least 2018, buying servers with chips manufactured by semiconductor fabrication behemoth TSMC, by the thousand.
Highly efficient, compute-intense data centres - sustainably powered, liquid-cooled and densely packed with ASIC / GPU / TPU accelerated servers. Our data centres are optimised to meet the demands of today's most highly specialised and demanding computational tasks - such as the training of generative Artificial Intelligence (AI) models.
DISTRIBUTED
0.5 - 10MW
We build small, distributed data centres - containerised, modular solutions deployable via the global logistics network. They're the perfect solution where a data centre footprint might need to be scaled up or down to match changing requirements.
CENTRALISED
10 - 100MW+
We also build large, centralised data centre facilities like this 60MW hyperscale project in Norway - 100% powered by clean, renewable hydroelectricity.
With static data centres we're really only limited in scale by the scope of available power and can scale to 100s of megawatts (MWs) or gigawatts (GWs).
Crucially we can also setup in remote locations, siting compute as close to power generation as is practicable - like here in the Arctic circle - where we can take advantage of 'stranded' hydropower.
DATA CENTRES: THE NEXT GENERATION
LEGACY
DATA CENTRES
HIGH OPEX & CAPEX
Legacy data centres - typically built to serve consumer facing software-as-a-service applications - lock-in high CAPEX and OPEX through a focus on:
- Low latency
- High uptime
- CPU-centric architectures
​
This legacy infrastructure:
- Is typically located close to urban centres - locking in high land costs
- By association is often encumbered with high energy costs, in part due to reliance on power generated by fossil fuels
- Is built with high levels of redundancy - in power and communications - with associated low levels of utilisation and high cost
​
All of the above contributes to data centres having a large physical footprint, further locking in both high CAPEX and OPEX.
NEXT-GENERATION DATA CENTRES
LOWER OPEX & CAPEX
Next-generation data centres by contrast - eg. built to train complex AI models - can benefit from significantly lower CAPEX & OPEX, by focusing on:
​
- Raw compute power
- High internal network speed for increased parallelisation
- GPU / TPU / ASIC-centric architectures
​
They are typically:
- Powered by low-cost, clean energy
- Co-located alongside the power generation eg. in remote locations
- Equipped with fewer redundant systems, each with higher utilisation
The high density of equipment leads to a smaller footprint, which together with out of city locations can enable dramatically lower CAPEX and OPEX vs legacy data centres
THE FOUNDERS
THE TEAM
CFO
Patrick Murray
-
15 years focused on the technology, FinTech and Life Sciences sectors.
-
Leader in closing and agreeing significant funding rounds, shareholder agreements with PE investors in FinTech startups.
-
Established & developed finance and compliance teams, robust accounting processes & controls, and financial reporting.