Big Interview

Snowhorn’s gigawatt data centre community

Josh Snowhorn Quantum Loophole.jpg

Former CyrusOne executive Josh Snowhorn is planning a data centre campus in northern Virginia. Power and fibre will be laid on: just build your own data centre on the site or take space in one of Quantum Loophole’s own units. By Alan Burkitt-Gray

 Former Terremark and CyrusOne executive Josh Snowhorn is planning to build a vast campus in northern Virginia to house hyperscalers and multi-tenant data centres. Snowhorn, who announced in February that he had raised an initial $13 million in seed funding for the project, gives more details in this exclusive interview.

He is calling the new company Quantum Loophole, and it will aim at creating “highly sustainable, environmentally sound data centre developments at scale”. Quantum Loophole has identified a 2,000 acre site that is already zoned as industrial land.

“We’re not trying to do a crazy assemblage of agricultural land and then trying to take that away from farmers in the community,” says Snowhorn.

The zoning will allow for “20 million square feet” of data centre space on the property, he estimates.

“Our land is almost equal to all data centre land that’s been built in Loudon County today. That’s the kind of scale of what we’re doing.” Loudon County is almost at the very tip of northern Virginia, north-west of Washington DC and just outside Ashburn.

“We’re under a hard contract on the land… We have escrowed funds and we have a contract with the seller,” says Snowhorn. “We are in the very last final throes of our funding process.”

Quantum Loophole will be providing a substation supplying 1,000MW of power, he says, and will be building a 100,000-strand fibre ring to connect the users. “We are extremely close to Ashburn and to the centre of the internet,” he added. The fibre network will connect the users to Ashburn “and we will be allowing fibre providers to connect to us”.

Snowhorn says: “We are in a location that doesn’t take tax on infrastructure.”

The site will be ready to start development in July, he tells Capacity. “There’s about a 12-month lead time on transmission hardware and transformers,” so it will be the end of 2022 before Quantum Loophole comes into action.


Every hyperscaler

There won’t be just one building being built, a number of sites will be under construction across the site at the same time. Snowhorn has talked to “every hyperscaler”, he says.

The name Quantum Loophole implies “freedom of choice”, as the company tries “to shoehorn science into our vision of what is good for our concrete box”, he says. “A lot of different boxes with a lot of variability, even though there are standards in the industry.”

It’s “a campus that is so large, and at a scale that’s beyond anything anybody’s ever built”, Snowhorn says, and this “allows us to sell land to customers if they want to build their own sites”, or sell land “to a multi-tenant data-centre provider, and they can build their own boxes and supply their own services”.


Master planning

It will be a “master-planned community”, he says, with “a common base-level security and conduits and things like that to tie it all up, but it really lets them be free to do what they want to do.”

As well as the multi-tenant buildings, Quantum Loophole will provide energy services. “We’re building our very first substation – it is going to be over 1,000MW of capacity.”

That’s “a key differentiator”, says Snowhorn. For most data centres, “you’re limited by the scope of the distribution substation”, he notes. “It has some inherent risks of reliability, because you’ve sort of added a layer back to the grid. And then that requires you to put in big honking generators and other things that are of course expensive.” And generators pollute the environment, he adds, and are “no longer acceptable, particularly at the scale we’re doing it”.

A bank of generators to back up a 1,000MW feed is like 500 diesel railway engines on the campus. “That’s not going to happen. We’d have Greenpeace surrounding us on day one.”

The answer – because back-up power is still needed – is utility-scale batteries. “We’ve filed patents on energy orchestration software that allows us to orchestrate between the grid, and our clients load and utilise the batteries to store either cheap energy or green electrons as much as those can be acquired,” Snowhorn says.

He will invite utilities to “start to scale around us”, he adds. “Of course, they are going to do it when the time is right, and they’ll be building solar and other green technologies around us to support the load.”

And he points to a curious fact. We’re using less electricity now than, say, a decade ago, thanks to LED and other low-energy lighting. “Every major power utility in the world is chasing load now,” says Snowhorn.

“The grids overall have declining loads – so they all look at data centres as the steady-state users of electricity.” And data centres are “magically efficient, and that’ll get better and better”, he adds. His vision for this data-centre park is “something like Google’s Mountain View campus where you have wonderful walkways and greenery”, and a community “instead of a bunch of big concrete ugly boxes”.

With the funding just about in place and the site identified – though he won’t say precisely where it is – the next stages are putting together “an amazing team and amazing product set”, with “the financial institutions fighting to win the deal”, he says.

When will we hear more? “We would anticipate further announcements around the July timeframe, related to the exact location, the exact size of the land.”

Why has no one tried this before? “A lot of people have attempted to build a big mega-campus,” but they’ve missed the main points, Snowhorn says. “How do we achieve benefits to our clients, outside of energy and sustainability and everything else?” Interconnection, he says.


Fibre ring

“We have a fibre ring that has capacity for 100,000 strands in each leg. It’s a complete ring.” That will be tied to the Ashburn infrastructure, connecting the “centre of the internet”, as he calls it, “back to the campus, building to the edge of the community”, where “all the fibre providers can connect to us”.

Rivals have “tried to take agricultural land and convert it”, he says. “That’s the big ugly thing for the community.” And “they didn’t look at the connectivity properly”, he adds.

“One of the biggest things has to do with taxes. We happen to be in an incredibly beneficial location. The clients coming out don’t actually pay taxes on that infrastructure [so] savings would be tremendous.”

But that’s just the start, says Snowhorn. “The real icing on the cake is taking away the chasing every year. Instead you can actually get a 15-year planning cycle.” You can stay in one location “that meets all of your power load demands, all of your connectivity demands, [and] has incredible cost and tax benefits”. It means you “can now change your planning cycle from one year to 10 to 15 years”.


Control your destiny

Expansions don’t mean “having to shoehorn yourself in” to an existing data centre. “Instead, you actually get to control your destiny.” If a hyperscaler “wants to come along and acquire 200 acres… and build their own destiny, they have that ability to do that, and that can really change their planning viewpoints.”

Is this a one-off for northern Virginia or is the model replicable in other parts of the US or other parts of the world? “Certainly not in the [London] Docklands,” Snowhorn smiles.

“We do have a North American focus for now. We do not plan to destroy a bunch of beautiful farmland in rural England, I promise you that. But, certainly, in some core North American markets where there’s rapidly rising land costs and a shortage of supply.”

He adds: “Our core mandate is going to be anything from 1,000 acres and above,” with “the right energy” and the right market. “They are not easy to find, I promise you.”

Why did Snowhorn have to set this up via a new company, not persuade, say, CyrusOne to do it? One of the problems is “they’re driven by their mandates as a public company to have success every quarter, and every year, and to show bookings”, and to be a “revenue engine”, he says. “It removes some of the ability to innovate.” 

Innovation is a “big risk”. With his new project, “let’s just say we get to, you know, three gigawatts on the single campus, just imagine that.” If it costs $6 million a megawatt to build, “well, holy crap, that’s $18 billion”. That’s why existing operators have refined projects down to modularity and driven down costs as much as they can, he says. “But it’s very much a rubber-stamp approach in the industry.”

With this project, it is taking “a leapfrog innovation”, says Snowhorn, and that “is a difficult thing, and takes a lot of work and that’s what we’re achieving now – really taking a leapfrog”.

But, he adds: “I will be clear, we are not doing this speculatively. The industry and the customers who will be taking capacity at large scale asked us to go and solve something like this, and we were the ones who were willing to do it right.”