The Search for Power in the Big City

Our panels are looking a little... full. Basically we've maxed out the amount of juice we can feed mining equipment and as such, the time to expand to a secondary location has arrived. It got to the point where I had to max the panel at my house and three of our employees also did the same. R&D is fun isn't it?

When setting up a mining operation its not now nice the building is that matters. It's how much power you can get and at what price you can get the space. Commercial power contracts are pretty straightforward. The more you buy, the cheaper it gets. To a point. When I started the hunt for a new space we looked all over the place. Brokers were pretty consistent in showing us the A class spaces off the get go, but they rarely had anything even close to what we wanted in terms of power. Sure, the space was nice and the offices were cozy. You got a good view and easy access to main highways and ample parking. But the thing about commercial space is, you don't need all the bells and whistles. You don't even need a lot of square footage.

What you need is power and lots of it. At an average of about 8.5 amps @110 per GPU mining rig (plus 20% overhead) you're going to fill the traditional 200a panels in a hurry. In fact, come to think of it 200a of 220v was actually pretty rare and even when we did find it that load was shared with HVAC, lighting and other crap.

Ok, so I'll have to come out right and admit we're not even in that big of a City. Canada as a whole has less people in it than New York City. Where we're located is an industrial hub that has seen a serious downturn over the past few years. So where others are suffering, we are find opportunity. I started asking the brokers to show us industrial sites with smaller square footage. Most of the stuff with high amperage service were large 40,000 square foot manufacturing plants that were sitting empty. So we started looking for smaller buildings that used to house welding shops and the like.

That's where we struck gold. When the industrial sector starts to take a nose dive, welding shops seem to be some of the first guys to bite the dust. It's not that I'm happy those guys are out of business. But it does provide a unique opportunity to capitalize on a Goldilocks building.

Welding shops are typically quite dirty, small and have a ton of power on the panel. It has been quite typical to find 600-2500a service in these buildings, though the higher end is a bit harder to find with the smaller footprints we want. And because these shops are not too visually appealing, you can walk away with a pretty competitive lease and you'll be hard pressed for them not to throw in the professional cleaning.

Lets be honest, mining rigs don't need ocean front views. We want rack density, power and somewhere to throw the obnoxious heat these monsters generate.

So that's how our hunt has been going. We're still looking at a few properties over the next few weeks to find the right fit of power, space and the ability to route heat outside the building. HVAC isn't a concern for us as it just doubles your power costs.

Every watt of energy burned is a watt of HVAC you require to cool (loosely speaking). We have some friends that run a pretty sizable operation in the next province over, and they're currently getting dinged $0.14kWh PLUS another $0.14kWh for the HVAC system to cool it. Whats interesting about this is you might think they don't make much money at $0.28, but the fact is they do. Lots of it. Obnoxious amounts of it. But if they could vent the heat outside efficiently without having to use air conditioning, the revenue would be that much greater.

So that's leading us to the next challenge in all of this. How to vent the heat. We've got a prototype being constructed right now of a way to direct heat out of a building efficiently through targeting the actual physical gpu's themselves. I've seen a couple people get close to what we're building, but nothing that has the rack density of what we're trying. I'm not an HVAC expert so we'll see if this works out. But if it does, we've got something that's going to help others in the $0.28 power cost range significantly reduce their overhead.

More to come on the journey. I hope this helps any of you in the same boat, or on your way to filling you panels.