Given that the container, immersion cooling, etc. are all based on existing technology within allied-control, and that two sites with low electricity costs and no legal problems have already been identified, why would it take so long (3-6m) to deploy containers? Surely the deployment time can be shortened and have less variance?
so from the Prospectus you mention "minimized manufacturing time and prefabricated DataTank mining systems... deployment can be carried out in hours, instead of ays or weeks." But the mining won't start for 3-6 months?
Short answer:
The goal is to deploy in the fastest time possible, but hardware and infrastructure deployments on a megawatt scale take time. It also has to be done internationally and safe without risks. Furthermore, it is not uncommon that manufacturers of electronic or infrastructure components take up to 12-16 weeks to deliver, or even longer on such a scale.
We have good people involved and hands-on experience building some of the largest and most efficient mines ourselves and will do so again. We also have our own production facility literally in-house, within our very own rooms.
Long answer:
In the case of the ASIC chips the delivery time is a price issue, in the case of other important components it's a quantity and industry issue. Example, a few hundred or a few thousand buck regulators can be purchased online, but for 100k+ back regulators Texas Instrument or any other company will take several months or do partial deliveries, even in China.
Insufficient time also leads to inefficient and unproven board designs. Board designers get it wrong all the time, not because they are bad engineers but because everyone wants the boards out the door ASAP. Ultimately this leads to higher cost, lost money and lost time.
We solve this problem by putting capacity in place first and foremost, and committing to hardware at the right price with the right performance in time for deployment.
Most mining businesses in the past focused on investing in mining hardware first, hosting and infrastructure cost afterwards was often accepted as collateral damage. At today's scale and economics (hosting and infrastructure cost exceeding HW cost), the focus needs to shift or ROI will be significantly affected. In addition to the arms race for the chips there is now an arms race for capacity as well.
Based on the forecast in pg 15 of the prospectus, deployment in 3 months would mean that DTMA will not reach ROI. 6m would be worse of course.
That is absolutely correct. The solution is to put capacity in place first and then buy hardware when it is guaranteed to be efficient and can be brought online in the shortest possible time.
The reference data in the prospectus is based on today's known facts and prices including safety margin. It can be assumed that future hardware prices will be lower, but the business plan cannot be based on assumptions alone, hence today's prices/economics are used to represent the business model.
This is indeed a departure from the usual "buy hardware first and get rich quick" schemes which often overlook important issue such as hosting, logistics, manufacturing etc - we firmly believe the long term approach in mining requires such a radical change in strategy.
We plan to be very profitable from the outset, but it's the next generation where it will make the biggest difference. Mining is to stay and our solution can be used again and again.
Secondary deployments literally take hours, with the complete structure being reused. This is true for DTMA investors (capacity + hardware) and DTMB investors who only invest in capacity.
When a better chip becomes available after DTMA starts operations, what is the procedure for swapping to the new chip and how would it be funded?
When hardware in DTMA capacity (hardware + capacity) stops being efficient, it can be switched off and DTMA effectively becomes the same as DMTB (capacity only).
DataTank Mining is working with hardware and ASIC manufactures (SHA and Scyrpt) to make immersion blade hardware available. DTMA or DTMB holders can chose to purchase (at internal cost) new hardware at any time, either by reinvesting paid dividends or raising new funds. Alternatively unit holders can send in their own hardware and get it deployed instead.
How would DataTank Mining's interests be aligned with shareholders'? As far as I can tell, once the container has been sold, DTM just makes a tiny cut of electricity bills, so your only incentive is to keep the container operating?
The cost of operation and building of DataTank infrastructure is passed through, there is no profit involved.
DataTank Mining's profit comes from deploying 20% of _identical_ hardware alongside the public capacity. Hence, we only earn money if we are successful. The 20% fee on operating cost is used to fund the day-to-day operation, including on-site staff providing various services such as ongoing hardware re-deployments, maintenance and security.
