Profitability of Bitcoin miners, are they still profitable?

Reddit's miner expert, praeluceo, on bitcoin mining:

Many individuals are still making a profit. Namely those who made smart hardware purchases and have access to free or inexpensive power.

  • If you're asking for currently commercially available hardware (I send X BTC to Joe from BitcoinTalk/112bit/etc, and in <14 days receive hardware) your answer is no, because the exchange rate of Bitcoin to USD is too depressed. Your USD is currently more valuable being converted directly into BTC without the hardware investment.
  • If you're open to used miners, you may be able to purchase some already-ROI-d hardware from a miner on ebay, Craigslist, etc, who is selling their kit at or below the current ROI. This is still pretty challenging as the price continues to trend below $400, but not impossible.
  • If you're looking at a pre-order, I don't care if it advertises 1 Petahash for $1,000 with delivery by January 1st, it will never, ever, ever ROI, and likely won't deliver. Never pre-order.
  • If you're talking about hardware in general, including stuff that can't be purchased, then no they are not all unprofitable. Without the markup for consumers (controller on every hashing unit, fancy enclosures, inefficient cooling systems that prioritize dB over Gh, shipping, etc) the hardware can be created relatively inexpensively. Especially if you have access to a captive foundry and some decent ASIC designs.


That said, difficulty has stagnated the last 2 adjustments, and it would no-longer surprise me if we see a negative change if price continues in this trend (especially if we see the $200s again for any appreciable length of time). If difficulty trends downward, eventually it'll get closer to that break-even point, and thus more sales will have a higher likelihood of breaking-even or gasp profiting!

Most people see a miner purchase as a NRE(non-recurring engineering) and once they have it, just run it. Sadly, most home miners also treat "ROI" as "I now I have more cash than I did when I started", which means they depend on arguments about Bitcoin's value going up to defend their purchase. I would argue that the current price is not capable of sustaining additional consumer hardware manufacturing as their USD-based costs are pretty fixed (heatsinks, power supplies, PCBs, manufacturing labor) likewise, the power to operate a miner is fixed, as is how many GH/s you can run on a single 110 outlet (at current ASIC efficiencies). At this point, for a 1 TH/s miner (with no electricity cost) to mine 0.8 BTC will take 6 months.

Bitcoin Math:

Using that 0.8 BTC as its effective mining potential, and assuming free electricity, that means an ideal miner can only cost $289.6 USD to produce. Since a 1000W power supply is about $150, lets assume you also have all infrastructure required to mine right now. Given an impossible non-existent "ideal miner" target, let's make some additional assumptions:

Using the BE200 (from ASICMiner) as likely the most efficient ASIC on the market at ~10 GH/s @$0.5/ea (estimated from posts on BitcoinTalk). To achieve 1 TH/s you need ~100 of these chips, so $50.
To place 100 chips (using current miners as a template) would require about 435 square inches of PCB (probably across 4 boards, but that's not relevant), a rough best-case scenario is $30 for the PCB itself (more realistic cases here in the US is $100 for that much PCB, but these are made in China).
Required components for a BOM for that board? I don't know, eyeballing existing miners I say maybe $1 of components per ASIC, so a broad guess of $100 in auxiliary components for the BOM (although this is a hugely crude estimate, it could be as high as $4, but I'm going low-end assuming they get great price breaks).
Assorted hardware for heat sinks (screws, heat sink, thermal conduction material), this is hard for me but I'll guess $8 per 25 ASICs, so $32 (same pricing goes here, I'm no good at pricing this kind of thermal management, but I doubt they're going all-out either).
So my total ballpark for a 1 TH/s miner would be $212 USD, or 0.58 BTC/TH, just in parts. That leaves 0.21 BTC for anything else (board layout, prototyping, assembly, shipping, bills, etc.). Unfortunately, today 0.21 BTC is only $77, that's a 36% overhead. When I design hardware, I always price to 200% over my raw materials cost to factor in paying the above expenses, and I target a 15% profit.

TL;DR: I don't think hardware manufacturers can profitably build and deliver consumer miners at the current exchange rate. It would not surprise me if some of the current manufacturers were also irrational actors, and that their balance sheets don't add up.