Dave Carlson is a software engineer and 10-year veteran startup entrepreneur. Beginning in 1999, Dave has played a primary role in several multimillion dollar startups and personal ventures.
In 2010, Dave discovered bitcoin and soon after founded “decentralized” bitcoin mining firm MegaBigPower.
2015 continued the trend: as the price of bitcoin fell and mining difficulty rose, miners were forced to sell both their bitcoin holdings and newly mined bitcoins to cover operating expenses and fuel growth.
This increased downward pressure on the price of bitcoin, and gives us all a painful reminder of how this network is designed with respect to miners: bitcoin punishes the greedy and the weak.
Just as in 2014, we saw attrition in mining operations (and among manufacturers) – the less favorable economic conditions meant many were unable to survive. The explosion of hashrate coming from China eclipsed operations in other parts of the world, and we quickly found the conditions at which many operations had to ‘blink first’.
Early leaders in the ASIC market found themselves lagging newer players, and in a declining market, struggling to find the funds to maintain their position.
To carry the same market share, industrial-sized operations needed to increase from 3-5MW to 30-50MW.
It becomes readily apparent that block rewards for mining are simply a subsidy designed to return on the investment in network security, not to produce long term revenues. Mining is now a race to see who will hold the biggest marketshare when the network transitions to a fee-based system.
Nobody loves bitcoin
In 2015, the market decided it didn’t like bitcoin anymore and the focus turned to the blockchain, its underlying distributed ledger.
This was silly, of course, because the two go hand in hand.
What the market was really saying was that it was unhappy that the bitcoin price didn’t hold up. In retrospect what happened to the price makes total sense: more mining was being done than was necessary to support the transactions being made.
Bitcoin production rates are essentially fixed, but difficulty is not.
So we drove ourselves to the brink in our attempts to grab marketshare, couldn’t (or wouldn’t) continue growing and waited for prices to rise again.
Mining is a commodity business and its inherent price sensitivity was never more apparent (let this be a lesson to you).
Usher in the blockchain: a distributed ledger system with super cheap transactions and almost costless auditing – recording exchanges of value via a digital token.
Indeed it’s very cool tech – but then I just described bitcoin, didn’t I?
Perhaps the most annoying question we asked ourselves this year was whether to raise the limit on the number of transactions that can be written into data blocks on the bitcoin blockchain.
I was against this can-kicking, central bank-esque meddling from the beginning. I stood before the Satoshi Roundtable in February and proclaimed my intent to just mine for rewards, leaving transactions out of my blocks completely.
This of course “stirred debate”, but I think I got my point across. With almost no revenue coming from fees, miners have literally no incentive to mess with transactions.
During the transaction spam “stress testing”, we witnessed this same idea as a defensive tactic – full blocks impact pool performance, causing orphans, or valid blocks that aren’t included on the main chain, meaning miners who mine them don’t get the rewards.
With such low fee revenue, these transactions now become a liability.
Think about what a single missed block costs a pool operator. At $450, we make $11,250 for processing a block. If instead those transactions cost us $11,250 well, you can see the motivation.
Now, speaking of motivation, bitcoin has always been a system whereby miners vote for protocol changes by taking (or not taking) a particular BIP. This provides a consensus mechanism for change to the network only when there is a measurable (read: financial) incentive to a majority of the mining hashpower.
Maybe now it’s clear why we don’t have consensus on the issue.
So what if we don’t take the BIP? The ever-rising daily transaction count will soon fill all the blocks and then we will have competition for transaction clearing. This will raise fees.
The shoe may be on the other foot at that point – startups with poorly conceived business models based on super-low transaction fees may be the next to fail. But it’s possible that projects like sidechains or the Lightning Network will provide aggregation solutions.
In a few years, perhaps only the largest aggregators will be able to afford to post transaction data to the main chain.
This all sounds very painful, but it’s actually a great and elegant and terrible evolutionary process happening all at once. It has the effect of taking out the garbage, so to speak.
This old saw again.
For miners, 2016 requires further optimization of operational expenditures, in addition to the need to transition away from less efficient hardware.
Even with very low power costs, miners need to move to the better gear in order to survive. Once you have very cheap power, where does a miner go?
I intend to spend 2016 pursuing more free power, and developing a ‘net-zero’ power cost business model. This can be achieved when you operate at megawatt scale and have a large ‘controllable power load’.
Where in the world can you run your old hardware that costs more in power than it makes in bitcoin? There are places.
MegaBigPower is rolling out a new offering for this old stuff to get a second life. This follows our continued efforts to find ways to monetize this equipment rather than see it heading to landfills.
Competition at industrial scale
My first industrial-scale ASIC project achieved 100 terahashes and we began the deployment when the global network itself was 100TH.
So, by the time we got it all running, we had something like 25% of the network. To attempt this same feat again will require some 750 petahashes, or 5.4 million of BitFury’s new chip.
Will it happen? Yes, definitely – and more no doubt. As bitcoin prices rise, the system has room for more mining.
But since we are greedy humans, we will overproduce, difficulty will rise, profits will fall, putting pressure on the price and the whole cycle will repeat again…
Don’t overlook what has happened, though. Bitcoin and its blockchain have become even more robust and more secure as a result.
With this newfound security and robustness, the network has greater value and eventually price will rise as a result of renewed investment and trust.
As the price rises, so does market capitalization and so does the amount of revenue to be had from processing transactions.
Right now, I estimate the 12-month revenue potential, if one could do all the mining, at about $700m.
When the mining market can pay billions, larger companies are going to take a serious look at this. Companies that already have assets such as power generation and capital resources will task their very smart power managers to build long-term strategies around large, controllable power loads, hedging and energy storage systems.
These companies will begin to see how mining (transaction processing) can ‘plug in’ to their existing business and even subsidize poorly performing assets. I’ve personally seen signs that this process is beginning already and will accelerate in 2016.
The good news here is that the security and value of the network can continue to grow even as it remains decentralized.
2015 was a year that tested our resolve as miners and forced us to adapt or die.
I believe 2016 carries more potential for the business of transaction processing than ever before, but it may not look anything like we thought it would back in the heyday of 2013.
The same goes for the blockchain. We are now looking at using the blockchain (and implicitly bitcoin) in ways that achieve a similar vision but are not nearly so straightforward as “Send x bitcoins from User A to User B”.
The potential increases for alternative chains to take up space in large mining centers. Private Custom Blockchain hosting is right around the corner.
I have a rule of thumb when it comes to new businesses: any good business takes 10 years to develop, and transaction processing is well on its way to reaching that milestone.
Mining image via Shutterstock
Disclaimer: The views expressed in this article are those of the author and do not necessarily represent the views of, and should not be attributed to, CoinDesk.