Proof-Of-Stake mining supports the most important cryptocurrencies today and is a new form of economic activity that emerged with the creation of Bitcoin: in 2017 alone, mining for Bitcoin generated ~$4B of revenues.
From an investor perspective, it is important to understand mining dynamics as they will continue to impact the security and price of cryptocurrencies. Almost 10 years after the creation of Bitcoin, we can draw a first set of conclusions on the natural consequences of Proof-Of-Work mining, and begin to see the traditional laws of economics apply to a new type of activity:
- Highly specialized hardware is developed and mass produced for the specific task of mining a cryptocurrency (Application Specific Integrated Circuits, or ASICs, are computer chips that are wired to do only one thing, but very efficiently).
- As one player develops leadership in this hardware, its advantage only strengthens, following an experience curve pattern (i.e. performance is correlated with accumulated experience). Ultimately it should be expected that one or very few players will control the entire supply of mining hardware.
- The leading hardware producer has a strong (irresistible?) incentive to mine themselves with new hardware, generating higher profits than everyone else (including their customers).
- Furthermore, hardware producers can arbitrage based on the price of the cryptocurrency: use more hardware in-house when price is high, offload inventory when price is low (and mining less profitable).
- Smaller miners are incentivized to pool their resources to smooth out their cash flow and outsource some of the overhead… creating large pools which ultimately “control” a very large share of computing resources (“hashing power”).
- As a result, the overall network looks less decentralized than intended, with a single or very few hardware producers controlling most of the market, and a few mining pools controlling most of the hashing power.
We believe that ultimately the same dynamics will play out for all Proof-of-Work algorithms (as long as the cryptocurrency reaches minimal levels of adoption and value). We are skeptical of ASICs-resistent algorithms: over time it is unlikely that there is no hardware optimization that will generate gains of at least 10-100x in efficiency. There is also attempts to create algorithms that dissuade pooling, with limited evidence for the viability of these approaches.
What is not very well understood is the impact on incentives and the risks created by these natural trends. But there are two risks that are clear at this point: (1) the main hardware manufacturers are in a position to capture the vast majority of profits from mining (smaller miners operating at very low margins), and (2) concentration of hashing power in large pools raises a risk of attacks on the network.
One hypothesis that appears to be proven so far is that these large players (hardware manufacturers and mining pools) have a strong disincentive to use their influence for straight malfeasance (i.e. 51% or double spend attacks): the price of Bitcoin would likely collapse following such an attack and undermine their whole business model. Similarly, miners that contribute their hashing power to a mining pool could easily start switching pools if their pool started to use their resources for an attack on the network, so it is not necessarily clear how much power mining pools really have.
Efforts to avoid and modify the side effects of Proof-of-Stake mechanisms will be interesting to follow, as innovation in the space could be central to the future of cryptocurrencies.