Note: I am a participant in the Numerai Tournament and hold/stake NMR.
Numeraire (NMR) is a cryptoasset created by Numerai to power the Erasure protocol. Erasure enables decentralized data marketplaces with various mechanisms for encrypting, staking, and burning on the Ethereum blockchain. Numerai is a global equities hedge fund that uses Erasure to run its data science tournament. Numerai encrypts and exports its own financial data to a group of data scientists in order to create a “meta-model.” Numerai can then use the meta-model's predictions to allocate fund capital. The meta-model is created by ensembling prediction submissions from these data scientist users. Then, users are able to stake NMR on the efficacy of their submissions and can earn or burn NMR based on the performance of the submission on live stock market data. The company is also developing adjacent tournaments like Numerai Signals (using Erasure). Signals will allow data scientists to bring their own data for predictions instead of having to rely on the Numerai encrypted data and targets. There are likely to be more Erasure on-ramps in the near future for 3rd parties to more easily incorporate staking and burning mechanisms into their own apps/dapps. See Placeholder and Joel Monegro's Erasure thesis for more background on the protocol and its possible use cases.
NMR token valuation frameworks
To uncontroversially set a few assumptions, the NMR price will be whatever someone else is willing to pay for it. If there is more buying volume, the price will go up. If there is more selling volume, the price will go down! What someone else is willing to pay for your NMR is a function of how much they think they can earn from holding or staking it. (So, the only way to “value” a token like NMR is with long-term expected supply/demand.)
A good way to figure out how long term supply/demand equilibrium of NMR will be set is to see what happens at logical extremes. In one possible extreme, imagine Numerai has already paid out all of its NMR and has to buy NMR on the open market to pay users based solely on the profits of the fund. In that case, is the supply-demand equilibrium met when the value of NMR is equal to the net present value (NPV) of all future payments from Numerai? To estimate that, you'll need to model out Numerai AUM, expected rate of return, etc, into perpetuity. In isolation, the NPV of future payments approach has a problem because it assumes the network and protocol will look exactly the same in the future as it is now, and that is obviously not realistic.
Currently, NMR represents a sort of license to do machine learning work for Numerai. Staking must be done with NMR. We've already seen a pivot to DAI staking and NMR as a "source of scarcity" in ErasureBay. We also know that Erasure can enable many more dapps to gather information with staking/griefing/burning primitives. Will that staking be done in NMR, DAI, or something else? All of those details matter a lot. Since NPV of future payments by Numerai doesn't scale if we assume other hedge funds or dapps start building or participating on Erasure, we need to generalize. For example, what if another hedge fund wanted to buy particularly good predictions on Numerai Signals? If we assume all staking and payments are done in NMR, then should the generalized value of NMR be the net present value that can be extracted from all future data staked on the Erasure protocol? Interestingly, the NPV extracted from all that data increases as more sophisticated players enter the NMR market to extract more value from it (e.g. if another hedge fund does buy Numerai Signals data, it may extract additional value from the network as compared to the value created solely from Numerai's use of that same data). More buyers, better value creation/extraction, more data, more buyers. Financial network effects.
Another extreme to imagine is what would happen if one user owned all Numeraire in circulation. In that case, that user would have a monopoly on the ability to earn rewards on Numerai, and so he/she can earn either 1) whatever price Numerai (or other data buyers) would be willing to buy NMR from the NMR monopolist (only to return it right away as payment) or 2) nothing because Numerai won't play this game and the price/value collapses to zero. This interesting dynamic shows the reflexivity at play in cryptoasset valuation. If someone (or even a concentrated few) were to own 100% of the Erasure Protocol, it would likely be worth $0. This is a good reminder that digital assets can't be valued like equities that represent claims on future cash flows. Instead, the value of a digital asset should be thought of as the abstract representation of value flows/extractions within/from the network. The network itself has value--the fact that many people are participating, coordinating, and exchanging--and so NMR does not only have value to the token holders who also stake on their holdings, but also to passive holders capitalizing the network who are earning residual value if usage in the network grows. What other extreme situations shed light on how someone might value the entire NMR/Erasure network? If all Erasure dapps had DAI staking and NMR was only a "source of scarcity" like in Erasure Bay, why would anybody want to own NMR? If burns make the supply of NMR approach 0, does NMR price approach infinity (joakim https://forum.numer.ai/t/nmr-price-and-value/745/2)?
Valuation
This all said, I believe the true, intrinsic Erasure network value is some combination of:
- expected future value extracted from all staked data on the Erasure protocol
- value as a "source of scarcity," where NMR can be burned but not minted
- value as a reserve currency for many staking applications (moneyness?)
- residual network value from capitalization, growth, speculation
To ballpark the NPV of future Numerai payouts alone (which we've already shown should be a conservative estimate of the entire Erasure network value), we can assume ~$60M in AUM for the company's master fund. The hedge fund runs a long/short, market-neutral strategy. The fund's target benchmark is the fed funds rate (risk free rate) and operates with 4 turns of leverage. Conservatively assume the firm can manage net 7% returns per year with a 10% operating profit that market buys NMR and pays users with that remaining profit, the NPV of those payment flows is ~$162M at an 8% discount rate. Like any DCF, this is going to be sensitive to any of these inputs but a rough estimate should at least show a floor valuation in the case the Numerai has only middling success and hardly grows.
In early stages of a cryptonetwork like Erasure, market structure, volume, liquidity, ease of access, etc. all may prove to be more important drivers of price than these fundamentals in any case. For example, recent price increases are the result of listings on major crypto exchanges which will allow quicker/easier transfers for NMR from fiat out of bank accounts, paypal, etc. After all, if long-term supply/demand is really the only way these networks reflect true value, increased access and demand is a very simple driver of price appreciation. This opportunity is still incredibly early and also risky.
NMR began with ~11M tokens in existence. Numerai still owns/retains over 50% of this supply in its treasury wallets. Excluding Numerai treasury tokens, 5.187M ($186.4M) tokens are in circulation. Currently 126.38K ($4.5M) are locked in Erasure contracts, so only ~2.4% of circulating supply is being staked. This is up from 9 months ago where staked to circulating supply was ~0.9%. Some large stakers have >$250,000 at risk. Once the network matures, this metric (staked to circulating supply) will likely be an important one to track for relative valuation. At this early stage, it's less informative, but 2.4% of circulating supply staked seems far too low to sustain. Other information about NMR supply, staking, treasury wallets, etc. can be found here (numer.ai/nmr). It's likely impossible to put any point estimate of value out there (aka I really just believe this is undervalued and misunderstood).
The current fully diluted market cap of NMR is $386M ($35.13/NMR) https://www.coingecko.com/en/coins/numeraire. A network that generates the information flows of a successfully implemented vision of Erasure should seemingly be worth more than $386M and likely valued in the billions. If Erasure captures even a small size of data flows in the financial industry, the ceiling is quite high. For example, AlternativeData.org estimates ~$1.7B of buy-side data spend per year from institutional funds. If Erasure generalizes to non-financial data, it's difficult to size those markets but the upside is obviously higher.
Catalysts
- Numerai Signals, an upcoming new tournament
- 3rd Party Erasure dapps
- Broader crypto adoption
Risks
- DAI staking becomes default behavior
- Poor returns at underlying hedge fund
- Regulatory risk on cryptoassets, tokens, digital assets
- Macro crypto & DeFi exposure
- Large "latent supply" from early VC buyers
