Liquid IP: IP-NFTs and Bonding Curves
Exploring A Universal Distribution of IP via NFTs + Curation Markets
This article explores the practical use of token bonding curves and curation markets to fund, curate and distribute ownership of intellectual property. In short, Token Curated Intellectual Property. The goal of the proposed design mechanism is to distribute risk, reward and ownership of intellectual assets and enable market participants to make early research and development stages liquid and tradable. This could have a wide range of applications in scientific development, pharmaceuticals, software development and engineering.
We outline how a combination of Non-Fungible-Tokens (NFTs) and Token Bonding Curves (TBCs) can create tokenised IP and how this could result in better innovation cycles. We also determine how a potential IP curve could look, as well as the actual utility of the issued tokens. This is rounded off by a practical example of the proposed mechanism in the pharmaceutical industry.
Curation markets were first proposed by Simon De La Rouviere. Some excellent reading material:
The concept of token curated intellectual property was first presented at a workshop at Dappcon 2018. Slides are here and also give a general intro.
Patents Were Introduced in the 1800s
The exclusivity of ideas and inventions — copyright — has given rise to some of the most successful and influential companies over the past centuries. Intellectual property can take many forms, most commonly patents, which depending on industry and use are issued for several decades. Patents were developed in the 1800s to allow individuals and organisations to take calculated bets on the pursuit of new technological inventions, allowing them to reap profits and recuperate their investments. However, state-enforced monopolies create problems like price gouging, patent trolls, rent extraction and depress scientific advancement and collaboration. In an age of open-source development and information sharing, tying the ownership of intellectual assets to individuals and corporations seems backwards. In the information age, contributors benefit most when information is freely shared.
Take the pharmaceutical industry, for example, where drugs are commonly sold at a thousand times their production value, creating spiralling cost structures for healthcare systems. Closed-source development requires higher investment costs by single entities, which in turn mean higher prices for consumers. In high-tech academic environments, this seems specifically problematic. Moreover, good concepts often don’t succeed, not because they are objectively bad, but because they are legally tied to a failing team or bad leadership. Lastly, access to capital remains unevenly distributed and is commonly tied more to relationships and geography than the actual invest-ability of the intellectual property.
What if, instead, we had a meritocracy of ideas and open-source development of property that is freely investable. A system to incentivise cross-collaboration and share the risks and rewards through distributed collective ownership. A system to curate the most valuable concepts and make them fail fast without tying them to individual organisations. Open-source innovation systems lead to faster development cycles, testing and feedback loops. Such a system could be applied to anything from molecular compounds, engineering plans and software with the latter already being widely applied in ecosystems like Ethereum. Open development of intellectual property could also lead to lower prices for consumers and increased market competition.
The following 3 step program outlines how this is feasible, today!
Step 1: Make IP Ownership Transferable via NFTs
Intellectual property tends to start as a singular and patents commonly have singular ownership. Patents in their simplest form comprise of some sort of proprietary data with a timestamped claim. If we can attach the unique data of a digital cat, or CryptoKitty, to a unique Non-Fungible Token (NFT, ERC721), why couldn’t we simply attach the data of a fusion reactor design or of a new molecular compound to an NFT? The data contained in chemical intellectual property, a so called Markush structure, is surprisingly not significantly more complex than the colour palette of the fur of a digital cat.
Step 2: Distribute Ownership via Token Bonding Curves
Now that we have an NFT that represents our IP we can freely trade it with anyone. But we now want to measure the attention and value of ideas and create Schelling points around them. We want to curate the best possible fusion reactors, molecules and kitties and allow anyone to purchase stake in them, sharing the risks, rewards and responsibilities that come with that ownership.
The NFT that represents the IP is now put into custodianship of a token bonding contract by setting the owner address to that of the bonding contract. This means that anyone wishing to purchase shares in the IP can simply send collateral to the contract and receive a fungible token in return, commonly an ERC20. The price is set by a pre-defined slope by the contract creator. Besides distributing ownership there is now also a way to measure attention in a specific concept. The concept of Re-Fungible NFTs via TBCs was first introduced by Billy Rennekamp early 2018.
But why are we attaching the IP rights to an NFT first, instead of straight to the token bonding contract? One, to fully transfer the legal rights of the IP (a singular) to the token bonding curve, where it becomes fully inert or non-transferable. Secondly, this structure works both ways: if all tokens issued by the TBC are sent back to the contract it could institute a mechanism to return the ownership to a single address.
Anyone can now monetise ideas or intellectual property from their earliest stages, to more advanced stages and have on-chain liquidity. In general, various stages of the development of a new invention require various levels of investment, risk and return. A token bonding curve allows us to measure this and trade ownership and attention at any stage. It also allows people to put their money where their mouth is and create skin in the game signals.
Example: The Pharmaceutical Industry
We examine what the innovation and development cycle in a TBC market could look like in the pharmaceutical sector. The development cycle to bring a new drug from discovery to market is typically 10–15 years. Let that sink in: 10–15 years of closed source development, during which no one else is allowed to work on or research a specific compound class. And the average cost is $2.5bn. Due to the fact that a single company retains the IP it takes on an enormous amount of risk in the process, as well as outsized returns.
Typically, pharmaceutical drug development goes through 5 distinct stages, each of which significantly increase the value of the IP, from early drug discovery to large-scale clinical trials that verify the safety and efficacy in humans.
Next, we will model this value creation process in a token bonding curve. Instead of using a linear or exponential function, as is often proposed, a Sigmoid function seems most suitable.
Step 3: Define a Price Curve that Makes Sense
S-shaped curves are commonly used for markets that stabilise after a certain period, or after a publicly perceivable inflection point. Sigmoid functions can be found in many natural processes such as population growth and population density. A Sigmoid function could be suitable for intellectual property or ideas, because once a certain idea has been proven to work and its features discovered, then its market value should become somewhat measurable and find a stable price. In our pharmaceutical example the curve could look like this:
In this example, market participants purchase more shares along the curve as more positive information about the compound is revealed through clinical trials, or vice-versa. Ultimately, when the majority of information is discovered and the drug is approved to go to market, the market value stabilises, enabling commercialisation and pricing. Of course, this is an oversimplification of the economic processes involved in revealing more information about a molecular compound — but it’s a first attempt at more stable price discovery and attention measurement. We are working on this use case at Molecule.
For further exploration of various curve models check out Wilson Lau’s article exploring rationality of linear functions, exponential functions and rule-based functions.
The Need for Dynamic Bonding Curves
The approach of choosing a fixed bonding curve, such as a Sigmoid Curve at the asset’s creation, is also highly problematic, as new information emerges. Take Sildenafil (brand name Viagra) for example. Sildenafil was initially developed by Pfizer to treat high-blood pressure with a relatively limited market potential. Assume that the initial curve was thus defined based on that potential. During the clinical trials scientists discovered that Viagra had its now famous side-effects, creating a massive multi-billion dollar market for the treatment of erectile dysfunction. How would the initial curve creator have been able to take that into account?
At the current stage of research, bonding curves have no mechanism to become dynamic, or change their curvature based on external inputs or new information. Changing the curve would also be highly disputed and hard to implement while the market is live, as proposed changes to it would bring turmoil as participants exploit information asymmetries about proposed changes. Heightening of the curve would cause buying pressure, whereas a proposed lowering would likely cause sell pressure. The two most suitable mechanism for dynamic curves could be:
- Oracles that change curvature based on set parameters. For example like interest rate changes, or scientific breakthroughs.
- Governance mechanisms enabled by token holders (potentially on a system like Aragon).
Token bonding markets could also be designed in a way to allow for incremental pre-defined development stages. Once a certain milestone is hit, the market locks up and the curve is re-evaluated and potentially adjusted.
Token Utility & The End of the Curve
A fundamental question remains as to what rights the issued token represents. Remember, the TBC is still custodian of the underlying patent or IP in the form of the NFT and now also holds on to all the capital. The token now represents a share in that intellectual property, and if the idea has a legal contractual real-world form, then they are a form of claim to that asset and a security. This is fantastic, because it actually enables the tokens issued by the TBC to have a real world use case.
Individual tokens could now represent license rights, enabling holders to receive royalties whenever the IP is used. This means beyond the token representing a part of the overall value of the asset, they could also provide a steady revenue stream in the form of royalty payments. Many such processes are already captured in legal licensure or royalty agreements today, so tokenising them is not far fetched.
Another consideration is for the token bonding curve to end once a certain development milestone is reached. For example, a pharmaceutical drug in development receives FDA approval. From this point onwards the total market value of the drug should be established and the risk is removed, meaning the token could become freely tradable on secondary markets. Should secondary markets ever fall below the value contained in the TBC contract then it acts as a fallback mechanism and permanent liquidity pool until the intellectual property becomes worthless. Imagine a TBC in the 90s where everyone is pulling out collateral from the Walkman curation market to invest into CD and MP3 markets.
Creator Rewards and Funding Mechanisms
Upon the creation of the IP market, the IP creator may deem to retain a specific share in the initial amount of tokens, essentially paying close to zero for the first 10% issued. As more collateral is staked into the market, the creator may opt to sell a small share of her tokens in order to redeem collateral from the contract and fund the development of the IP. In addition, she may opt to transfer tokens to other parties who could also fund themselves and so.
Another useful funding mechanism is curve taxation. Here, every time a participant purchases tokens from the TBC a small extra share (the tax) is created and sent to a set of core contributors. The contract imposes a lower sell-out curve than the buy-in curve, meaning that a participant would lose a small amount of collateral if he immediately sold tokens back to the contract. This collateral is effectively retained for the contributors. For example, Alice purchases 1000 tokens of Bob’s project for $1000, which has 1% curve taxation. The contract sends 1000 tokens to Alice and 10 tokens to Bob. Alice’s sell out value is now $990. Bob’s sell out value is $10 (note: math grossly simplified). The process of selecting contributors could be managed by a Token Curated Registry.
Conclusion
The combination of token bonding curves and NFTs present an entirely new way to trade ownership of intellectual property in a variety of industries. Specifically, the use of TBCs alleviates uncertainty around IP investment with a clear pricing model and a safer collateralised, transparent funding model. This could fuel new innovation cycles, as token holders have incentive and legal rights to contribute to its development. In the scientific realm, the effects of this could be significant as this mechanism creates economic incentives for open-sourcing data about IP via skin-in-the-game rewards. Positive data outcomes lead to more attention and staked value into specific IP and vice-versa. Negative data leads to the opposite.
In the future, we see such an architecture interfacing with data exchange protocols such as Ocean, creating a double incentive layer for market participants: 1) create open source datasets about IP due to skin-in-the-game incentives and 2) receive rewards for open sourcing the dataset on a data exchange protocol.
Thanks for direct, indirect and inspirational input from Simon De La Rouviere, Billy Rennekamp, Florian Buehringer, Balazs Nemethi, Slava Balasanov, Wilson Lau and Trent McConaghy!
