AI crypto. What happens when artificial intelligence enters the world of blockchain? To what extent is it a reality, or something we won’t see until years into the horizon?
While the future of AI may be unknown, we do have a better idea of how crypto works. If there’s one reason blockchain became as big as it is, it’s because of the (lack of) entry barrier. Accessibility.
Without this freedom, Bitcoin wouldn’t be as secure or decentralized. But is also why there are ~23,000 crypto projects, most of which are now inactive or worthless. Whether it’s lack of utility or speculation, it keeps repeating whether it’s DeFi apps, NFTs, Metaverse tokens, or the next big thing.
With that said, there will still be AI cryptocurrencies worth investing in. Some will lead the markets for being the first ones in their industry, others for their utility, and sometimes long before having a native token (like the Uniswap exchange).
How do you find such projects in an endless sea of questionable tokens? First, let’s identify what AI crypto is, and what the best examples look like.
- AI crypto is the automation of complex tasks under the principles of blockchain: decentralized, trustless, and autonomous. Unlike smart contracts, crypto AI is adaptive.
- Data scientists can use AI tokens to buy data sets and services or stake crypto to support an AI and get rewards or penalties based on its performance.
- There are already use cases of AI crypto in cyber-security, trading, supply chains, healthcare, and construction.
What Exactly Is AI Crypto?
Given how broad both terms can be, let’s start from common sense and then expand:
AI cryptocurrencies are those that power AI services when interacting with their platform.
This is similar to the way Ethereum powers smart contracts, which are autonomous programs that execute an agreement reformatted into code. There’s no human intervention, and they can follow complex operations, but smart contracts are not AI.
The missing piece is adaptation.
Smart contracts follow preset rules and can’t “think.” Humans do the decision-making, learning, contextualizing, and problem-solving. AI would be able to do these complex tasks with similar intelligence to humans and the consistency of smart contracts.
Applied to blockchains, AI services are supposedly decentralized, trustless, autonomous, and adaptive.
Note that most AI tokens themselves don’t have AI components, but rather support or bridge AI services to on-chain records.
E.g. You could associate that ETH tokens serve to secure the Ethereum blockchain. But Ethereum has seen many forks like the Merge, and every time it updates, it can redefine what those tokens mean (like the switch from proof-of-work, PoW, to proof-of-stake, PoS).
The point is, there’s no such thing as “smart tokens” (yet). AI cryptocurrencies are generally used for:
- Marketplace trading. The AI token is the only currency allowed to buy AI services or data sets (used for AI training)
- Staking. Developers can stake AI tokens on the data sets and models they create to prove confidence in their integrity and quality. There’s often a positive incentive for being right (AI token rewards) and a negative one for being wrong (stake slashing).
AI has largely remained a niche industry that primarily serves data scientists, tech companies, and dApp developers. A lot of use cases are conceptual, such as data marketplaces and predictive analytics. However, as blockchain merges with AI, we may soon see more real-world applications.
Some known AI examples are:
- Supply chain optimization
- Digital twins
- Personalized healthcare
- Threat intelligence
- Fraud detection
- Predictive market analytics
- AI marketplaces
Examples of AI Crypto Projects
The following 5 AI crypto examples may sound abstract to nontechnical investors. That’s because they’re targeted at developers and researchers, whereas for traders, AI tokens aren’t that different from other altcoins. Long-term, however, they’re relevant for their broad variety of use cases.
The Graph (GRT)
The Graph works as a data-sorting layer between the blockchain and the dApp ecosystem. Developers use it to quickly fetch complex data from the blockchain to their platform, which otherwise would take days to find and even longer to program.
The Graph has a complex economy involving data sets (called sub-graphs), indexers who create subgraphs, delegators who stake on indexers, curators who stake on subgraphs, and developers who use the GraphQL language to query relevant data. Because there are GRT tokens at stake, each role ensures that the sourced data is relevant and up-to-date.
The Graph is a Certik-audited cryptocurrency with a circulating supply of ~9B GRT. It launched at $0.03 each on its public sale in December 2020, and its 24h volume now ranges from $10M to $45M. You can buy GRT for ~$0.15 on Ethereum, Solana, Avalanche, and other blockchains.
SingularityNET is an AI marketplace for developers to publish and promote their services. AGIX is the currency used by diverse consumers, such as healthcare companies, financial platforms, and manufacturers. Service providers can leverage this marketplace with a service that’s fast, reliable, and cost-efficient, which will turn into more calls (sales) and fee revenue for the marketplace.
SingularityNET had its ICO (initial coin offering) in December 2017 at $0.10 per token. Today’s average volume is $15M to $65M for a circulating supply of 1.2B tokens around $0.40 each. You can buy AGIX using the Ethereum or Cardano contract.
Fetch.ai is a custom blockchain built on CosmosSDK with the goal of automating complex tasks on-chain. Developers use FET tokens to buy data sets for different use cases (automotive, manufacturing, social networks, urban mobility, smart grids). Other companies also use it to pay for on-chain IoT services, which are the processing and storage of data via “autonomous agents.”
E.g, Warehouses can take several days to process incoming inventory. This company could implement IoT devices and smart contracts to quickly verify and count inventory, so it’s available for sale within hours of arriving. From the logistics side, the shipping company might use both IoT and Fetch.ai to verify the shipping conditions at all times (such as time, speed, or temperature).
Fetch.ai has been audited by Certik and Slowmist in 2019-2020. It has a stable circulating supply of 800M FET that sold for $0.086 each on its ICO in March 2017. Daily volume is irregular, $10M – $20M with frequent spikes well over $100M. Fetch.ai sells at ~$0.38 per token on Ethereum, BnB Chain, and Cosmos.
Ocean Protocol (OCEAN)
Ocean Protocol is a data marketplace and set of tools for researchers. It allows users to monetize their data through encrypted, private transfers. Data “consumers” like scientists may buy these sets to train their own AIs.
Users interact with OCEAN tokens when buying from the marketplace or staking on data pools. The latter is to reward smaller providers and prevent big ones from dominating the marketplace.
Ocean Protocol is also a Certirk-audited cryptocurrency that held its ICO in March 2019 ($0.12/OCEAN). Its circulating supply started at ~230M and stayed at 613M since June 2021. Its daily volume rarely goes below $3M or over $20M (at least since the bear market). The only major chain where to buy OCEAN tokens is Ethereum.
Numerai (or Numeraire) is a decentralized, AI-powered hedge fund. It shares “data sets” for a decentralized network of data scientists, who use it to code their own models for algorithmic investing. Users only submit predictions from their model, while the actual code is never revealed.
Scientists can then stake NMR tokens to prove that they trust their prediction model. If it loses long-term, NMR doesn’t return to the protocol but is rather destroyed. If the model is profitable, then Numerai rewards the owner with NMR tokens, and more users also want to stake in it.
Those who rank in the Top 100 in this data science tournament receive the most rewards and confidence from users. Numerai also benefits because users outperforming the market means more revenue from maintenance and performance fees.
Numerai sold each token for $0.60 on its ICO in June 2017. It’s only available on Ethereum, it has no recent audits, and its circulating supply changes sharply. It jumped from ~1.2M NMR to 2.2M in 2019, 5M in July 2021, 10M the next month, and back to 5.9M within a week by November 2021. While it’s now ~$18 each, NMR consistently trades below $1M per day with $10M spikes every other week.
What’s Next For AI Crypto?
That’s about all notable examples of AI crypto. Except for two projects, the others (and likely most small projects we didn’t mention) didn’t start until 2021. AI crypto has barely seen two years of market evolution.
What is the investors’ reception of AI projects now that ChatGPT has reignited this trend?
Now that the general market shows signs of recovery?
It all suggests that AI crypto has a lot left to go in the price discovery phase for at least another two years.
The question is, how quickly will it occur? Until we see more ordinary use cases like ChatGPT, AI cryptos remain a niche industry and will take longer to develop than more approachable ones like DeFi or NFTs. At least when it comes to tokens, as they’re not always necessary to implement AI for blockchain tech.