AI+DeFi is ready to go, are there any AI+ stablecoins?

Original author: 0x Jeff

Compiled by: CryptoLeo, Odaily Planet Daily

Editor's note: 2025 is the year of the rise of the stablecoin track, with the United States launching the GENIUS Act, and South Korea's new president Lee Jae-ming is also fulfilling his campaign promise to allow local companies to issue stablecoins. In addition to the national level, there are Standard Chartered Bank and Flowering Period first, and JD.com and Ant Group later. Domestic and foreign corporate giants are exploring various stablecoins for hairstyles.

Encryption AI researcher Jeff published an article analyzing the existing problems of existing encryption + AI projects, which are too "pro-AI and scattered to encrypt", which makes it unable to gain a foothold in DeFi. In addition, Jeff also listed existing AI+ stablecoin projects worthy of attention, compiled as follows.

Stablecoins are one of the most important infrastructures created for cryptocurrencies. Without stablecoins, we wouldn't have stable units of money for investors to invest (which would make it very difficult to build CEX, DEX, Perps, money markets, and any other verticals).

Stablecoins are rapidly gaining popularity - from 2023 to 2025, the total supply, transaction volume, circulation speed (stablecoin transaction frequency frequency) of stablecoins has increased sharply, especially in payments and cross-border transactions.

In addition, we have also seen clearer regulatory regulations and further institutional stablecoin adoption, such as Stripe launching stablecoin financial accounts in 101 countries, Societe Generale will launch a US dollar-backed stablecoin, major banks (Bank of the United States, JPMorgan Chase, Citibank, Wells Fargo) are exploring joint issuance of stablecoins, and large companies are exploring stablecoin payment options to reduce transaction fees for Visa and Mastercard, etc.

Recently, CRCL (Circle) IPO has also set off a stablecoin craze, attracting more stakeholders. While we are seeing TradFi being further adopted, we are also seeing some stablecoin innovations emerging in the field of artificial intelligence, aiming to solve the challenges encountered by service providers and users in Web3 AI.

The first challenge

Although AI teams often design AI tokens as a key part of the AI ​​ecosystem (payment, governance, utilities), they usually invest less resources in DeFi and more resources in AI products.

example:

• Virtuals Protocol uses their VIRTUAL/AGENT LP, which brings good value-added to VIRTUAL, but also makes it difficult for agency teams and liquidity providers to provide liquidity (due to impermanent losses);

• Aethir uses ATH tokens as computing power payment, which plays a push role for the token, but also increases its volatility;

• Bittensor uses dTAO (alpha subnet tokens) to pay miners, validators and subnet owners, and participants must sell alpha tokens in exchange for stablecoins to maintain their operations;

While both examples can be considered flywheels for AI tokens, they will also prevent some key players from participating due to the volatility brought by the design. (BTW, these 3 examples are relatively excellent, but there are also many AI teams that do a pretty bad job of token design, especially some fair launch teams).

The increase in the number of tokens on the market, coupled with the not optimal design, leads to low liquidity and making it difficult to build DeFi Lego bricks.

Projects to deal with the problem of “uneven resource allocation”

MAITRIX —AI Stable Coin Layer

Maitrix has launched an over-solidated AI native stablecoin (AI USD) tailored to each independent ecosystem, essentially transforming an unstable (but highly productive) AI economies into a predictable, composable and dynamic economy with AI native stablecoins.

Key components of Maitrix:

CDP: Users deposit AI tokens and their derivative assets (current pledged or pledged AI tokens) through CDP, thereby minting and destroying AI USD;

Stablecoin Launchpad: AI projects can create their own AI stablecoins using their native tokens and their derivatives;

Curve War ve(3, 3) Incentives: ve governance, emission redirection and bribery mechanisms similar to ve(3, 3)

StableSwap DEX: Implement transactions between various AI USD tokens.

Supported AI dollar assets (so far)

• Aethir USD (AUSD) - a stable payment method for computing power

• Vana USD (VanaUSD) – Data-based stablecoin

• Virtual USD (vUSD)

• ai16z USD (ai16z USD)

• 0G USD (0 USD)

• Nillion USD

There are more partners in talks.

There are not many detailed documents on the use cases of each AI stablecoin. After their white paper comes out, I will introduce the technical content in detail. But for the moment, Maitrix is ​​the only team to build this layer for AI projects and has established partnerships with the top AI ecosystem.

Maitrix is ​​currently attracting much attention on the test network. The public main network is about to be launched.

The second challenge

With the continuous accelerated development of AI and the expansion of application scope, the demand for computing resources has also increased. Data centers and cloud operators need to plan their expansion in advance to meet future needs.

Enterprise-grade GPUs such as NVIDIA's H100 and H200 are often expensive and require a lot of money.

Traditional financing methods, such as bank loans or equity investments, are often slow and complex, causing data centers to not scale quickly to meet demand, which is where the next two AI projects, Gaib and USDAI, focus.

Respond to “demand issues” projects

The first project is Gaib.

GAIB AiFi – the first economic layer of AI and computing power, Gaib tokenize the future cash flow of GPUs to help data centers raise funds efficiently while providing investors with earning assets powered by real assets (GPUs).

It basically works:

• Cloud/data centers package future GPU cash flow into financial products;

• These cash flows are tokenized in a cycle of 6-12 months;

• Investors buy these tokens and start receiving regular rewards;

• They call this AI-synthesis dollar “AID.”

Each AID token is backed by a portfolio of GPU financing transactions and Treasury bond bonds or other current assets reserves

Floating returns are expected to be around 40% APY, which depends largely on the portfolio of GPU transactions, whether it is debt financing or equity transactions (equity accounts for more than 60-80% while debt annual yields are 10-20%).

So far, they have accumulated about $22 million in TVL incentive deposits, which exist in the form of “Spice” points, and investors will be entitled to future airdrop rewards

In addition, Gaib has also reached a partnership with Aethir, conducting its first pilot GPU tokenization earlier this year. The pilot is just GPU tokenization/segmentation as part of its roadmap, expanding to the GPU-backed stablecoin “AID”.

The second project is USD.AI

USDAI is an interest-generating synthetic stablecoin powered by RWA launched by Permian Labs. It is very similar to Gaib to some extent, but it is also different. USDAI is a stablecoin secured by hardware assets (such as GPUs, telecommunications equipment, solar panels) as collateral. Its operations are debt financing transactions where the borrower (asset owner) receives loans from USDAI and pays interest, and these interest income belongs to the USDAI token holders.

Behind Permian Labs is metastreet , a top structured credit market that provides NFT-secured loans, structured credit for non-liquid assets/risk-weighted assets (watches, artworks), and NFT earnings transfer (PT YT) similar to Pendle.

USDAI is not online yet, but its target yield is 15-25% APY, and its portfolio is divided into three stages, from 100% Treasury bonds to 100% hardware assets. USDAI adopts CALIBER, a system that simplifies the loan/issuance process and complies with legal standards for bringing GPUs onto the chain.

Odaily Note: CALIBER: Collateralized Asset Ledger: Insurance, Bailment, Evaluation, Redemption (guaranteeding asset ledger: insurance, custody, evaluation, redemption). This system is based on Article 7 of the United Commercial Code (UCC), and through asset tokenization and legal frameworks, converts real-world assets (such as infrastructure) into legal collateral that can be used for on-chain financing.

To clearly distinguish, USDAI focuses on debt and has a wider range of assets. With its CALIBER model, they can cover a variety of use cases (regardless of demand), while Gaib is more focused on equity, providing higher expected yields.

You can fill out the form to apply to become an early user and USDAI will provide additional rewards for early participants.

Other AI-related stablecoin products

Almanak recently launched the alUSD, ERC-7540 version token (an expansion of ERC-4626), which is a tokenized AI earnings optimization strategy designed to maximize the risk-adjusted returns of stablecoin investments such as Aave, Compound, Curve, and Yearn.

The Almanak team will soon launch the points campaign to channel liquidity and continue to expand DeFi composability so that people can use alUSD as collateral or recycle to maximize profits.

AIxFI project, a vault that can automatically deploy USDC in the DeFi protocol. Initially, it will be based on rules and then gradually introduced to AI for decision-making. Will be launched this month on Virtuals Protocol.

Future trends

We’ll likely see the rise of another Ethena project focused on leveraging GPUs to bring high returns to stablecoins. More importantly, how they manage their 1:1 dollar anchorage and make sure the price returns to $1 in case of a critical situation.

In the future, we will also see more tokenized AI strategies emerging. We have seen that AI can better optimize returns when considering Gas fees, rebalancing fees, slippages, and other dynamic variables. Imagine tokenizing these strategies into highly composable "treasuries" that can be used as collateral or recycled to achieve 5-10 times leverage returns.

As participants such as Maitrix build stablecoin infrastructure for top AI ecosystems, we will begin to see an increase in Web3 AI liquidity. More AI value will begin to become more composable and flow to DeFi, thereby increasing the value-added value of the entire Web3 ecosystem.

Although these teams are very interesting, the risk/peg management/redemption/liquidation mechanism is crucial when it comes to stablecoins. Before deciding on investment, do a risk assessment.