BitGo Holdings has created BitGo Labs, a research and innovation unit focused on developing digital-asset applications for an economy where artificial-intelligence systems may initiate payments, move funds and require custody services without waiting for traditional banking processes.
The company said BitGo Labs will initially concentrate on combining AI with crypto-native financial tools, with work spanning payments, settlement and custody. Its first projects are expected to explore micropayments and instant settlement, two functions that could support frequent, low-value transactions generated by software agents rather than people.
The initiative places BitGo’s existing custody and infrastructure business closer to an emerging question for digital assets: how autonomous or semi-autonomous AI systems would pay for data, computing resources and online services. Conventional card and bank-payment networks can be inefficient for transactions worth only a few cents, particularly when software may need to make many payments in rapid succession.
BitGo Labs starts with AI card application
BitGo Labs has already begun work on several products, including Gokubu, a consumer application built around an AI-powered payment card. BitGo said the product is live and assigns users an AI financial assistant designed to help optimize spending, rewards and transaction value.
The company has not provided detailed performance data for Gokubu or disclosed how its assistant determines which spending choices offer the greatest value. Its launch nevertheless gives BitGo Labs an immediate consumer-facing test case, rather than limiting the unit to protocol research or enterprise infrastructure.
BitGo said premium Gokubu users can receive a 6% return in the product’s native token on the first $2,000 of spending, subject to a monthly subscription fee. Such token-based reward structures can increase engagement with an application, though their value depends on the token’s liquidity, price stability and the conditions attached to earning and redeeming rewards.
The card product also illustrates the practical challenge behind BitGo’s AI focus. Financial assistants need permissioned access to payment tools while users retain meaningful control over balances, spending limits and account recovery. Those requirements bring custody, transaction authorization and fraud prevention into the product design from the outset.
Payments and custody form the initial focus
BitGo said its research unit will build applications where AI systems can interact with digital assets through payments, settlement and custody features. Instant settlement would allow a completed transaction to be recorded and finalized quickly on a blockchain, reducing the delay that can arise when funds must clear through several intermediaries.
Micropayments are particularly relevant to machine-driven commerce. An AI agent purchasing a small unit of data, an application programming interface call or a limited amount of computing capacity may need to make a payment that is far below the minimum economical size for many legacy payment rails. Digital assets and stablecoins could provide a mechanism for those transfers if the underlying network costs remain sufficiently low and the payment process can be controlled securely.
BitGo’s role in custody gives the company a distinct angle in that market. AI agents may be able to propose or execute transactions, but the digital-asset keys controlling funds must be protected and governed. A system intended for autonomous payments would need rules defining how much an agent can spend, which counterparties it may use and when human approval is required.
The company said BitGo Labs will assess potential projects using two criteria: whether an application can perform a function that existing systems cannot, and whether it delivers measurable benefits. That standard suggests the unit will seek to avoid AI features added mainly as marketing tools, although BitGo did not specify the metrics it will use to judge results.
Belshe sees stablecoins as machine money
BitGo Chief Executive Officer Mike Belshe has argued that stablecoins are likely to become a native currency for AI agents. Earlier this month, he said people will probably retain control over AI systems longer than some forecasts assume, a view that aligns with products built around constrained automation rather than fully independent financial agents.
Belshe’s comments responded to an argument by Arthur Hayes, chief investment officer of Maelstrom, who said stablecoins and other digital currencies cannot become AI agents’ native money because they cannot be converted directly into computing power.
The disagreement turns on the difference between a payment instrument and the resource being purchased. Hayes’s argument treats compute as the essential input for machine intelligence and raises the prospect of assets tied more directly to processing capacity. Belshe’s position leaves room for stablecoins to function as a practical settlement layer even if AI systems ultimately spend those tokens to obtain compute, data or other services.
That debate has commercial consequences for infrastructure providers. If stablecoins become a common way for agents to pay, providers of wallets, custody, compliance controls and on-chain settlement could become part of the operating stack for AI services. If AI commerce instead coalesces around specialized compute-linked assets, conventional stablecoin infrastructure may have a more limited role.
Research program extends beyond consumer products
BitGo said BitGo Labs will publish research, create protocol prototypes and reference implementations, and collaborate with outside crypto and AI partners. Reference implementations are working examples that developers can inspect and adapt when building compatible software, helping move an idea from a research concept toward usable infrastructure.
That approach could allow BitGo to test how its custody systems perform under AI-oriented payment workflows before committing to a single product architecture. It also gives the company a way to participate in technical standards that may shape how agents identify counterparties, receive spending permissions and settle transactions.
The immediate test will be whether BitGo Labs can produce applications that solve a specific payment or custody problem better than existing products. Gokubu offers an early consumer experiment, while the planned work on micropayments and settlement targets the less visible infrastructure required if AI-generated transactions become routine.
For deeper insight into AI and blockchain’s future in finance, explore our guide on how AI complements blockchain today.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
