Key Notes
- Bitwise’s Hougan sees NEAR’s existing trading business as a foundation for a larger opportunity in AI-agent commerce.
- NEAR Intents has processed over $30 billion in cumulative volume, while Hougan projects about $45 million in 2026 fees.
- NEAR Intents uses competing solvers, while NEAR AI Cloud runs models in protected hardware with verifiable execution.
Bitwise Chief Investment Officer Matt Hougan sees NEAR’s existing cross-chain trading business as a foundation for its AI ambitions. His October 7 memo argues the investment case extends beyond a future dominated by autonomous agents.
Hougan’s assessment centers on NEAR Intents and confidential AI infrastructure, while acknowledging that autonomous finance remains at an early stage.
An AI Background Behind the Blockchain
NEAR’s connection to artificial intelligence begins with co-founder Illia Polosukhin. He was one of eight authors of the 2017 Transformer paper, “Attention Is All You Need,” which introduced an attention-based neural-network architecture that became foundational to modern language models.
Polosukhin and Alex Skidanov co-founded NEAR in 2018, according to Bitwise’s accompanying research guide. The project now combines blockchain settlement, cross-chain execution and confidential computing. These address different requirements for software that could eventually manage assets, purchase services and transact on a user’s behalf.
The Existing Business Is Cross-Chain Trading
Bitwise’s October 7 guide puts NEAR Intents’ cumulative transaction volume above $30 billion across more than 35 blockchains. That measures the value routed through the execution system over time. It is neither revenue earned by the protocol nor money deposited into NEAR investment products.
Hougan projects roughly $45 million in Intents fees for 2026. That is an annual estimate, rather than a completed-year result or cash distributed to token holders. He also notes that human users currently dominate Intents activity, making today’s adoption distinct from its AI-agent ambitions.
The distinction between volume and fees matters when assessing the business. A routing system can handle a large value of transactions while retaining a much smaller amount as revenue. Its economics depend on the charges collected, the portion retained and how that revenue is used.
Bitwise says a fee mechanism activated in February directs NEAR’s share of Intents fees into open-market token buybacks. The relevant amount is the protocol’s retained share, rather than every dollar paid across the ecosystem. Purchases create a connection between usage and token demand, but do not establish a guaranteed return or a fixed cash payment to holders.
What Intents Actually Does
NEAR’s documentation describes a system in which users specify an outcome and third parties compete to fulfill it. Instead of manually arranging each step of a cross-chain trade, a user or AI agent can request an exchange from one asset into another.
Market makers known as solvers compete offchain to provide a suitable route and price. The selected quote goes back to the user or agent for approval. Once accepted, a verifier smart contract on NEAR checks and settles the transaction. The interface simplifies the process, while the underlying execution still relies on market makers and settlement infrastructure.
The project’s contract repository illustrates the settlement logic with two counterparties exchanging USDT and USDC. Each signs the terms, and the verifier evaluates whether both instructions can be satisfied together. This atomic settlement means the exchange completes as one operation, avoiding a situation in which one party pays while the other fails to deliver inside that contract.
That model is useful to human traders today. Its potential relevance to agents is the same ability to express a desired result without hard-coding every intermediate transaction into the application.
Confidential AI Addresses a Different Problem
The computing side concerns what happens to information supplied to an AI model. NEAR AI introduced its Cloud and Private Chat products in December 2025, using trusted execution environments built on Intel TDX and NVIDIA confidential-computing hardware.
In the design described by NEAR AI, a prompt is encrypted before entering an isolated computing environment. The model processes it inside that environment and returns an encrypted response. A hardware-backed attestation lets the customer check that the expected code ran inside genuine protected hardware.
NEAR AI says this prevents the infrastructure operator, including NEAR AI itself, from reading the protected data. The aim is to let businesses use sensitive information for inference without exposing it to the cloud host. The protection concerns data processed in that environment and depends on the hardware and verification process described by the provider.
Confidential inference and blockchain settlement are separate parts of the system. Running a model in protected hardware does not mean its computation takes place inside every NEAR blockchain transaction, and private inference alone does not make an agent’s subsequent payments or trades confidential.
Adoption Must Connect the Products
A recent Robinhood Chain integration provides an example of Intents reaching another ecosystem. CoinScreamer reported that ETH, USDG, PONS and CASHCAT were the initial assets supported through near.com, with cross-chain swaps, deposits and withdrawals among the announced functions.
The investment thesis ultimately depends on how these services develop together. Trading activity, retained fees and buyback spending test the existing execution business. Sustained use of confidential computing and agent-initiated transactions would test the AI opportunity. A working product in the first category is evidence of present demand; it does not settle how large the second category will become.
Disclaimer: CoinScreamer is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and market insights on digital assets and related technologies. NuvexMedia LLC invests in and collaborates with companies across the digital asset, blockchain, and technology sectors. These relationships do not influence CoinScreamer’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2025 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
