Bitwise CIO Matt Hougan Calls NEAR the Blockchain for AI After a 198% Quarter

The Reasoning Before the 198%

The argument comes first and the figure trails it. Matt Hougan, chief investment officer at Bitwise, set out in a memo posted to Bitwise's CIO memos page on October 7 why he considers NEAR Protocol (NEAR) the blockchain built for artificial intelligence: a network designed to let software agents transact with each other rather than with people. Hougan names NEAR as one of the projects he is watching closely, and his premise rests on how the quarter actually traded. Bitcoin (BTC) gained 43% in the third quarter and Ethereum (ETH) 71%, while NEAR climbed 198%, a lead over the wider altcoin field that Hougan reads as investors repricing a specific use case rather than riding general momentum. That use case is machine-to-machine settlement. The chain's engineering follows from it: processing capacity sized for machine-frequency activity and an AI stack that computes on encrypted data. Throughput alone was never the pitch; the pitch is that agents, not human traders, become the marginal counterparty. Hougan is explicit about the gap between design and reality, conceding that the network's users today are still human. Where the design shows up in numbers is NEAR Protocol (NEAR) Intents, the execution layer that simplifies trades across chains, which has cleared more than $30 billion in cumulative volume, with fee revenue the network expects to reach roughly $45 million this year. An agent running an AI crypto wallet needs a chain that settles without human sign-off, and that is the demand NEAR's architecture was drawn to meet. What the memo leaves open, and what the NEAR Protocol community is now testing, is whether a demonstrably revenue-generating business can pull AI-focused capital into the asset.

From “Attention Is All You Need” to NRR

The lineage behind that argument is the part most coverage skips. NEAR was co-founded by Illia Polosukhin, one of the authors of the 2017 paper “Attention Is All You Need”, the work that became the foundation for generative AI, and the network launched in 2020 as one of the more advanced designs of its cycle: dynamic sharding that parallelizes transaction processing, plus signature schemes built to survive quantum computers. Sharding, briefly, splits the chain into segments that process transactions at the same time, which is how the network chases the throughput agents would need. The commercial layer sits above the engineering. NEAR Protocol (NEAR) Intents routes a trade to wherever execution is best across connected chains, so a machine quoting both sides of a book functions as a market maker without the chain becoming the bottleneck. Bitwise also wrapped the thesis in a tradable product: the firm's first US NEAR spot ETF, ticker NRR, listed on NYSE Arca on September 29, a debut we covered as the NEAR ETF debut on NYSE Arca. That makes Bitwise both the commentator and the issuer, a dual role worth keeping in mind when weighing the memo. The same month exposed the operational side. On October 1 a bug in NEAR Intents drained roughly $3.8 million, and the team pledged to make users whole; our follow-up confirmed the funds returned in full. One bug, one guarantee and one ETF listing inside a single month is the environment the memo lands in, a network scoring a 198% quarter while its newest product path was still being stress-tested.

The Human Gap in the AI Thesis

Our read is that the memo works as a positioning document as much as a valuation case, and its weakest joint is the one Hougan names himself: the users are still human. The $30 billion of Intents volume and the $45 million fee estimate describe a machine-ready system waiting for machine customers, so the thesis holds only if agent activity arrives at scale, on a timetable the memo does not commit to. The October 1 bug is a reminder that the plumbing meant to serve those agents was still being debugged. The market has started pricing the distance between story and usage: our live monitoring shows the NEAR price down 15.9% over the past 24 hours. Earlier in the week, coverage of the memo sent the asset toward $5.53 resistance. Whether AI-oriented capital treats the 198% quarter as an entry point, or as proof the repricing already happened, decides what the memo was worth.