# Into: HashiChain A self-described Layer 1 for AI agents that leans on Bittensor as its consensus engine, currently documented as a whitepaper rather than shipped as running code. // An agent Layer 1, still at whitepaper stage --- > New to Bittensor? Start here. Experienced users can skip to the full analysis. ### What is HashiChain? HashiChain (Subnet 115) describes itself as a Layer 1 blockchain built for AI agents rather than for people. According to the project's README, the goal is infrastructure where autonomous AI programs can find each other, agree on interactions, and settle value without a human in the loop. As of June 2026 that vision lives mostly in documentation: the public repository holds a README, a license, and an assets folder, with no implementation code committed yet. **The simple version:** Imagine a blockchain whose "transactions" are AI agents agreeing to work together, instead of people sending tokens around. **Centralized equivalent:** No direct equivalent. The nearest reference points are agent-coordination frameworks and general Layer 1 chains, but HashiChain's stated pitch is narrower: a chain whose ledger records intents between agents. **How it works:** (the following is the design described in the repository README, not observed behavior of a live network) - **Miners** are described as "Solver Nodes" that run distributed AI models inside Trusted Execution Environments (TEEs), simulating potential agent interactions inside secure sandboxes. - **Validators** are not given a separate scoring loop in the README. The document leans on Bittensor's Yuma Consensus to reach network-wide agreement on the validity of these probabilistic interactions, but it does not spell out a distinct validator task, so that part of the mechanism is not established in the repo. --- ### Why This Matters - **The problem it solves:** Conventional smart contracts handle deterministic, arithmetic operations. They cannot natively process the fuzzy, semantic "does agent A's intent match agent B's?" question. HashiChain's stated aim is to make that intent matching a first-class on-chain primitive, via what the README calls "Proof of Entanglement," the probabilistic verification of semantic compatibility. - **The opportunity:** If autonomous agents become meaningful economic actors, neutral infrastructure for them to coordinate and settle would matter. This is a bet on a future that has not arrived yet, so treat the upside as a thesis rather than a measured market. - **The Bittensor advantage:** Rather than bootstrapping a validator set and incentive layer from zero, HashiChain proposes to use Bittensor's Yuma Consensus as both its consensus mechanism and its computational substrate. - **Traction signals:** Limited and worth stating plainly. Per a live GitHub check, the repository's last commit was on December 8, 2025, made by a single contributor across five commits on that one day. Third-party reporting (SimplyTao) notes there is no public demo, working testnet, or independent audit. The subnet currently receives 0% of network emissions. --- ## Full Analysis **Category:** Other (agent infrastructure / Layer 1) | **Centralized Competitor:** No direct equivalent HashiChain sits in an early and speculative corner of Bittensor: subnets that propose a new base layer rather than a discrete AI service. The README frames the project around the metaphor of "Hashi" (bridge and chopsticks): a bridge connecting isolated agent silos through privacy-preserving tunnels, and chopsticks as a metaphor for coordination, the idea that a solitary agent cannot capture value alone. **Mechanism:** What the repository actually contains today is a whitepaper-style README, a license, and image assets. The design it lays out: miners run AI models inside TEEs as Solver Nodes and compute the probability of a successful match between private agent intents; the protocol then uses Yuma Consensus to agree on the validity of those probabilistic state transitions and finalize them on what the README calls the Intent Ledger. How miner output is scored and how validators specifically participate beyond the consensus framing is not detailed in the repo, so those specifics are not established here. Because there is no implementation code, demo, or testnet published, none of these mechanics can be verified as running. On-chain, the subnet is small and quiet. Price sits around 0.00647 TAO with a market cap near 15,275 TAO and roughly 5,590 TAO of depth backing the pool. Net staking flow over the trailing 7 days is slightly negative (about -7 TAO), and under Bittensor's flow-based Taoflow model a subnet with non-positive net flows receives no emissions, which matches its current 0% emission share. The price is down roughly 14% over the past month. One episode is worth recording because it shaped the subnet's public profile. On February 25, 2026, a single wallet staked roughly 12,000 TAO into Subnet 115 in one transaction. Against the subnet's thin liquidity that buy moved the alpha price by more than 1,000% within hours and briefly lifted HashiChain into the top tier of subnets by emission share, according to on-chain analysts and SimplyTao's reporting. That move has since fully unwound, and the price now trades near its 60-day low. Several experienced community members publicly flagged the event as opaque and a manipulation risk; those are third-party characterizations, recorded here for context rather than endorsed. --- ### Risk Factors - **Execution and implementation:** The repository is at the whitepaper stage. There is no implementation code, public demo, working testnet, or independent audit (per SimplyTao and confirmed by the repo contents), so the design remains unproven. - **Development activity:** Per a live GitHub check, the last public commit was December 8, 2025, from a single contributor. This may reflect a move to private development, an internal build phase, or a pause; the public repository alone does not settle which, so the gap is noted rather than interpreted. - **Deregistration:** The subnet receives 0% of emissions. With a registration dating to around December 2025, it is past the four-month immunity window, which makes it a candidate for automatic deregistration if its EMA price stays among the lowest of non-immune subnets. - **Liquidity and volatility:** The pool is thin, around 5,590 TAO of depth. The February event showed how a single large stake can swing the price by orders of magnitude on this little liquidity. The stake distribution is also concentrated, with a top-100 Gini coefficient of 0.82, so large positions can dominate pool dynamics. --- Into the next one.