Project Research Dossier: Entropy (entropy.io) - the oracle question

✅ Eligibility note: Entropy is pre-TGE with no native token, ICO or public token sale reported. Funding is a private seed round only. Research date: 28 September 2026.

Full PDF dossier attached (squad research). This post is my own angle on it: the oracle. Built from secondary sources and official Hyperliquid docs; anything I could not confirm is labelled as such.

TL;DR Entropy lets you trade perps on private-company equity, for example io:ANTH (Anthropic), on Hyperliquid via HIP-3. The interesting question is not the product. It is how you price something that has no public price. Everything else in this project depends on that answer.

Why the oracle is the whole story A perp needs a reference price to mark positions, trigger liquidations and settle funding. Anthropic is private, so there is no continuous public price, no major-exchange order book and no closing auction. Any oracle for it has to be built from indirect signals. That makes three things depend on one component: fair pricing, liquidation fairness and the deployer's own collateral. If the reference is wrong, users can be liquidated at a bad price, or the deployer can be penalized.

How pricing reportedly works (unconfirmed)** Secondary sources describe Entropy's approach as weighting external price signals by executable depth on Hyperliquid's own order book. I could not find a primary specification, so treat this as reported, not verified. If accurate, there is a circularity worth asking about: part of the reference comes from the same market the oracle is supposed to keep honest. On a thin book, a relatively small order can move the number. Hyperliquid's own docs also warn that most price indices are poor perp oracle sources, and that deployers must think through manipulation edge cases.

Guardrails that do exist (per Hyperliquid docs)**

• A HIP-3 deployer must keep 500,000 HYPE staked.
• If an oracle price moves more than 50% versus start-of-day, validators can review whether to slash the deployer.
• The stake stays slashable during the 7-day unstaking queue.
• Oracle updates are subject to rules and price clamps (see deployer-actions docs). Please check current values in the docs before relying on them. So there is real skin in the game, but it also means a badly handled oracle event lands on the deployer's own balance sheet.

Failure modes I would want answered**

• Stale inputs: what happens if a reference source stops updating during a volatile move?
• Thin books: what is the minimum depth before the weighting stops being meaningful?
• Extreme moves: how does the system behave near the 50% review threshold, and who decides?
• Fallbacks: is there a documented fallback price, and is it audited? None of these are answered in the sources I checked.

What would change my view**

• A published oracle specification and stress-test results.
• Depth, spread, volume and open-interest data per market over a meaningful period.
• Reasons and settlement details for the delisted markets io:OAI and io:IONQ.
• Named founders with public profiles. The team is described as alumni of Citadel, Optiver, Polymarket and Millennium, but I could not confirm names, so I treat that as unverified.

Other facts, with their limits**

• No token, so FDV, unlocks and vesting do not apply.
• $14M seed, reportedly led by Ribbit Capital, from CryptoRank / DropsTab. I found no primary financing documents.
• About $40M in HYPE reportedly secured. This is not operating cash, and quantity and custody are unconfirmed.

About $40M in HYPE reportedly secured. This is not operating cash, and quantity and custody are unconfirmed. More importantly, without primary documentation, it is difficult to determine whether this represents treasury assets, deployer collateral, or another form of exposure. I would therefore treat the figure as a reported data point rather than evidence of financial strength.