Morpheus

What it does

Morpheus is a decentralised marketplace for AI inferenceInferenceRunning a trained AI model to produce an answer. Inference is what happens when you type a prompt into ChatGPT and get a response. The model takes your input, computes a best guess, and returns it.Like asking an expert for their opinion. The training was the decades they spent becoming an expert. The inference is the 30 seconds it takes them to answer your specific question.Inference in the glossary → and agents. Not a provider. It doesn’t run GPUs or host models. It connects consumers to independent providers who do, handles routing and payment via smart contracts, and lets anyone participate on either side. No company in the middle. No terms of service dictating what your agent can do.

Four participant groups make up the network. Compute providers supply GPUGPUGraphics Processing Unit. Originally designed to render video game graphics, GPUs turned out to be exceptionally good at the massively parallel math that AI models need. Modern AI training and inference runs almost entirely on GPUs.Like a factory with 10,000 workers doing the same simple task in parallel, versus a CPU which is more like 10 workers each doing different complex tasks. AI training involves doing simple math a million times per second on a million numbers, which is exactly what the GPU factory is designed for.GPU in the glossary → and CPU resources. Code contributors build and maintain the software. Capital providers stake assets to fund protocol development. Community builders operate builder subnets that fund Smart Agents and end-user applications. Each group earns MOR tokens proportional to their contribution.

The compute marketplace is the critical piece. It matches inference requests with available compute providers, handles P2P routing (encrypted, per Morpheus’s docs) and payment, and maintains quality standards through a reputation system. MOR transitions from a tokenTokenA digital unit of value or access rights tracked on a blockchain. Tokens can represent ownership in a project, a right to use a service, a share of future revenue, or simply a tradable asset with no underlying claim.Like a physical poker chip a casino issues. The chip itself has no value. What makes it worth something is what it lets you do at the casino, what the casino has promised, and how much other people will pay you for it.Token in the glossary → project to functional infrastructure here. Privacy level depends on the provider. v6.0.0 (March 2026) added Phase 1, in which consumers cryptographically verify the provider’s proxy-router binary inside Intel TDX; our own refresher runs that check against every TEETEETrusted Execution Environment. A hardware-secured region of a CPU or GPU where code runs in isolation, so even the machine's operator can't read what's happening inside. TEEs give decentralised AI inference privacy guarantees.Like a bank vault inside a bank. The bank owns the building, staffs the lobby, and runs the security cameras. But what's inside the vault is invisible to everyone, including the bank staff, unless the customer opens it.TEE in the glossary →-tagged model on chain (5 at the last read), and the TDX quote binds to the key in the provider’s TLS certificate. v7.0.0 (April 2026) added Phase 2, shipped as open-source code in the proxy-router (internal/attestation/backend_verifier.go and nras_verifier.go): the provider’s proxy-router attests its own backend LLM on every prompt, binding CPU TDX to NVIDIA NRAS GPU evidence. Consumers never see Phase 2 directly, but the provider node publishes its verdict per model on a public health endpoint, which we read: 2 of the TEE models were passing backend attestation and 0 were flagged unverified at the last read; a model whose backend is down, rather than failing attestation, counts in neither. That verdict is the provider’s own report, produced by the published TEE image whose identity our Phase 1 check proves.

We then tested it from our own consumer node on 24 September 2026. A session on the model passing Phase 2 opened, the node logged the provider’s attestation as passed, and it re-checked the quote on each prompt. A session on a model the provider had flagged was refused by the provider before any MOR was staked, so a backend that fails attestation can’t be used.

One practical catch: consumers on proxy-router versions before v7.9.0 (August 2026) can’t open TEE sessions at the moment. SecretAI’s quote-verification service now returns a field in a format those versions reject, so their Phase 1 check fails closed. We hit this on our own node, which ran v7.2.0, and upgrading to v7.11.0 cleared it.

The product in practice

Grounded in our on-chain inference index, hands-on time with the consumer node, and our Morpheus deep-dives.

The USP. Morpheus is the rare consumer AI whose privacy claim you can check yourself. Connect to a TEE-tagged provider and, by Morpheus’s account, each prompt triggers a roughly 50ms hardware check: a fresh Intel TDX and NVIDIA certificate proving the model runs in a sealed enclave with logging sealed off, and if it fails the session never opens. The host-level half of that (Phase 1) is the part we check ourselves, on every TEE-tagged model; the backend half (Phase 2) is reported per model by the provider’s node. That sits above Venice’s policy-based privacy, and it’s the whole reason to pick Morpheus over a slicker app.

Who it’s for. Consumers who want private inference they can actually check, developers building censorship-resistant apps, and builders of sovereign AI agents, which are central to what Morpheus set out to enable. Agents are more vision than shipping product today: the crypto-native case, a wallet-holding agent acting across DeFi with no platform in the middle, is live but early, with personal assistants the longer-term aim. What ships now is the verifiable-private inference those agents would run on. Compute providers supply the inference; capital providers and subnet stakers fund the economics; the customer is whoever, or whatever, sends the prompt.

Using it. The Morpheus core is infrastructure: the on-chain marketplace, the node protocol, the attestation. Everything you touch to consume inference sits on top, including the api.mor.org gateway, the self-hosted Uplink gateway in front of your own node, and consumer apps like NodeNeo, Mordiem and DRM3’s Pistachio (which agents run on which route). The easy route is straightforward: stake MOR to the Marketplace API subnet for a standing inference quota, or top up USD credits by card with no wallet at all. The full hardware-attested guarantee needs the self-hosted desktop node connecting to a TEE provider, a heavier setup, and any MOR you stake for access is returnable collateral rather than a per-use spend.

Demand, honestly. Every session is on-chain, so unlike self-reported networks you can audit exactly what’s flowing, and we track it daily. Throughput, meaning tokens processed, built through the first half of 2026 to a peak in July, fell about 38% in August (a figure our index and the project’s own tech.mor.org dashboard agreed on to the token), then edged up about 8% in September despite the shorter month, leaving it about a third below July.

Sessions rose by about a third on August, so the average session got smaller again. Underneath the total, the direct side hit a record: sessions opened outside the API Gateway (self-custody consumer nodes, OpenClaw and Node Neo) processed about 7.25B tokens in September, up from 4.25B in August and the highest month yet on tech.mor.org’s native segment, while gateway volume fell. The traffic is mostly coding work on open-weight models, still led by DeepSeek V4 Flash, and supply stays thin: 27 providers closed sessions over the last 30 days and the top three handled 66.1% of them, with the protocol’s revenue still capital-provider yield rather than compute fees. For most of the year the work sat with the core team: an address registered under provider.mor.org was the busiest provider every month from February to August, and supply only spread from July (who served the tokens, month by month).

The vision. As providers deploy the v7.0 TEE image, verifiable-private inference shifts from niche to default, and Morpheus becomes the permissionless base layer for private AI that no company can switch off, with the consumer apps riding on top competing on experience. Provider supply is the gate; the technology already ships.

How Morpheus works

No company runs Morpheus. No CEO, no office, no employees. Four participant types each earn MOR tokens proportional to their contribution, with smart contracts coordinating everything. Daily emissions split into five buckets: 24% each to the four participant groups, plus 4% to a community-governed Protection Fund.

Capital providers are the economic engine. You stake stETH (or since September 2025, USDC, USDT or WBTC via Aave integration) into the Morpheus smart contracts on Ethereum. The yield generated by your stETH goes to the protocol, and in return you receive your proportional share of the 24% capital provider emission allocation, distributed on Arbitrum. This is the primary participation method today and the lowest-friction entry point.

Effectively all of that yield purchases MOR on Arbitrum. Half of what it buys is burned and half goes to a 16-year lock. MRC43 specifies a 75/25 buy-and-pair split instead, but the last MOR reached the protocol-owned liquidityProtocol-Owned LiquidityLiquidity that a protocol owns directly instead of renting from outside providers. The protocol funds and holds its own AMM position, so the trading depth is permanent and can't be pulled when farming rewards dry up.Like a marketplace that owns the building its traders work in, rather than renting stalls week to week. Renters pack up the moment a better deal appears elsewhere. An owner stays put, so the floor never empties out from under the people trying to buy and sell.Protocol-Owned Liquidity in the glossary → position on 3 January 2025.

Compute providers run the Lumerin Node software and serve AI inference requests on the Base network. When a user sends a promptPromptThe text you give an AI model to tell it what to generate. A prompt can be a simple question, a long instruction, a chunk of context plus a task, or a conversation history the model uses to produce its response.Like a brief you give to a junior designer. A vague brief gets a vague result. A detailed brief with context, constraints, and examples gets something usable. The quality of the output depends heavily on the quality of the brief.Prompt in the glossary →, the Smart Agent Router matches the request to available compute, handles payment, and maintains quality through a reputation system. Compute providers earn from the 24% Compute pillar based on session activity (bid price multiplied by session length), paid from a protocol funding account, not from burning the user’s MOR. The staked-access path returns the user’s MOR in full; only the direct-pay route spends it. How Morpheus Pays for Inference traces the full flow.

Code contributors build and maintain the Morpheus software: smart agents, core protocol, tooling. Contributions are assessed and weighted for the 24% code allocation. This is a meritocratic modelModelA trained neural network that takes inputs (text, images, audio) and produces outputs (more text, classifications, generated content). In DeAI the model is the thing that actually does the work.Like a very experienced apprentice who has spent years watching thousands of masters make furniture. They can't explain how they know when a joint is right, but they can make a chair that looks and functions like a Chippendale. The training is invisible. The output is what matters.Model in the glossary → where shipping code earns tokens.

Community builders operate builder subnets: on-chain pools on Base that attract staked MOR and fund Smart Agents and end-user applications. The community bucket is 24% of emissionsEmissionsNew tokens created and distributed by a blockchain protocol over time as rewards to validators, stakers, or miners. Emissions fund network security and participation at the cost of diluting existing holders.Like a company that pays employees partly in newly printed shares. Every year the total number of shares goes up, which means existing shareholders own a slightly smaller slice of the same company unless the company grows faster than the printing.Emissions in the glossary →, paid to whichever subnet a builder operates in proportion to the MOR staked to it. The protocol pays the operator; what reaches stakers is set off-chain. See How Morpheus builder subnets work.

Protection Fund receives the remaining 4% of daily emissions. Community-governed reserve for compensating users affected by smart contractSmart ContractA program stored on a blockchain that runs automatically when its conditions are met. Smart contracts are how blockchains do anything beyond just transferring tokens — DeFi, NFTs, DAOs, and DeAI infrastructure all run on smart contracts.Like a vending machine. You put in the right input and it produces the expected output, no human operator required. The rules are fixed in the machine itself, anyone can use it, and nobody can stop a transaction in the middle.Smart Contract in the glossary → bugs or failures. Effectively self-insurance for the protocol.

The critical mechanism is the Power Factor (MRC42). Participants can lock their earned MOR rewards in exchange for a multiplier capped at approximately 10.7x. Locked MOR cannot be withdrawn early and lock periods can only be increased, never decreased. The multiplier is a curve anchored to a fixed 2024 to 2040 window rather than a function of duration, so the same lock is worth less the later you start it: a six-year lock reached the cap at launch and reaches about 8.4x today. On chain, 257 of 6,453 capital stakers hold a live lock, and the pool’s blended multiplier is 1.89 times. See the full tokenomics analysis for worked examples and detailed numbers.

It’s the compute marketplace where Morpheus transitions from a token project to functional infrastructure. Users stake MOR, into a session or an access subnet, to gain a daily compute ration. The more MOR staked, the more compute accessible. The system works today, though the provider base is still small relative to the long-term demand target.

Live inference activity

OYM indexes every SessionClosed event on the Morpheus Inference Contract on Base, plus the daily MOR locked into the Morpheus Marketplace API builder subnet for IPS-quota access. Two demand signals, refreshed daily.

OYM Live Data

98.2K

Sessions (last 60 days)

Morpheus Live Inference Activity

On-chain MOR settled per day, provider-attested tokens served, and stake-for-access subnet flows. 44 unique providers all-time. Indexed from SessionClosed events on Base.

Open the live tracker → Updated 2026-10-23

For a second lens on the same contracts, the snapshot below reads the network live from MorScan, a signed third-party explorer that publishes an Ed25519 provenance receipt with every figure. It earns MOR emissions as a builder subnet, so treat it as verifiable, not neutral, and cross-check anything load-bearing against BaseScan.

53

Compute providers

15

Active sessions

627

Live model bids

$0.83

MOR / USD (Base DEX)

MorScan reports, indexed from the Morpheus contracts on Base, as of 20 Oct 2026. Third-party signed source, not independent of Morpheus.

Verify the signature →

Value proposition

Straightforward pitch. Centralised AI assistants (Siri, Alexa, ChatGPT) run on someone else’s servers, follow someone else’s rules and harvest your data. A Morpheus agent runs on infrastructure you choose, uses models you select and follows instructions only you set.

Right now the practical application is crypto-native: an agent that can interact with DeFiDeFiDecentralised Finance. Financial services like lending, trading, and yield farming built on smart contracts instead of traditional banks or brokerages. DeFi protocols are usually permissionless and global.Like a vending machine that can give you a loan, swap your currencies, or invest your savings. Nobody is behind the counter, the rules are written into the machine itself, and anyone with money in the right format can use it.DeFi in the glossary → protocols, manage wallets, execute trades and handle multi-step operations across chains without trusting a centralised intermediary. It sees your full portfolio, understands your risk parametersParametersThe internal numbers (weights and biases) inside a neural network that get adjusted during training. A 70-billion-parameter model has 70 billion adjustable internal numbers encoding everything it has learned.Like the synapses in a human brain. Each parameter is a tiny dial that gets nudged a little during training. With enough dials, the network can represent surprisingly complex patterns. The total parameter count is roughly how much "brain" the model has.Parameters in the glossary → and executes according to your strategy. No platform in the middle deciding what your agent is allowed to do.

Longer term, the vision extends beyond crypto to personal agents handling scheduling, research, communication and decision support, all on sovereign infrastructure. Whether that materialises depends on execution over the next 12-18 months. I think the compute marketplace being live and functional already puts Morpheus ahead of most projects still shipping decks.

Tokenomics

Morpheus tokenomics come down to stake: the MOR token is needed as collateral to open inference sessions on the compute marketplace, returned in full when the session closes, and as stake in the API gateway’s builder subnet for usage credits, so inference itself burns no MOR. Value reaches holders through one live route, and it’s funded by yield on capital-provider deposits rather than inference revenue: that yield buys MOR, about 8.3% of daily emissions over the last 30 days, and half of each purchase is burned (309,313 MOR so far) while half is locked for the post-2040 tail. The burn is a multisig transfer under MRC43 rather than a contract rule, the liquidity-pairing leg stopped in January 2025, and the tail re-emission exists only in the whitepaper. Dilution is the 16-year emission schedule, 11,961 MOR a day now, plus 3.4 million MOR emitted into protocol Safes but not yet distributed (how MOR actually works).

MOR is a fair-launch token. No pre-mine. No VCVCVenture Capital. Private investors who fund projects at an early stage in exchange for equity or token allocations. VC rounds are typically pre-launch, at steep discounts to any future public price, with multi-year vesting.Like angel investors in a startup who buy shares before the company goes public. They take more risk because the company might fail, so they get a better price. Once the company IPOs they can sell, and the public market pays whatever price it thinks is fair.VC in the glossary → allocation. No insider distribution. Every MOR in existence was earned through contribution to the network. This matters because it means no insider vesting schedule waiting to dump on participants.

Emissions follow a linear decline. Day one, on 8 February 2024, minted 14,400 MOR, and each day emits roughly 2.47 MOR less than the one before, reaching zero around day 5,833 (early 2040). The split across the four contributor groups and the 4% Protection Fund is the donut shown earlier in How Morpheus works.

After Epoch 1: the post-2040 tail

Day 5,833 ends Epoch 1, and emissions don’t stop there. The whitepaper specifies a tail emission: 50% of the MOR burned during the previous 5,833 days is re-emitted over the next 5,833 days, capped at 16% of the MOR then in circulation. That calculation repeats each epoch, and because every tail is a fraction of what was burned, the series converges toward roughly 1% of circulating supply a year rather than zero.

The funding side is already running, and it’s measurable. MRC43, marked Implemented, routes protocol-owned-liquidity MOR to a multisig, burns half of it permanently and locks the other half for 16 years to seed the second epoch. Both halves are visible on Arbitrum, and they track each other exactly. Read on 16 August 2026, the burn address held 287,795 MOR across 60 transfers and the lock contract 0xb1972e86… held the same 287,795 MOR across 60 transfers, with nothing ever withdrawn. Both keep growing with each liquidity event, so Epoch 2 is sized by how much MOR gets burned before 2040 rather than by anything fixed today.

A second and smaller mechanism sits alongside it. MRC22, still In Progress, would bank the Community 24% emissions that go undistributed, since those rewards pay pro rata to MOR staked against circulating supply and Protection Fund and POL holdings can never be staked. Those banked tokens would restart the same method in Epoch 2. That is a proposal about one pillar, not the network-wide tail.

The restart itself is the missing piece. Query the schedule on the Lumerin Diamond for any future date and it flattens after February 2040, with the compute pillar totalling 10.08M MOR against 31.92M for everything else. That split then moves with usage rather than time, because getComputeBalance subtracts every MOR ever claimed by providers and totalMORSupply adds the same amount back. Re-emission needs a governance action and a contract that doesn’t exist yet, so read the tail as funded intent with an on-chain balance behind it, not as shipped code.

Since September 2025, Capital V2 accepts USDC, USDT and WBTC via Aave integration on top of stETH, with Chainlink oracles normalising yields across asset types.

The burnBurnPermanently removing tokens from circulation by sending them to an address that no one controls. Burns reduce total supply, which (all else equal) makes each remaining token worth more of the network's value.Like a company buying back its own shares and shredding them. The company's total value stays the same, but each remaining share now represents a slightly bigger slice of that value.Burn in the glossary → runs through MRC43: the multisig that receives the bought MOR sends half to the burn address and locks half for the tail, by transfer rather than by contract rule, and the liquidity-pairing leg hasn’t run since January 2025. See the full tokenomics analysis for worked examples and detailed numbers.

How to participate

Provide capital. Stake stETH, USDC, USDT or WBTC through the Morpheus smart contracts and earn daily MOR emissions. This is the lowest-friction entry point. You need a supported asset and a walletWalletSoftware that stores the private keys needed to control tokens on a blockchain. A wallet does not actually hold any tokens. The tokens live on the chain. The wallet holds the keys that prove you own them.Like the key to a safe deposit box. The key doesn't contain your valuables. The valuables sit in the bank's vault. The key is what proves you're allowed to open the box and take them.Wallet in the glossary →. The 7-day lock on deposits is the only constraint. After that, your capital is freely withdrawable (though you stop earning MOR). Consider locking earned MOR via the Power Factor if you have long-term conviction, noting the multiplier is lower the later you start.

Provide compute. Run a Morpheus compute node and serve inference requests on the Base network. Requires GPU hardware and technical setup. Earnings depend on the volume of inference requests routed to your node.

Contribute code. Submit pull requests to the Morpheus GitHub repositories. Code contributions are assessed and weighted for emissions allocation. This is for developers.

Build community. Create content, documentation, tutorials or tooling. Community contributions are assessed by the community builder group.

Staking MOR: step-by-step guide

Capital provision is the most accessible way to participate. Here is the practical walkthrough.

What you need: stETH (stake ETH via Lido first if you hold ETH), a wallet like MetaMask, and ETH on Ethereum mainnet for gasGasThe fee paid to a blockchain to process a transaction. Gas is denominated in the chain's native token and varies with network demand. Sending a transaction without enough gas means the transaction fails and the gas is still consumed.Like the petrol that powers a car. You need to put petrol in to make the engine run. The amount of petrol you need depends on how far you're driving and how much you're carrying. If you run out, the car stops.Gas in the glossary →. Since September 2025, Morpheus also accepts USDC, USDT and WBTC via Aave integration with Chainlink oracles normalising yields across asset types.

Step 1: Get stETH. If you hold ETH, stake it through Lido at stake.lido.fi to receive stETH. If you do not hold ETH, buy stETH directly on a DEXDEXDecentralised Exchange. A trading venue where token swaps happen entirely through smart contracts, with no central operator holding user funds. The largest DEXes are Uniswap, Aerodrome, Raydium, PancakeSwap, and Curve.Like a self-service vending machine that lets you swap one type of coin for another. The machine sets the exchange rate based on its current stock, anyone can deposit coins to refill it, and there's no clerk behind the counter.DEX in the glossary →. stETH earns approximately 3-4% APY from Ethereum validatorValidatorA computer that runs the full blockchain protocol, verifies transactions, and proposes new blocks. Validators are the workers that keep a Proof of Stake network running, and they earn rewards for doing the work correctly.Like a notary public who witnesses and stamps legal documents. Validators witness transactions, check they follow the rules, and stamp them into the permanent record. A notary who commits fraud loses their license. Validators work the same way, except the license is staked tokens that get slashed on misbehaviour.Validator in the glossary → rewards. This yield is what funds Morpheus.

Step 2: Connect to Morpheus. Go to the Morpheus stakingStakingLocking up a cryptocurrency to help secure a blockchain network, usually in exchange for rewards. The locked tokens act as a security deposit that can be taken away if the staker misbehaves.Like putting down a large rental deposit for an apartment. You get the money back if you behave, you earn interest while it's locked, and the landlord takes it if you trash the place.Staking in the glossary → interface. Connect your wallet. You are interacting with smart contracts on Ethereum mainnet.

Step 3: Deposit stETH. Approve the contract to spend your stETH, then deposit. There is a 7-day lock on deposits. After that, your stETH is freely withdrawable at any time (though you stop earning MOR when you withdraw). Your stETH yield goes to the protocol; you earn daily MOR emissions proportional to your share of the capital pool.

Step 4: Consider the Power Factor. Once you earn MOR, you can lock it for a multiplier on your emissions. The contract sets no maximum duration; the multiplier caps at 10.7x and stops growing at the January 2040 anchor.

Because that anchor is fixed, a six-year lock that reached the cap at launch reaches about 8.4x today, and the cap now takes about 10 years. This is irreversible: locked MOR cannot be withdrawn early and the period can only be extended. Only do this if your horizon genuinely matches it.

Step 5: Claim MOR. Earned MOR is distributed on Arbitrum and can be claimed at any time with no vesting or lock (unless you choose the Power Factor). BridgeBridgeA protocol that lets you move assets from one blockchain to another. Bridges typically lock the asset on the source chain and mint a wrapped version on the destination chain. Bridges are notoriously the most-attacked component in crypto.Like a coat check at a club. You hand over your coat, get a numbered ticket, and the club promises to return the coat when you bring back the ticket. The trust assumption is that the coat check doesn't lose your coat or run away with it.Bridge in the glossary → to your preferred chain or hold on Arbitrum.

The honest economics: Your stETH continues earning Lido’s ~3-4% APY for the Morpheus protocol, and you receive MOR in return. The value of that MOR depends entirely on the token price. With daily DEX volume dangerously thin, the MOR you earn has limited liquidityLiquidityHow easily a token can be bought or sold without moving the price. High liquidity means you can enter or exit large positions quickly at the quoted price. Low liquidity means even small trades can swing the market.Like the difference between selling a house and selling a share of Apple stock. The house might be worth more on paper, but finding a buyer at that price takes weeks. The Apple share converts to cash in one click.Liquidity in the glossary →. This is an early-stage infrastructure bet with a multi-year time horizon. I sized my position accordingly, meaningful enough to be invested in the outcome, not so large that thin liquidity would be a problem on exit.

Honest assessment

What works

The fair launchFair LaunchA token launch where everyone has the same access from day one. No private sale, no insider allocation, no VC discount. Tokens are distributed by mining, staking, or open public sale at a single price.Like a 100m sprint where everyone starts behind the same line at the same time. Some runners are faster, but nobody gets to start 10 metres ahead because they paid extra. The race is decided by the run, not by who bought the best position.Fair Launch in the glossary → is rare in this space. Most “DeAIDeAIDecentralised AI. An umbrella term for blockchain-based projects that build AI infrastructure (compute, data, inference, models, agents) without a single central provider controlling the system.Like the difference between streaming a movie from Netflix and sharing it via BitTorrent. Netflix is fast and polished but one company controls what you can watch and what you pay. BitTorrent is messier but no single operator can shut you out.DeAI in the glossary →” projects have significant insider allocations that create misaligned incentives. Morpheus doesn’t. Code ships consistently. The compute marketplace is functional, not vapourware. Community engagement is active and growing.

What doesn’t work yet

Agent capabilities are early-stage. Current smart agents can handle basic crypto operations but aren’t yet competitive with centralised alternatives for general tasks. It needs significantly more compute providers to achieve reliable latency and availability.

The risk

Morpheus is betting on the agent paradigm becoming the primary way people interact with AI. If the market decides that agents are a feature of existing platforms rather than an independent infrastructure layer, the thesis weakens. Your stETH capital is withdrawable after 7 days, but if you have locked MOR via the Power Factor, those rewards are committed for the full lock duration.

My position

I am a capital provider with stETH staked in the Morpheus contracts. I have also staked MOR into the Morpheus Marketplace API builder subnet at app.mor.org so I can use the inference API directly via the daily-credits flow. So I’m exposed on both sides: yield-redirect via stETH on the supply side, and stake-for-access via the subnet on the demand side. I believe in the thesis and sized the positions for a multi-year commitment. This is early-stage infrastructure, not a blue chip.

Freedom Score: 76/100

Morpheus scores 76/100 (B grade). Full methodology at Freedom Score Methodology.

Infrastructure decentralisation (15/20): Compute marketplace with independent providers running physical hardware. Permissionless participation. No geographic restrictions. But the network is early-stage and provider count is still small relative to the demand target. Critical smart contract infrastructure runs on Ethereum/Arbitrum/Base, inheriting those chains’ decentralisation properties.

Governance decentralisation (10/20): Four-group emission model distributes influence across compute, code, capital and community contributors. But one 5-of-9 Gnosis Safe holds owner and upgrade authority over the core deposit, staking and inference contracts, with no timelock, and the core development team retains significant influence over protocol direction. The morpheusai.eth Snapshot space no longer loads, so there is no Snapshot voting venue to observe.

Token distribution fairness (13/15): Clean fair launch. No pre-mine, no ICOICOInitial Coin Offering. A token sale where a project sells tokens directly to the public, usually before any product exists. ICOs dominated 2017-2018 funding and are now mostly replaced by airdrops, IDOs, or fair launches.Like a company selling shares to the public before going public, except with no SEC oversight, no audited financials, and often no product at all. The 2017 ICO boom showed why those guardrails exist in traditional finance.ICO in the glossary →, no VC allocation, no insider distribution. Every MOR earned through contribution. This is the cleanest distribution in the DeAI space. The Power Factor mechanic further rewards long-term commitment, with participants voluntarily locking claims for multiplied emissions.

Censorship resistance (12/15): No content policy on inference requests. Permissionless compute providers serve any workload. No KYC requirement. But the network’s small scale means fewer alternative providers if any individual node refuses a request.

Data sovereignty (14/15): Local agent execution. Self-custodial wallets. No platform surveillance or data collection. Agent interactions with DeFi protocols don’t route through a centralised intermediary. Morpheus documents P2P encrypted connections between consumer and provider. v6.0.0 (March 2026) shipped Phase 1: consumers cryptographically verify the provider’s proxy-router binary inside Intel TDX, which our refresher re-checks on every TEE-tagged model on chain. Phase 2 shipped in v7.0.0 (April 2026, in partnership with Secret Labs / SecretVM) as open-source code in the proxy-router: the provider’s proxy-router attests its own backend LLM on every prompt, combining CPU TDX, NVIDIA NRAS GPU attestation, anti-replay CPU-GPU nonce binding, TLS certificate pinning, and RTMR3 workload replay (proves loaded models match what the operator declared). Logging inside the enclave is locked in production mode and can’t be increased to capture prompts. v6+ consumers automatically inherit Phase 2 guarantees when they route to v7+ providers, per Morpheus’s docs. Phase 2 runs inside the provider; its node reports 2 TEE model passing and 0 flagged unverified at the last read, and privacy on models without a TEE tag rests on the provider’s conduct. Remaining gaps: NVIDIA-only GPU attestation (AMD SEV-SNP host attestation joined in v7.2.0, 2 June 2026, but per the v7.15.0 release notes it failed open until 14 October 2026: an empty measurement was skipped, so any SEV host passed Phase 1, and consumer nodes older than v7.15.0 still accept one), and the apiAPIApplication Programming Interface. A structured way for one piece of software to talk to another. In DeAI, APIs let applications request inference from a model without running the model themselves.Like a waiter in a restaurant. You don't walk into the kitchen and cook your own meal. You tell the waiter what you want, they tell the kitchen, the kitchen cooks it, and the waiter brings it back. The API is the waiter.API in the glossary →.mor.org gateway, whose own node verifies the provider’s attestation but whose hop is private only by policy, per Morpheus’s TEE page. On-chain transactions are publicly visible (standard blockchain transparency).

Open source and transparency (12/15): Core code public and auditable, though 21 of 50 non-fork public repositories carry no licence file (GitHub API, 18 September 2026). Active GitHub development. Transparent emission mechanics verifiable on-chain. Code4rena completed a formal security audit in August 2025 covering 6 core contracts (ChainLinkDataConsumer, DepositPool, Distributor, RewardPool, L1SenderV2, L2TokenReceiverV2), finding 4 medium-severity issues around stETH rounding and Aave pool migration. The audit covers capital and distribution contracts, but compute marketplace contracts on Base and the agent execution layer remain unaudited. Bug bounty programme scope could be broader.

Path to improvement

Three changes would materially increase Morpheus’s score:

1. Transition smart contract controls to community governance. The bootstrap-phase multisig controls are the primary freedom score deduction. Publishing a concrete timeline for transitioning key contract controls to on-chain governanceDAODecentralised Autonomous Organisation. A way to coordinate decisions and manage a treasury using token-weighted voting instead of a traditional company structure. Token holders propose and vote on changes directly.Like a shareholder-run company where every shareholder can vote on every decision, the votes are public, and the company can't do anything the shareholders don't approve. The coordination is messier than a normal company but nobody has unilateral control.DAO in the glossary → would signal commitment to full decentralisation.
2. Expand security audit coverage. The August 2025 Code4rena audit covered 6 core capital and distribution contracts. Extending formal audits to the compute marketplace contracts on Base and the agent execution layer would close the remaining gap.
3. Grow the compute provider base. More providers means more redundancy, lower latency and stronger censorship resistance. The compute marketplace is live on Base but needs scale to deliver on the sovereign agent vision.

Returns Score: 58/100

MOR scores 58/100 (C grade). Full methodology at Returns Score Methodology.

Token utility (18/20): MOR does productive work across the protocol. Staking stETH earns daily emissions, and compute access requires MOR as stake. It’s not a governance-only token bolted onto an existing product. MOR is structurally embedded in every interaction layer. Capital providers, compute providers, code contributors and community builders all earn and spend MOR through distinct mechanisms. The multi-sided utility is among the strongest in the DeAI space.

Value accrual (11/20): MOR accrues value through two mechanisms: users stake MOR for compute access, which locks float while sessions run but returns it in full at close, and the MRC43 burn step. Capital-provider yield is spent on MOR in full, and half of every purchase is permanently burned while the other half is locked for tail emissions. MRC43’s 25% liquidity-pairing leg hasn’t run since January 2025.

The capital-yield engine is the quantifiable part: DeFiLlama tracks $7.3 million of MOR bought back over the protocol’s life from harvested stETH and Aave yield. That is an on-chain value-accrual flywheel, and the strongest evidence in this dimension. But it has shrunk hard with the deposit base, so the current buy pressure is a fraction of daily emissions. The mechanisms are sound; the scale, on both compute and yield, isn’t there.

Supply dynamics (17/20): The fair launch with zero insider allocation is rare. No VC vesting schedule, no team vesting cliffCliffA waiting period at the start of a token vesting schedule during which no tokens unlock at all. After the cliff ends, tokens begin releasing according to the vesting schedule.Like a probationary period at a new job. You don't get your stock options on day one. You wait 12 months to prove you'll stick around, then everything starts unlocking normally.Cliff in the glossary →, no foundation treasury dump risk. A 16-year linear emission decline (14,400 MOR/day at launch, falling by a fixed ~2.47 MOR/day to zero on day 5,833, the end of Epoch 1) means circulating supplyCirculating SupplyThe number of tokens currently in circulation and tradeable on the open market. Differs from total supply (which includes locked or unvested tokens) and max supply (the upper limit, if there is one).Like the number of cars on the road today versus the number ever produced. Some are in showrooms, some in junkyards, some still at the factory. Only the ones on the road count toward what people are actually driving.Circulating Supply in the glossary → grows via emissions in the early years against a hard 42M cap. The Power Factor mechanic (MRC42) constrains some sell pressure: participants can voluntarily lock claims for a multiplier capped at ~10.7x, and locks can only be extended, never shortened. Measured on chain in August 2026, 265 of 6,428 capital stakers hold a live lock, covering 917,127 MOR or 2.2% of the terminal cap, and the multiplier decays with start date rather than being a function of duration. Real, and smaller than the original scoring assumed. One concern: current emission rates still run ahead of demand-driven burns by a wide margin.

Revenue sustainability (8/25): This is the weakest part of the thesis, and the data has worsened. Morpheus’s only material revenue is the yield it captures on capital-provider deposits, which DeFiLlama tracks on-chain: stETH rebasing plus Aave V3 interest, with about 75% funding MOR buybacks. That engine has collapsed. The deposit base fell from a mid-2024 peak near $579 million to roughly $13 million by June 2026, and captured yield fell with it, from a lifetime $9.8 million to a run-rate near $0.4 million a year. The dollar figure has since risen to $18.3 million, but the deposit base has not recovered: the stETH pool holds 7,386 stETH, flat since the June reading, so the move is the ETH price rather than returning capital, and DeFiLlama’s trailing-year revenue stood near $0.54 million in late September 2026, with the 30-day figure annualising to about the same.

So the rising floor this dimension was bumped for in April 2026 has not held; most of that capital was emission-chasing stETH that left as MOR fell sharply from its high. The buyback mechanism is sound and on-chain, $7.3 million of MOR repurchased over the protocol’s life, but at the current rate it is a trickle against emissions. Compute revenue from external customers, the AI-native stream that would actually back the token, stays negligible. The score reverts from 10 to 8: a proven but shrinking yield-funded buyback, and no compute revenue yet.

Liquidity and access (4/15): MOR trades only on decentralised exchanges: every venue CoinGecko tracks is a DEX pool, and no centralised exchange lists it. Volume is dangerously thin. A position of any meaningful size is difficult to enter or exit without moving the price. This is a practical constraint that most project assessments gloss over: it doesn’t matter how good the fundamentals are if you can’t trade the token without 5-10% slippageSlippageThe difference between the expected price of a trade and the price you actually get when the trade executes. Slippage usually goes against the trader and gets worse with bigger trades or thinner markets.Like trying to buy 1000 bananas at the corner shop. The first few are at the marked price, but by the time you've bought them all you've moved the price up because there are no more bananas left at the original level. The shop has to restock at higher cost.Slippage in the glossary →.

Path to improvement

Three changes would materially increase Morpheus’s returns score:

1. Secure centralised exchange listings. DEX-only trading with dangerously thin volume is the most immediate constraint on capital formation. Even a single tier-2 CEX listing would meaningfully improve liquidity and discoverability.
2. Grow compute marketplace revenue from external customers. The capital-yield engine that funded buybacks has collapsed with TVL, so the buyback flywheel now runs on fumes. Paying compute customers who are not already MOR holders would create external demand and a revenue stream that does not depend on mercenary yield-farming deposits.
3. Distribute compute fees to holders. As the compute marketplace scales, directing a portion of compute fees to MOR holders or capital providers would create yield backed by compute revenue, complementing the existing stETH-funded buyback and burn.

Score change log

Date	Score	Change	Reason
2026-10-20	Data	N/A	The TEE provider moved host on 14 October (v7.15.0) and now serves two new TEE models, both passing Phase 2; our Phase 1 check re-verified on the new host. Scores unchanged.
2026-10-17	Data	N/A	Added that SEV-SNP host attestation failed open until v7.15.0 (14 October 2026), per its release notes: any SEV measurement passed Phase 1. Scores unchanged.
2026-10-15	Data	N/A	Buyback MOR a day and its share of emissions are now measured on Arbitrum from the burn and 16-year lock balances. The old figure converted DeFiLlama’s 30-day dollar total at today’s price and read about 60% high after the price fall. Scores unchanged.
2026-10-13	Data	N/A	Added the monthly provider-concentration series from our inference index: the core team’s provider.mor.org was the busiest address every month February to August. Scores unchanged.
2026-10-13	Data	N/A	Verdict summary corrected: since 9 October the TEE provider reports no model passing Phase 2 (two flagged, gpt-oss:120B:tee down), so “passing on some” was stale. Scores unchanged.
2026-10-09	Data	N/A	Six-year Power Factor (was 8.6x, now ~8.4x), years to the cap, lifetime buyback (was $7.16M) and provider concentration now bind to live data. Scores unchanged.
2026-10-06	Data	N/A	Tokenomics now opens on stake as the demand, the yield-funded buyback and burn as the only live route, and dilution; lifetime buyback and Power Factor bound live, “persistent” access demand and the unsourced “agent services consume it” dropped. Scores unchanged.
2026-10-04	Data	N/A	Added that native (non-gateway) tokens set a monthly record in September, about 7.25B against 4.25B in August per tech.mor.org, while the network total stayed a third below July. Scores unchanged.
2026-10-04	Data	N/A	Demand framing refreshed from our index: September tokens edged up about 8% on August while sessions rose about a third, so the ‘fewer but larger sessions’ line was stale. USP line reworded to a checkable privacy claim. Scores unchanged.
2026-09-25	Data	N/A	Corrected two stale TEE gaps: AMD SEV-SNP attestation shipped in v7.2.0 (2 June 2026), and the api.mor.org gateway now verifies provider attestation, with its own hop policy-based. Scores unchanged.
2026-09-24	Data	N/A	Tested TEE from our own consumer node: the Phase 2-passing model opened and attested, a flagged model was refused before staking. Noted that routers before v7.9.0 fail TEE sessions against the current SecretAI portal. Scores unchanged.
2026-09-24	Data	N/A	Phase 2 TEE observed live: the provider’s public healthcheck (v7.11.0) reports backend attestation per model, now tracked daily with our Phase 1 re-check across all TEE models. Data Sovereignty held at 14/15 after review.
2026-09-24	Score	Freedom 78 to 76	Governance 12 to 10: one 5-of-9 Safe holds upgrade authority with no timelock and the Snapshot space is gone. Phase 2 TEE wording corrected to credit the shipped proxy-router code.
2026-09-24	Data	N/A	TEE prose scoped to the verified Phase 1 (Phase 2 attributed to release notes); access is by staked MOR, not holding; governance names the 5-of-9 upgrade Safe and the lost Snapshot space; no CEX lists MOR; unlicensed repos and current revenue added. Scores unchanged.
2026-09-23	Data	N/A	Integrity records re-cut: 16 claims now quote Morpheus’s own pages verbatim, 26 are our own measurements, and one third-party session count was removed. Scores unchanged.
2026-09-19	Data	N/A	tee_attestation and thesis_permissionless_private moved from unverified to verified: a live Intel TDX quote from the TEE-tagged provider binds to its TLS certificate (computed by us) and passes the vendor signature check. Scores unchanged.
2026-09-18	Data	N/A	mrc_snapshot_voting moved match → outdated: the morpheusai.eth Snapshot space we cite fails to load, confirmed four ways with positive controls. Governance flagged for the monthly review.
2026-09-17	Data	N/A	self_custodial_wallets moved from unverified to verified: we run a provider from a key we hold, with no custody contract in the flow. bug_bounty converted to permanently unre-checkable, since a published programme is not evidence of a paid bounty. Scores unchanged.
2026-09-17	Data	N/A	no_protocol_content_policy moved from unverified to verified on-chain: the deployed LumerinDiamond’s full 73-selector map carries no pause, blacklist, ban or content-policy function. Scores unchanged.
2026-09-07	Data	N/A	Both permissionless-provider checks moved from unverified to verified on-chain. providerRegister on the source-verified ProviderRegistry facet carries no owner modifier and no allowlist, and we registered our own node from a fresh address with no approval step. Scores unchanged.
2026-09-05	Data	N/A	Correction to the row below: it gave the August fall as 56.6%, the SESSION count, under the word “throughput”, which means tokens. Tokens fell 38.2% (16.12B to 9.96B), matching tech.mor.org exactly; thesis_inference_demand moved match to outdated.
2026-09-05	Data	N/A	§8e sweep correction. The demand paragraph said throughput had climbed every month since February and peaked mid-July; August reversed it. Leading model corrected to DeepSeek V4 Flash.
2026-09-05	Data	N/A	Bug Bounty Program v2 published 7 Aug 2026, superseding all prior versions: $100K ceiling from the Protection Fund capped at 10% of funds at risk, scope narrowed to deployed bytecode at a named address list, and a mandatory mainnet-fork proof of concept. Read from mor.org/bug-bounty via Chrome (the page bot-gates automated fetches). No score change.
2026-08-21	Data	N/A	Corrected the Power Factor against LockMultiplierMath: it is anchored to a fixed 2024 to 2040 window, so a six-year lock reaching 10.7x at launch reaches ~8.6x today, and the contract sets no maximum lock. Scores unchanged; effect deferred to the September review.
2026-08-16	Data	N/A	Rewrote the post-2040 section. The epoch-2 mechanism is the whitepaper’s burn-funded tail emission implemented by MRC43, not MRC22’s undistributed-emissions bank, and the earlier text framed a live mechanism as spec-stage intent. Verified on Arbitrum: the burn address and the 16-year lock contract each hold 287,795 MOR across 60 transfers. Scores unchanged.
2026-08-15	Data	N/A	Corrected five access-mechanism claims that described the Yellowstone design doc rather than the deployed Diamond: the budget rule is on-chain (getTodaysBudget, stakeToStipend), rationing is by session stake against total supply rather than wallet balance, and direct pay is stipend-limited so it buys no access beyond the ration. Scores unchanged.
2026-08-15	Data	N/A	Compute pool peaked near 2,805,000 MOR in mid-July 2026 and fell to 2,778,186 by 15 August, as provider claims (~3,800 MOR/day, 30d) overtook the pillar’s ~2,912 MOR/day accrual. MOR committed to access rose from 123,259 to 777,023 MOR since 16 June. Revenue Sustainability unchanged: access demand is token demand, not revenue.
2026-08-13	Data	N/A	Re-pointed all 27 Morpheus claim sources from the bare mor.org homepage to the page carrying each claim (mor.org/mor-token, the GitBook docs, the bug-bounty programme), so a reader can trace every verdict to where the project actually says it.
2026-08-13	Data	N/A	The re-point caught a stale figure of ours: the bug bounty was rewritten on 7 August 2026 to a $100,000 ceiling capped at 10% of funds at risk, superseding the “$500-$150k” we carried. Corrected; verdict stays unverified since a published programme is not a paid bounty.
2026-07-19	Data	N/A	Refreshed the “demand, honestly” framing from the stale “still modest”. Our on-chain inference index shows attested throughput up every month since February and an all-time daily record on 15 July 2026, led by open-weight coding traffic (GLM 5.2); supply stays thin (~7-12 active providers) and revenue is still capital yield, so scores unchanged.
2026-07-14	Data	N/A	Reconciled the displayed token contract to the live Base MOR address (0x7431aDa8…); the research JSON previously pointed at the Arbitrum-legacy OFT as MOR migrates onto Base. Scores unchanged.
2026-07-10	Data	N/A	Wired MorScan (signed third-party Morpheus indexer) as a data_sources provider plus a live network component, cross-checked against our own Alchemy holder count and SessionClosed index (both reconcile). Scores unchanged.
2026-07-07	Data	N/A	A DefiLlama LlamaAI forensic (23 June 2026) flagged the Builders proxy implementation as unverified; no longer true. It’s source-verified on BaseScan as BuildersV4 (EIP-1967 slot read); upgrade authority stays a 5-of-9 Safe. Added a reconciliation and wired the slot as an onchain-rpc source so a later upgrade trips the drift guard. Scores unchanged.
2026-07-06	Editorial	N/A	Broadened “Who it’s for” to name the agent-builder audience (central to Morpheus’s design), framed honestly as more vision than shipping product today. No score change.
2026-07-05	Editorial	N/A	Added a “The product in practice” section (per specs/product-comparison-framework.md) from our on-chain inference index and hands-on node use; the USP is verifiable hardware-attested privacy (v7.0 TEE), demand is modest and on-chain-auditable, and revenue today is capital-provider yield rather than compute fees. No score change.
2026-07-04	Data	N/A	Reconciled the yield-revenue record. The trailing-12-month revenue_annual_usd had gone stale ($1.03M, metrics_date still 2026-03-01); refreshed to ~$976K against the 1 July 2026 DeFiLlama primary_data pull (total1y $975,979). It was never in conflict with the ~$0.4M/yr run-rate cited in Returns evidence: those are different windows (trailing-12m vs current run-rate), now made explicit with an integrity.reconciliations[] entry (verdict: match). Lifetime ~$9.8M; run-rate ~$0.4M/yr and falling with TVL. Scores unchanged.
2026-08-19	Data	N/A	The 75/25 buy-and-pair split is MRC43 design intent, not deployed behaviour. On Arbitrum the yield buys MOR in full, then splits exactly 50/50 between the burn address and the 16-year lock; the last MOR reached the Uniswap position on 3 January 2025. Corrected here and in both tokenomics articles. Scores unchanged.
2026-06-18	Returns	60 → 58	Revenue Sustainability 10 → 8, reverting the April 2026 bump. DeFiLlama’s morpheus-ai adapter (verified live and against the adapter source) shows the protocol’s only material revenue is captured capital yield (stETH rebasing + Aave V3 interest), not compute, and it has collapsed with TVL: ~$579M peak (mid-2024) to ~$13M (Jun 2026), lifetime revenue ~$9.8M but run-rate ~$0.4M/yr. The “rising floor” premise the bump rewarded did not hold; most of that was emission-chasing stETH that fled. The yield-funded buyback is real (~$7.16M of MOR repurchased lifetime, ~75% of yield) but now a trickle vs emissions; compute revenue still negligible. Corrected the rising-floor/deepens-daily framing across Revenue Sustainability, Value Accrual, the yield-engine pillar, pros/cons and snapshot; populated revenue_annual_usd (~$1.03M trailing-12m) + the DeFiLlama manifest note.
2026-06-15	Data	N/A	POL buy/LP split corrected from the stale 50/50 bootstrap pattern to ~75% MOR-buy / 25% LP per MRC43 (yield-engine cards, capital-provider prose, value-accrual rationale). Burn-split phrasing tightened: half the MOR left after each POL generation event is burned, the other half locked for tail emissions. Added a verified on-chain POL figure to the tokenomics deep dive (21 events / ~290 wstETH in the year to 15 June 2026; ~2,677 wstETH all-time since May 2024), summed from wstETH inflows to the Arbitrum executor L2TokenReceiverV2 and independently reproduced via Etherscan v2. Scores unchanged.
2026-06-09	Data	N/A	Compute-payment mechanics clarified and a conflation corrected. Sharpened the compute-provider line: providers earn from the 24% Compute pillar by session activity (bid price x session length), paid from a protocol funding account, not by burning the user’s MOR; staked access returns MOR in full, only direct-pay spends it. Fixed the matching “burned as transaction fees” conflation in the research JSON (burn_mechanism, token_utility, token_model_classification, revenue_source) and the summary. Linked the new How Morpheus Pays for Inference deep dive. Verified against tech.mor.org session lifecycle, the Lumerin Diamond on Base, and our own SessionClosed indexer. Scores unchanged.
2026-06-09	Data	N/A	Added Epoch 2 / post-2040 emission structure to Tokenomics (“After Epoch 1: the post-2040 bank”): undistributed Epoch 1 emissions bank and re-emit on the same declining curve in Epoch 2 (years 17-32) per MRC22, structurally guaranteed by Protection Fund + POL MOR counting in circulating supply and being unstakeable; reschedules within the 42M cap, no new mint; each epoch smaller than the last. Flagged as MRC22 design intent (In Progress), not deployed code; live curve is hard-coded LinearDistributionIntervalDecrease to zero. Tagged the two “16-year decay to zero” mentions (TokenFactsStrip + Supply dynamics) as Epoch 1. Verified against MRC22. Scores unchanged.
2026-05-28	Data	N/A	Builder-pillar framing corrected: Community builders operate builder subnets (Smart Agents / end-user apps), not content/docs (post-MRC22). Emission curve corrected from “halving every four years” to linear decline (2.47 MOR/day to zero on day 5,833). Verified against MRC22 + deployed BuildersV4 contract. Scores unchanged.
2026-04-26	Freedom	76 → 78	v7.0.0 Full TEE Capability shipped 23 April 2026. Data Sovereignty 12→14 (Phase 2 backend LLM attestation closes the explicitly-flagged GPU attestation gap; Intel TDX only and NVIDIA-only GPU support remain).
2026-04-06	Freedom	76 → 76	Code4rena audit (Aug 2025) reviewed. Open Source & Transparency unchanged at 12. Compute/agent contracts still unaudited.
2026-04-06	Returns	57 → 60	Supply Dynamics 16→17 (Power Factor MRC42 underweighted). Revenue Sustainability 8→10 (stETH yield is verifiable on-chain external revenue, underscored vs peers).
2025-03-01	Both	N/A	Initial publish. Freedom 76/100, Returns 57/100.