DIEM: How Venice Turned Staking into Perpetual AI Credit

What DIEM actually is

Most 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.Read more → tokens give you yield. DIEM gives you compute.

Each DIEM staked on Venice generates $1 per day in 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.Read more → 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.Read more → credit. Not a coupon. Not a discount code. A perpetual, daily credit that works against Venice’s full model catalogue, from Llama to Stable Diffusion. Stake 10 DIEM, get $10/day in inference. The credit resets daily, doesn’t roll over, and works exactly like paying cash through the API.

That’s the pitch. The mechanism underneath is where it gets interesting.

The minting pipeline

Getting from VVV (Venice’s native 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.Read more →) to usable inference credit takes four steps. Each one locks capital deeper into the system.

Step 1: Buy VVV. On any exchange or 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.Read more → where it trades. Base chain is the primary market.

Step 2: Stake VVV to get sVVV. Standard staking. You earn 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.Read more → yield (currently around 20-30% APY depending on lock duration). Your VVV becomes sVVV, a staked receipt token.

Step 3: Lock sVVV to mint DIEM. Your sVVV doesn’t get destroyed. It moves into escrowEscrowA contract that holds tokens on behalf of a user under a defined release condition. The tokens are not destroyed and not freely tradeable. They sit locked until the condition is met (a burn, a time elapsing, a counterparty action).Like leaving the deeds to your house with a solicitor while a sale completes. You still own the property, but you can't sell or remortgage it until the escrow releases.Read more → against the DIEM you mint. Venice’s own protocol post puts it plainly: “Burn DIEM to unlock sVVV whenever you want your original tokens back.” Burns can be partial. Burn half your DIEM, get half your sVVV released. While the sVVV sits in escrow, you keep 80% of the staking yield. The other 20% flows to Venice as protocol revenue. Your VVV is removed from 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.Read more → for as long as the DIEM remains minted, but the lock is reversible at any time by burning the DIEM.

Step 4: Stake DIEM. Once minted, stake your DIEM on Venice to activate the $1/day inference credit. Unstaked DIEM earns nothing. It’s a tradeable ERC-20 on Base, so you can sell it, but it only generates credit when staked.

The result: a liquidity lock, not a sacrifice. VVV holders who want inference access park their stake in escrow and accept the illiquidity. Venice gets a contractionary force on circulating supply that scales with DIEM demand. The holder gets a tradeable, income-producing position they can unwind by burning the DIEM, provided they still have it.

The exponential lock curve

Here’s where most people’s eyes glaze over. Don’t let them. This curve is the entire mechanism.

How the formula works

The amount of sVVV you have to lock per DIEM minted (Venice’s term: the mint rate) follows an exponential function:

A note on language. The mint rate isn’t a “cost” in any conventional sense. You don’t spend sVVV. You lock it in escrow. You can burn the DIEM later and get the sVVV back. So we’re going to call it the lock requirement when we’re talking about user economics, and mint rate when we mean Venice’s published parameter. What you actually carry when you mint is illiquidity on that sVVV, price risk on the locked VVV, and a 20% yield share to Venice. Calling it a “cost” anchors readers in the wrong frame.

In plain terms: the first DIEM required 90 sVVV. As supply approaches the target (38,000), the lock requirement rises on an exponential curve. Not linear. Exponential. The difference matters, and it matters more now that supply has crossed the target.

At ~685 sVVV per DIEM today, the lock requirement is steep, and it accelerates from here. Past the target, the curve goes near-vertical: the same 0.5% supply increment from 100% to 100.5% pushed the requirement up by roughly 3%. A run to 40,000 supply (only 5% past target) would push it to ~865.

What the curve looked like in practice

What this curve actually looked like in practice, week by week since DIEM launched, has not been published anywhere else. We rebuilt it from on-chain data:

DIEM Mint Rate History

Weekly snapshots of the sVVV collateral required to mint 1 DIEM. Bars show the rate at each weekly checkpoint; the faint band behind them is the trailing 30-day low-high envelope. Tracked exclusively by OYM from on-chain Mint and Burn events.

Current rate

743 sVVV

DIEM supply

38,687 /38,000

Lock requirement (sVVV per DIEM)

2.1415586472655752e+57 1.0707793236328308e+57 8.648173248330999e+43

Rate at target ~665 sVVV

Aug 2025 Oct 2025 Dec 2025 Jan 2026 Mar 2026 May 2026

Week (last 52 weeks)

When the rate crossed each threshold

100 sVVV

2025-08-20

@ supply 20,534

150 sVVV

2025-08-20

@ supply 30,801

200 sVVV

2025-08-20

@ supply 30,801

300 sVVV

2025-08-20

@ supply 41,068

500 sVVV

2025-08-20

@ supply 41,068

665 sVVV

2025-08-20

@ supply 41,068

Two structural observations from that chart. First, launch day (20 August 2025) was a frenzy. Supply rocketed from zero through 30,000 in hours, dragging the lock requirement from 90 sVVV all the way to ~370 in a single session. Most of the early DIEM was minted at favourable rates only because participants moved fast on that one day. Second, after launch the lock requirement actually fell for several weeks: net burning exceeded net minting as users released their sVVV. That elasticity is invisible if you only watch the DIEM price chart. The full event log, methodology and live data are at /projects/venice/diem/.

The secondary market

DIEM also has a secondary market. It trades on Aerodrome SlipStream on Base, and Coinbase added a DIEM/USD spot pair in late April 2026. Today DIEM trades around $1,082 (per DexScreener). That market price is a live signal of how the market values the perpetual $1/day credit. It moves on a different rhythm to the lock requirement: DIEM has roughly tripled in 2026 (from ~$340 on 12 February to ~$1,082 today) while the lock requirement climbed only modestly in the same window (~620 sVVV in mid-February to ~685 today). The lock curve had already done most of its work by January, so since then it’s the secondary market doing the revaluing.

The flywheel

Venice designed DIEM to create a self-reinforcing cycle. It works like this:

1. Demand for inference drives demand for DIEM (cheaper than paying cash if you use enough compute)
2. Demand for DIEM drives demand for VVV (you need sVVV to mint)
3. Minting DIEM locks VVV in escrow, removing it from circulation
4. Reduced circulating VVV pushes VVV price higher (all else equal)
5. Higher VVV price makes minting more expensive in dollar terms, increasing DIEM’s relative value
6. Higher DIEM value attracts more minters, locking more VVV

It’s a contractionary loop, but an elastic one. The supply lock is real while DIEM exists, but DIEM holders can unwind it by burning their tokens to reclaim the underlying sVVV. So the circulating VVV supply shrinks as inference demand rises, and can expand again if a wave of holders decides to redeem. The pressure is one-way only as long as nobody wants their sVVV back.

Currently 8.26 million VVV is locked for DIEM. That’s roughly 25% of all staked VVV and about 7% of total supply, held in escrow against the outstanding DIEM. With emissions dropping from 6 million to 5 million VVV per year on 1 May 2026, and again to 3 million on 1 July 2026, the structural squeeze on new issuance tightens further. Burns can still release locked VVV back to circulation, so the net direction depends on mint flow minus burn flow.

29,340 DIEM (77% of supply) is currently staked. That generates demand for 29,340 credits per day against Venice’s network capacity of 18,148 DIEM per day. Oversubscribed by about 62%. Venice hasn’t disclosed how it handles the oversubscription, whether credits are pro-rated, queued, or if the capacity figure refers to something else entirely.

How DIEM compares

DIEM sits in an unusual spot. It’s not a simple staking yield. It’s not a consumable credit. It’s a perpetual, tradeable claim on daily compute output.

The closest comparison is Morpheus. Both let you preserve principal while diverting (or in DIEM’s case, partially diverting) the staking yield in exchange for product access. MOR staking diverts your stETH yield to the protocol. DIEM gives Venice 20% of your sVVV yield while you keep the other 80%, and instead of MOR-style emissions you get the daily compute credit. Both use a receipt token that can be redeemed for the original collateral, although DIEM redemption requires you to still hold the DIEM you minted.

Akash’s ACT is a consumable credit: you 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.Read more → AKT to mint ACT, use the compute, ACT gets burned, AKT gets reminted to the provider. It’s a flow-through mechanism. DIEM is a stock, not a flow. You hold it and it produces daily output indefinitely.

TAO staking is the simplest of the three: stake for emissions yield, no product access attached. DIEM bundles yield (80% staking rewards retained) with product access ($1/day credit) into a single position.

DIEM sits at the high-illiquidity end of this set, but the principal is preserved. You’re not converting your VVV into a different asset class. You’re parking it in escrow against a receipt that produces daily compute credit, and the path back is conditional on still holding the DIEM you minted (or buying it back from the market). The commitment is to illiquidity and to operational risk in the receipt itself, not to surrendering the underlying capital.

The centralisation risks

Two things about DIEM that aren’t discussed enough.

The target supply is not on-chain. The exponential mint formula, including the 38,000 target supply parameter, lives off-chain. Venice can change it. If they raise the target to 50,000, the mint curve flattens and existing DIEM holders face dilution from cheaper new supply. If they lower it, minting becomes prohibitively expensive faster.

Venice has stated they plan to raise the target as network capacity grows. That’s rational from an infrastructure planning perspective. But it means your DIEM’s scarcity is a function of a company decision, not a protocol guarantee. There is no 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.Read more → vote, no multisig requirement, no timelock. Venice adjusts it when they decide to.

The $1/day credit has no pricing protection. Venice controls model pricing unilaterally. Today, $1 buys a meaningful amount of inference. If Venice raises prices (or if the models it offers become more expensive to run), your $1 buys less compute tomorrow. There’s no on-chain guarantee that $1 purchases a fixed number of tokens or a fixed amount of 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.Read more → time.

This isn’t theoretical. AI 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.Read more → pricing shifts constantly. If Venice migrates to more expensive frontier models or adjusts its pricing tiers, every DIEM holder’s daily allocation changes without any on-chain event. The $1 is nominal, not real.

Put these two together and you have a token where both the supply curve and the utility value are controlled by a single company. That’s a different risk profile from protocols where the rules are immutable smart contracts.

The maths for holders

Is DIEM a good deal? Depends on your assumptions, and more importantly, on which path you take to get one. The two paths look identical in product (a DIEM is a DIEM), but they sit on completely different financial structures.

The secondary market path

Buy DIEM for ~$1,082. This is an outright capital expenditure. Cash leaves your wallet, a DIEM arrives in it. Payback against the $1/day credit is the price you paid, in days: 1,082 days at today’s market, roughly three years. After that, it’s pure profit in inference terms. If you spend $300/month on Venice’s API, 10 DIEM ($10,820) replaces about three years of that spend before it starts compounding. Exit is trivial: sell the DIEM on the secondary market and recover whatever it’s paying that day.

The mint path

Lock ~685 sVVV. This is not a payment. You’re not spending sVVV. You’re posting it as collateral against the DIEM the mint contract issues you. The sVVV continues to exist, in your name, in the DIEM escrow. To release it later, you burn the DIEM. So the mint “cost” is misleading shorthand. What you actually accept when you mint is:

1. Illiquidity. Your sVVV can’t be sold or moved while it backs DIEM.
2. Price risk on the locked VVV. If VVV falls 50% while you’re locked, your unlock claim is worth half of what you posted. You can still burn DIEM to recover the same number of sVVV, but those sVVV represent fewer dollars.
3. A 20% yield share to Venice. Locked sVVV continues earning emissions, but Venice keeps a fifth of that yield instead of the staker getting all of it.
4. Operational dependency on holding the DIEM you minted. If you sell the DIEM, you can’t unlock without buying one back from the market. (More on this below.)

That last bit is what makes the mint path materially different from a normal collateralised position.

If you were planning to stake VVV anyway and you intend to keep your DIEM, minting is closer to a yield-share with Venice than a capital sacrifice. The annual cash cost is roughly 20% of the sVVV staking yield, not the $6,030 USD-equivalent of 685 sVVV.

That reframing also explains why the secondary market price isn’t going to converge on the lock requirement converted to dollars. They’re priced against different things. A buyer is paying cash for a perpetual credit stream. A minter is posting collateral for the same credit stream and accepting illiquidity in exchange. Two different cost bases produce two different prices for the same token, and that’s structural, not a market mistake.

The short-squeeze risk on your own collateral

Here’s the part that doesn’t get discussed elsewhere. If you mint a DIEM (locking 685 sVVV) and then sell that DIEM on the secondary market for $1,082, you’ve trapped your sVVV. To ever release it, you have to burn a DIEM, which means buying one back. Today that costs $1,082. If DIEM’s market price runs to $3,000 or $4,000, unlocking your own collateral suddenly costs you several times what you sold it for. The escrow only releases against burning a DIEM you hold, which functionally means burning any DIEM bought from the market. The minter who sells is structurally short DIEM against their own escrowed sVVV.

That dynamic creates conditions for a real short-squeeze on the token. If Venice raises model pricing, or a new use case spikes demand for the credit, every minter-who-sold becomes a forced buyer if they ever want their sVVV back. As DIEM has roughly tripled in 2026, anyone in that position has watched their unlock cost track the market all the way up. It’s a quiet incentive structure the existing DIEM coverage skates past.

For pure inference access, buying DIEM on the secondary market is the simpler path. For a long-term VVV holder who plans to keep the DIEM, the mint path can make sense as a yield-share. The path that doesn’t make sense is “mint then immediately sell” unless you’ve separately decided you never want your sVVV back.

What I’m watching

Emission cuts. VVV emissions drop from 6 million to 5 million on 1 May 2026, then to 3 million on 1 July. Each cut reduces new VVV entering circulation, while DIEM minting continues to pull existing supply into escrow. Net direction depends on mint flow versus burn flow, but the structural backdrop is tighter issuance.

Target supply adjustments. When Venice raises the 38,000 target (and they will, if demand continues), the mint curve resets. Early DIEM holders benefit from scarcity, but only until Venice decides to expand the cap. Watch for announcements.

Capacity vs staked DIEM. Currently 29,340 DIEM staked against 18,148 daily capacity. That ratio matters. If Venice doesn’t scale capacity to match staked supply, credit quality degrades, either through pro-rating or congestion.

Net mint flow. Track whether escrow is filling or draining each week. We publish the live trajectory at /projects/venice/diem/. Sustained net burning would mean the supply lock is unwinding, which loosens the contractionary pressure on VVV.

Lock requirement vs market price. The lock requirement is determined mechanically by the curve. The market price moves on demand for the credit. They drift apart and compress at different rates. The longer-term trajectory of both is the cleanest signal of whether the bonding curve is working as intended.

The honest assessment

DIEM is one of the more creative token designs in DeAI. Converting staked capital into a perpetual compute annuity, with an exponential cost curve that rewards early participants and a burn-to-unlock release condition that preserves principal, is legitimately novel. The flywheel mechanics are sound in theory: every DIEM minted pulls VVV into escrow, the exponential curve makes each subsequent mint progressively more expensive, and the elastic supply lock means the contractionary pressure is real but not irreversible.

The execution has some substance behind it. 38,182 DIEM minted (now past the 38,000 target), 8.26 million VVV in escrow, 77% staking rate. People are using this thing, not just speculating on it.

But I keep coming back to the centralisation. Both the supply target and the credit value are controlled by Venice, not by smart contracts. In a space that talks constantly about decentralisation, DIEM’s core economic 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.Read more → sit behind a company dashboard. Erik Voorhees has earned trust through a long track record in crypto (ShapeShift, SatoshiDice), and Venice’s incentives are aligned with holders for now. “Trust the team” is a reasonable position for some investors. It’s not a protocol guarantee.

If you’re a heavy Venice API user, DIEM at ~$1,082 breaks even on credits in roughly three years, and the underlying token is still yours to sell after that. Defensible if you genuinely use $1/day in inference and trust Venice operationally; harder to justify on a pure cash basis if you don’t. If you’re a long-term VVV holder considering the mint path, the calculus is closer to a yield-share with Venice than a capital sacrifice, but you take on the short-squeeze risk if you ever sell the minted DIEM. If you’re buying as a pure investment, you’re betting on Venice’s continued growth, the stability of the $1/day value, and a company’s willingness to keep the parameters favourable. That’s a bet on Venice the company, not on an immutable protocol.

I hold staked VVV and a small DIEM position. I minted that DIEM during the initial low-rate window and stake it for daily inference. The burn-to-unlock mechanism removes the worst version of the lock-up risk, but I want to see how Venice handles the first target supply adjustment before I lock more capital in a system where the rules can change without my inputPromptThe 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.Read more →.

For full context on Venice’s architecture, privacy modes, and scoring, see our Venice project review.