Celestia’s product loop and where TIA sits

Celestia sells one scarce thing: data availability bandwidth, packaged as blobspace, and secured by a proof-of-stake validator set. Mainnet Beta (Lemon Mint) went live on October 31, 2023.

TIA is wired into that loop in three concrete ways:

1) Pay for blobspace. Rollups post data to Celestia via PayForBlobs transactions, paying fees in TIA.

2) Secure consensus. Celestia is a Cosmos SDK chain using proof-of-stake with an initial validator set size of 100. Stakers delegate TIA to validators and earn rewards.

3) Govern a bounded parameter surface. Governance exists, but it is not “you can tune anything whenever.” Some parameters are explicitly changeable via governance and many are not.

From a mechanism design standpoint, Celestia’s key bet is that blobspace demand grows into a meaningful fee stream. If it does not, the token economy defaults to a familiar Cosmos shape: inflation funds security, and non-stakers are diluted to pay stakers.

For a Cosmos baseline comparison, see our ATOM tokenomics review.

TIA supply: genesis, allocations, and vesting

Genesis supply was 1,000,000,000 TIA, with unlock schedules by allocation bucket.

On the public market, the relevant reality is that TIA has ongoing issuance. CoinGecko lists max supply as ∞, and reports circulating supply, total supply, and outstanding supply (values change over time).

Celestia’s Genesis Drop allocated 60,000,000 TIA (6% of total supply) across eligibility categories, with tokens accessible “from the first block.”

Allocations at genesis (amounts below are calculated directly from the 1,000,000,000 TIA genesis supply and the published percentages):

One detail that matters for market microstructure: the docs note that due to 2024 being a leap year, yearly unlock intervals occur on October 30 of each year, and give October 30, 2024 as the “year 1” unlock example.

Issuance: a “strict schedule” that still moved twice

Celestia’s stated design intent is predictable issuance. TIA inflation started at 8% annually, and was originally set to decrease by 10% per year until reaching a 1.5% floor.

The mechanism behind that predictability is explicit in Celestia’s strict inflation schedule. It contrasts Cosmos SDK’s typical “bonded-ratio-targeting” inflation with Celestia’s approach, and defines constants for initial inflation, disinflation rate, and target inflation.

It also specifies an important accounting rule: annual provisions are based on total supply at the beginning of each year, so intra-year minting does not increase that year’s provision base.

Then reality intervened. Two upgrades rewired the schedule:

July 2025 (v4 Lotus, CIP-29): inflation dropped from ~7.2% to ~5.0% and the annual decrease was set to 6.7% until 1.5%.

November 2025 (v6, CIP-41): inflation dropped again to ~2.5% and continues decreasing by 6.7% until stabilizing at 1.5%.

Aave tokenomics is a useful comparison point for how “collateral friendliness” pressures designs to reduce opportunity costs.

CIP-41 is explicit about motivation: reducing the opportunity cost of using TIA as collateral in financial applications, while balancing validator economics through a commission floor.

From my bias toward deterministic rules, Celestia did two things at once:

It made issuance computable on a calendar. That increases modelability.

It also demonstrated that issuance is still a policy lever, just one pulled through coordinated upgrades and CIPs rather than day-to-day parameter governance. That lowers confidence in long-horizon parameter stability.

Fees, blobspace pricing, and where value actually accrues

Blobspace demand becomes economic demand for TIA through transaction fees. The docs describe a standard gas-priced mempool where higher gas price transactions are prioritized by validators.

For PayForBlobs, Celestia gives a clear gas limit formula with a fixed cost plus a size-based component.

Gas limit formula: Gas Limit = FC + Σ SparseSharesNeeded(blob) × ShareSize × GasCostPerBlobByte, with FC = 65,000 gas as the default fixed cost approximation.

The immediate mechanism-design implication is that blobspace pricing is gas-parameter mediated. That is good. It is legible. It is tunable.

But it is not fully “onchain-policy tunable.” In the app’s v6 parameter sheet, blob.GasPerBlobByte = 8 and is marked not changeable via governance.

That creates a mild but real documentation tension. The submit-data docs describe gas-per-byte style parameters as potentially adjustable “through governance mechanisms.” The specifications say that at least some of those knobs are currently filtered or hardcoded.

Value flow is also explicit. ADR-019 describes minting and distribution flow: newly minted tokens are minted into the mint module account and moved to the fee collector, then distribution logic applies.

Crucially, Celestia sets proposer bonuses to zero. ADR-019 states BaseProposerReward = 0 and BonusProposerReward = 0. The v6 parameters list the same.

The community pool is funded via a community tax. Celestia’s docs describe the community pool receiving 2% of block rewards. ADR-019 describes community tax applied as a fraction of fees, and explicitly states CommunityTax = 0.02. Cosmos SDK distribution docs define the mechanism as community_tax × fees (plus rounding dust), and the remainder to bonded validators by power.

If you're modeling rollup token economics, contrast this fee loop with the Optimism tokenomics approach.

Two takeaways that matter for token value capture:

No automatic burn. The described flow is redistribution to validators and the community pool, not a sink that reduces supply.

Fee-driven security is the endgame, not the current default. If blob fees are small relative to issuance, security spend is inflation-funded and value accrual is a dilution transfer from non-stakers to stakers.

Staking, slashing, and validator economics as the “security budget API”

Celestia’s token economy is ultimately a security budget mechanism with a DA metering sidecar.

The chain uses utia as the bond denomination, with 6 decimals (1 TIA = 1,000,000 utia).

Staking has a meaningful exit friction. The v6 parameters list UnbondingTime = 21 days. That is a standard Cosmos trade. It reduces fast-exit security failures, and it increases the capital cost of risk management for stakers.

Slashing parameters define who eats tail risk. In v6, double-sign slashing is 2%, and downtime slashing is 0%, with MinSignedPerWindow = 75% over SignedBlocksWindow = 10,000.

The most opinionated validator-econ change in recent history is CIP-41’s commission floor move. CIP-41 proposes increasing minimum validator commission from 5% to 10%. In the v6 parameter sheet, staking.MinCommissionRate = 10%.

Mechanically, that is a redistribution between two groups that both call themselves “stakers.” Validators receive a larger guaranteed cut of rewards, while delegators eat the reduction. The system is trying to prevent a predictable failure mode of “issuance compresses, validator revenue collapses, validator set quality degrades, security declines.” That is the right failure mode to worry about on a DA layer.

Governance and parameter control: constrained onchain, flexible offchain

Celestia’s parameter control is best described as bounded onchain governance plus upgrade-driven policy.

The v6 parameter registry explicitly labels which values are changeable via governance and which require a hard fork or are blocked by filters. It also notes that governance-modifiable parameters can be updated by submitting governance proposals containing module MsgUpdateParams messages.

Two governance facts matter more than all the usual “token holders vote” talk:

Mint parameters are hardcoded. The v6 parameter sheet states that none of the mint module parameters are governance modifiable because they have been converted into hardcoded constants. That aligns with ADR-019’s design choice to remove governance-modifiable parameters from mint.

Onchain governance does not fully bind protocol evolution. The CIP process blog states that on-chain governance is limited to certain protocol parameters, and can be overridden or updated with upgrades by the off-chain social layer.

This is the core trade-off:

Deterministic rules improve predictability inside a given protocol version. Celestia leans hard into that, especially for issuance.

Upgrade-driven changes preserve governance adaptability. Celestia also uses that path, as shown by CIP-29 and CIP-41 materially changing issuance dynamics.

For builders, “predictable tokenomics” therefore means “predictable conditional on no major upgrade changing the constants.” That is still useful. It is not the same thing as immutability.

If you are doing tokenomics consulting for an app that wants to rely on TIA as a long-duration collateral asset, treat “upgrade governance” as a first-class risk factor. The parameter sheet is the starting point, not the guarantee.

Risk analysis: the dominant risk is security budget compression

Celestia’s design is coherent, but it concentrates risk in a few places that are measurable and manageable.

Top 3 risks

  1. Security budget compression and validator centralization (dominant risk). Trigger: sustained issuance reductions (via upgrades) without compensating growth in fee revenue from blobspace demand. Mechanism: validator revenue is a function of inflation-funded rewards plus fees, shaped by commission floors and bonded ratio. CIP-41 explicitly reduces inflation and raises the minimum commission floor, which protects operators but changes delegator economics. Who bears it: rollups and users posting blobs bear liveness and safety tail risk; delegators bear opportunity cost and potential governance power concentration; non-stakers bear dilution if issuance remains material. Indicators: bonded ratio and its trend, validator set concentration, median validator profitability proxies, and the ratio of fee revenue to inflation issuance (fees become visible onchain, and distribution flow is defined).

    Dominant risk: Celestia is a DA layer. Its value proposition to rollups is that data is available, verifiable, and credibly neutral. That promise is only as strong as the validator set’s honesty threshold and operational quality. In a PoS system, you buy that security with rewards.

    Celestia’s issuance policy has already been moved downward twice. The rationale is understandable. High inflation can make the token unusable as collateral because staking yield sets a high hurdle rate for every other use. CIP-29 and CIP-41 both explicitly target that problem.

    The mechanical question is what replaces issuance as the security budget. The only sustainable replacement is fees from blobspace usage, since Celestia does not describe a burn-and-scarcity value capture loop. The fee flow is redistribution to validators and the community pool. That is fine, but it has a sharp implication: if fees stay low, reducing issuance is not “making the token better.” It is reducing the protocol’s ability to pay for a high-quality validator set.

    Celestia partially mitigates this by raising the minimum commission floor to protect operators. But that shifts burden to delegators, which can reduce bonding participation, which can reduce security, which can increase perceived risk for rollups, which can reduce blob demand. That is a feedback loop worth naming.

    From a mechanism design perspective, the cleanest long-term equilibrium is one where blob fees are an increasingly large share of validator revenue, issuance is low and stable, and governance no longer needs to touch inflation for “token usability” reasons. Celestia is directionally steering there. The uncertainty is whether demand arrives fast enough, and whether pricing parameters and blockspace supply policies can respond without turning into discretionary governance.

  2. Policy instability via upgrades (rule predictability is version-scoped). Trigger: macro shifts in desired token positioning (collateral friendliness, DeFi integration, ecosystem incentives) that motivate more changes to issuance, commissions, or resource pricing. Mechanism: Celestia’s own CIP process framing acknowledges that on-chain governance is limited and that the off-chain social layer can override parameters via upgrades. Who bears it: anyone building a long-dated economic model around “TIA yield,” “TIA dilution,” or “blob fee floors.” Indicators: frequency of CIPs touching economics, upgrade cadence, and divergence between documentation claims about “governance adjustable” parameters and the parameter registry marking values as not changeable.

  3. Value capture softness if TIA is only a fee token for DA and nothing else. Trigger: rollups minimize TIA exposure, treating it purely as a passthrough fee asset, while DA competition pressures fees downward. Mechanism: TIA demand is structurally tied to paying for blobspace in TIA, but rollups can avoid using TIA as their internal gas or money. The docs describe using TIA to bootstrap rollups as an option, not a requirement. Who bears it: long-term holders expecting fee-driven value accrual bear it; validators bear it if fees do not grow. Indicators: share of validator rewards coming from fees vs inflation, onchain blob volume vs fee levels, and whether fee market parameters are adjusted through governance or left static while usage shifts.

We publish related measurement frameworks and breakdowns in our crypto research.

If you want a single sentence verdict: Celestia’s token design is legible and mostly deterministic within a version, but the system has already shown a willingness to re-parameterize issuance via upgrades, and the long-term health of TIA hinges on blob fees growing into the primary security budget.

If you want definitions before going deeper, start with our tokenomics FAQ.

If you want help stress-testing these assumptions for your protocol, see our tokenomics services.



This article is part of our Tokenomics Deep Dive series.