A discussion starting-point — facts in a chain; ▸ opens the granular data

From "how Cardano wins" to what an incentive manager can actually deliver — the facts, in one chain

Five links of one argument, each carrying only measured facts — what moves a chain's market cap, the demand foundation beneath it, what AlphaGrowth's "Cardano Prime" proposal offers, what AlphaGrowth's past programs did on-chain, and what Cardano's own incentive programs did on-chain. Key figures are visible on the page; each ▸ opens the granular table behind them. A starting-board for discussion — figures cited, opinions and vote-leanings left out.

Sources throughout: Dune (Analyst plan), DefiLlama, CoinMetrics, ADAtool's Cardano DBSync, comp.xyz governance, Tally, and the official proposal — June–July 2026.

①

What actually moves a chain's market cap

Across the assets ranked above ADA, transaction count, active addresses and TVL show almost no correlation with market cap (Pearson r(mcap, active addr) = −0.17; r(mcap, TVL) = −0.15). The assets that climbed each ran one of three engines — and the two fastest re-rated ~10× (Hyperliquid, on real buyback/burn revenue) and +650–1000% (Zcash, on a spot-ETF + supply-squeeze narrative) in ~18 months.

Own a dominant use-case

Payments (TRON, Stellar), privacy (Monero): real usage → real fees → a defensible cap.

Category-killer product + tokenomics

Hyperliquid: best-in-class product + real revenue → buyback/burn + airdrop loyalty.

Narrative wave + structural catalyst

Zcash: privacy revival + SEC clearance + ETF filing + halving supply squeeze.

A corollary sharpened in the current debate: what a chain can durably charge for matters twice — once as fuel for the two replicable engines (fees fund flagships and buybacks), and once as reduced structural sell pressure, because holders with on-chain avenues to earn on the native asset have fewer reasons to exit it. The first half is measured in ④–⑤; the second is directionally real but not measurable in advance, so this page treats it as upside rather than as part of any program's price. One further observation belongs here: a treasury whose ecosystem generates near-zero fees is not preserving value by saving — its purchasing power is levered to the chain's relevance, so the counterfactual to any growth spend is never neutral.

Sources: CoinGecko, CoinMetrics, DefiLlama (Jun 2026); reporting on Hyperliquid & Zcash.

②

The foundation beneath the two replicable engines: demand that stays

Two of those three engines (a revenue flagship, a capital-markets catalyst) rest on real DeFi demand — capital and users with a reason to arrive and stay. On that measure Cardano's rails are laid but nearly empty: USDCx (Circle's official USD) went live 27 Feb 2026, with LayerZero v2, Pyth and Dune — yet Cardano stablecoin supply is only ~$45–60M. The one durable stablecoin flip on record — Aptos, $111M→$1.88B — was paid BD work (BlackRock BUIDL + native USDT/USDC onboarding), not emissions.

$45–60M

Cardano stablecoin supply

$14.9B

Solana stablecoin supply

$3.9B

Arbitrum stablecoin supply

The rail is laid; almost nothing flows through it yet — so every growth program faces the same question: can it create demand that outlives its own funding? Source: DefiLlama (Jun 2026).

③

What AlphaGrowth's "Cardano Prime" proposal offers

AlphaGrowth proposes to manage a treasury-funded incentive + BD program to grow Cardano DeFi. Stated framing (AG, on X): "TVL is the headline, organic APR is the goal." Read against ④–⑤ below, four things define what is actually on the table:

1 · The honest unit is annual

Measurements and the published response both treat incentive TVL as management-dependent (‘a permanent operator, not a 12-month visitor’). ₳120M = year one of a continuing function; year 2 needs its own action; large onboardings (>$1M) are separate withdrawals — a serious multi-year commitment plausibly totals several hundred million ADA.

2 · …and that is the market price

Arbitrum ~$100M+ (STIP/LTIPP) · Optimism $100M+ (RetroPGF) · Avalanche $180M (Rush) · Solana Foundation: continuous · Base: Coinbase-subsidised. No top-tier DeFi chain spent zero — and spend guaranteed nothing either (④).

3 · The exit exists at scale — but is unproven

End-state: profitable dapps fund their own growth from fees and emissions — Ethereum apps generate ≈$271M/month in fees (DefiLlama, Jul 2026). But no L1-treasury-kickstarted program has completed that hand-off anywhere, and dapp emissions are themselves a treasury (the dilution moves to dapp token holders).

4 · Year one is a cancellable trial

Month-4 gate (~75% of budget), six return triggers, $4.64M fee cap. The month-6 falsification trigger (+6-month persistence reporting) is, in effect, the test of whether Cardano becomes the first measured hand-off — note its enforcement is a stated commitment, not a contract.

Net framing: with management-dependence now common ground, the live question is not ‘does incentive TVL stick on its own’ (measured below: no) but ‘fund a multi-year growth function — and can its cost hand off to dapp economics before the treasury tires of paying?’

$19.2M / ₳120M

Total ask

~72.5%

AG pay tied to verified TVL (at-risk)

~75%

Budget gated to Phase 3

$0.073

Spend per $1 TVL grown

▸Details: the seven budget lines + stated KPIs

Line	Amount	When / gating
Performance Fee Reserve	$4.64M	Phase 3 · gated · unearned returned
AG Fixed Advisory Fee	$1.76M	Monthly tranches · M5–12 gated
Ecosystem Grants	$5.6M	$3M Phase 2 + $3M Phase 3
LP Incentives	$4.32M	All Phase 3 · gated
Marketing & BD	$2.4M	$0.5M / $0.5M / $1.65M
Legal & Compliance	$0.16M	Phase 1 upfront
Independent Audit	$0.32M	Phase 2
TOTAL	$19.2M	₳120M · ~75% Phase-3 gated

The Performance Fee is calculated by a stated formula (declining 3%/2%/1% across TVL bands, capped at $264M growth). Per the proposal, the granular breakdown of the other envelopes is finalized in the subsequent contract. Stated KPIs: +$500M TVL · $10M revenue · 25k users

Source: official Cardano Prime proposal PDF (Jun 2026) + @alphagrowth1. Reference benchmark: Arbitrum STIP ≈ $0.085 per $1 TVL grown.

④

AlphaGrowth's past programs — what the on-chain data shows

AlphaGrowth's headline record: securing ecosystem grants (1.8M ARB from Arbitrum DAO; 515k OP from Optimism) and distributing them as APR incentives on Compound V3. We verified the whole cycle on Dune at address level; AlphaGrowth has published a point-by-point response (alphagrowth.io/cardano-prime-track-record-response). Its position, in three lines:

Accepts the numbers

+49% tenure TVL, the 12–18-month DiD window, the drawdown table, the $64.9M/$192M market sizes.

Contests the reading

Chain-beta, a management vacuum after handover, utilization over APY as the yardstick.

Names its own failures

The BTC mainnet market (late-recognized Babylon bet), wstETH (lost Instadapp to Aave), Ronin (no capital).

Below, each card runs the same three steps: what the data shows → the strongest other reading of the same curve → what survives both.

Program	Baseline→Peak	Multiple	Half-life	1yr retained	Now vs baseline
Compound V3 · Arbitrum (AG)	$56M→$213M	3.8×	463d	~78%	~1.09×
Compound V3 · Optimism (AG)	$5.8M→$56M	9.8×	—	—	~1.8×
Radiant · Arbitrum (emissions)	—	—	74d	~0%	0×
Velodrome · Optimism (emissions)	—	—	—	0.1% at end	—

Protocol-level TVL: AG's Compound programs were far stickier than typical pure-emission programs (still above their pre-grant baselines ~2 years on, where Radiant reverted to 0× in 74 days). The four cards below look underneath that protocol-level number.

Fact 1 — The effect was real: it beat every control, and retention led the incentive-program field

463d

half-life (vs Radiant 74d)

~78%

1yr retention (vs Radiant ~0%)

$1.41B→$2.10B

Compound total TVL over AG's tenure

At the TVL peak, each series vs its own grant-end level (=100) — the lift was specific to the granted market:

Compound-Arbitrum

230

Arbitrum chain TVL

Compound-Ethereum

Aave-Arbitrum

Compound-Base

The outperformance vs all four controls held through the grant and ~12 months after — a genuine market-specific effect, not just the rising tide. Independent OpenBlock analysis rated the program ~7.7× more TVL-efficient per ARB than a competitor's LTIPP program (AG's own figure: '$139 of TVL per $1 of ARB').

⇄ THE OTHER READING OF THE SAME CURVE

· A 12–18-month, controls-beating window is exactly what an incentive program is designed to buy — a window of competitiveness to build integrations and partnerships in.
· Baseline context: mid-2023 Compound was an ‘orphan protocol’ (Labs stepped back under legal pressure; no development, marketing or partnerships); tenure TVL went $1.41B→$2.10B.

What survives both: the readings part on whether the window was used to build things that outlive it — which is precisely what Facts 2–4 measure.

▸Granular: the 5-series difference-in-differences at every snapshot

Series	Start	End	Peak	+180d	+365d	Now
Compound-Arbitrumtreated (got grant)	45	100	230	188	179	64
Aave-Arbitrumcontrol — same chain	68	100	82	137	127	49
Arbitrum chain TVLcontrol — whole chain	86	100	93	102	93	42
Compound-Ethereumcontrol — same brand, no grant	104	100	85	134	154	72
Compound-Basecontrol — same brand, no grant	152	100	77	182	188	52

Dune + DefiLlama per-protocol/chain TVL. Each series indexed to its own value at grant-end = 100. The 'Now' column (64 vs controls 42–72) is read together with Fact 3.

Fact 2 — Underneath the flat total, the supplier base fully turned over

During the window: 27,506 wallets supplied (gross $140.7M)

At window end: 531 wallets held >$100 (net $31.7M)

Now (~2 years on): 217 wallets, $7.9M — 0.8% of the gross address count

Yet total net TVL stayed roughly flat ($60.5M → $58.0M) — because new suppliers replaced the departed. Who holds today's $58.0M:

61%

25%

13.7%

entered AFTER the grant window (950 addrs / $35.5M)entered beforeentered during (the grant cohort)

Top-5 whales account for 92% of what the grant cohort still holds. The replacement suppliers were themselves still incentivized: Compound's own COMP emissions (~$1–2M/yr) kept flowing until they tapered to ~zero in early 2026 — claimant counts (~500→125/mo) and TVL then fell together.

⇄ THE OTHER READING OF THE SAME CURVE

· 0.8% counts addresses, and large accounts rotate them — TVL-weighted retention would read differently (our one-hop rotation check: 2 wallets / $80.5k; CEX-mediated rotation unmeasured on both sides, so both statements can coexist).
· Whale concentration is the DeFi-lending norm; COMP emissions are set by Gauntlet (a governance-controlled tool, not an AG subsidy).
· One program-built relationship demonstrably persists: the OKX Earn route (possibly the industry's first exchange-DeFi channel, later mainstreamed by Morpho×Coinbase, Aave×Binance; zero distribution fees) — OKX DeFi/CeFi Earn remain major Compound depositors today.

What survives both: OKX is the clearest single data point for ‘programs buy relationships, not resting TVL’ — which is also an argument for judging on relationships and fees, not the TVL headline.

▸Granular: full decay table · USDT robustness · rotation check · era table

Snapshot	Net held	Addrs >$100	Note
Inflow during window (gross)	$140.7M	27,506	new suppliers
Window end (2024-07-16)	$31.7M	531	net retained
+90d	$10.4M	274
+180d	$15.5M	260	a few re-supplied
+365d	$7.7M	230
Now	$7.9M	217	≈25% of window-end · 0.8% of gross addrs

USDT robustness: re-running across USDC.e+USDC+USDT leaves the conclusion unchanged — 245 addresses hold >$100 (0.9%); the higher dollar figure ($42.0M) is 92% five whales that incidentally first-touched Compound in-window.

Rotation check: could exits be wallets ROTATING to fresh addresses? One-hop USDC linkage over 27,295 EOAs finds 2 rotated wallets / $80.5k (≈ +1.7% to retained TVL). A naive pass 'found' $7.1M, but 98% traced to the LI.FI Diamond router (0x1231deb6…) = new users bridging in, not rotated whales. Limit: exits via CEX/bridge that return from unlinked addresses are invisible to one-hop linkage.

Net TVL by entry era at each snapshot:

Net TVL by entry era	Window end	+90d	+180d	+365d	Now
Pre-grant entrants	$28.7M (590)	$24.0M (525)	$23.9M (534)	$19.8M (507)	$14.6M (470)
Grant-era entrants	$31.7M (531)	$10.4M (274)	$15.5M (260)	$7.7M (230)	$7.9M (217)
Post-grant entrants	$0M (0)	$7.2M (242)	$18.9M (663)	$35.0M (825)	$35.5M (950)
TOTAL net ($M)	$60.5M	$41.6M	$58.3M	$62.5M	$58.0M

Dune — Compound V3 Arbitrum, USDC.e+USDC base markets, net supply (Σsupply−Σwithdraw, floored ≥0 per address). "Era" = the address's first-ever supply date. Invariant check: withdrawals attributed to the Comet event's 'to' field make positive net balances sum to the live market's ~$58M. Split today: post 61% / pre 25% / grant 13.7%.

Fact 3 — Yields: ≈8–13% with incentives → ≈2.6% after; the TVL premium mean-reverted

≈8–13%

grant-era all-in yield (base+COMP+ARB)

≈2.6%

now — base only, rewards paused

≈2.54%

Aave-Arbitrum (never incentivized)

TVL vs its own level at grant-end (=100), Compound-Arbitrum:

start

100

end

230

peak

188

+180d

179

+365d

now

The outperformance of Fact 1 lasted ~12–18 months; as rewards tapered and macro conditions dominated, the premium mean-reverted toward the sister-chain baseline, and the organic base APY settled at essentially the never-incentivized Aave-Arbitrum rate.

⇄ THE OTHER READING OF THE SAME CURVE

· Wrong yardstick: single-market base APY is designed by risk managers (Gauntlet/LlamaRisk) to land ~3–4% even at optimal ~90% utilization — converging to the benchmark IS the healthy end state. The better lens: utilization, 87–90% during tenure vs ~78% today.
· Attractive organic APR comes from structured-product stacking (CDP / LST-LRT loops, >10% net at peak) — which a chain-level program can compose more freely than one protocol.
· The reversion itself: chain-beta + a management vacuum (Aave/Morpho also lost most L2 TVL; Dencun undercut the L2 pitch; the Compound Foundation pivoted its $20M budget to V4) — counterexample where the chain held: the USDT mainnet market, $192M today.

What survives both: the DiD design already nets out chain-beta, so the vacuum argument carries the weight — and it concedes the central premise measured here: this TVL holds only while someone actively manages it.

▸Granular: monthly APY table (base / +COMP / +ARB)

Month	Base	+COMP	+ARB (est.)	Phase
2024-04	6.5%	2.0%	≈3–5pt	window opens
2024-06	6.8%	1.4%	≈3–5pt	ARB distribution ramps
2024-07	3.2%	0.7%	≈3–5pt	TVL peak
2024-09	3.7%	0.9%	—	distribution ends
2024-12	11.0%	1.6%	—	post-grant rate spike (utilization)
2025-06	3.3%	1.1%	—	program over; campaigns only
2025-12	3.0%	0.8%	—	COMP rewards decaying
2026-07	2.6%	0.0%	—	rewards paused — base only

base+COMP: DefiLlama daily history (monthly means). ARB boost estimated from ~$1.4–2.7M ARB over ~3–4 months on $90–260M TVL (off-contract, not in yield feeds).

Fact 4 — The demonstrated skill is winning external money; on Cardano the funding structure differs

User incentives — other ecosystems' tokens, won competitively

1.8M ARB (Arbitrum DAO — the largest grant of its LTIPP cycle) + 515K OP (Optimism) + MNT/Ronin/Sky ≈ $2.7M + $1.2M + $2M

AG's service fee — Compound's treasury

77,208 COMP ≈ $4.25M（2024）＋$340K（2-month interim 2025）

In the Compound programs the incentive tokens came from Arbitrum/Optimism/Mantle DAOs; Compound paid only the service fee. Under Cardano Prime, Cardano's treasury (₳120M) is both the incentive source and the fee payer. Sequence facts: the 2025 renewal was rejected on-chain (#416: For 100.3K vs Against 517K), the reduced V4 also failed, a 2-month interim followed, and in Jun 2026 the DAO deprecated 2 of the 4 AG-launched chains (Linea, Mantle).

⇄ THE OTHER READING — AND ITS HEAVIEST SENTENCE

· The reversion is ‘a property of what came after’ — a management vacuum — and TVL needs a permanent operator, not a 12-month visitor.
· PRIME is built around that admission: 30-day fee holds, a declining rate under the $4.64M cap, a months-7–12 taper with cliff-vesting and vault products, the month-4 gate (~75%), six return triggers, the month-6 falsification trigger, 6-month persistence reporting.

What survives both: taken seriously, the admission reframes the ask itself — a permanent function funded year by year (③) — and sharpens the structural difference: at Compound another DAO paid the incentives; on Cardano the treasury funds both sides until (and unless) the cost hands off to profitable dapps.

▸Granular: the full money map + program timeline + monthly cohort flows

What	Amount	Funded by	Period / status
ARB user incentives (LTIPP)	1.8M ARB ≈ $2.7M	Arbitrum DAO	2024 Q2–Q3 (3-month mandate)
OP user incentives	515K OP ≈ $1.2M	Optimism	2024-05 →
MNT / Ronin / Sky	≈$1M + $1M + rolling	Mantle / Ronin / Sky	2025→ (Linea/Mantle comets deprecated 2026)
Ongoing COMP rewards (protocol)	≈$1–2M/yr	Compound treasury	decayed → top-ups PAUSED 2026
AG service fee 2024	77,208 COMP ≈ $4.25M	Compound treasury	2024-05 → 2025-04
AG 2025 ask (V4)	127,426 COMP ≈ $5.35M	Compound treasury	NOT passed (V3 #416 rejected; V4 failed)
AG interim extension	$340K USDC	Compound treasury	2025-05 → 07 (then program ends)

2023-12AG engagement with Compound begins

2024-03Wins 1.8M ARB from Arbitrum DAO (asked 5M, cut to 1.8M; largest of the cycle)

2024-05CGP-2024 passes: 77,208 COMP (≈$4.25M) AG service budget; KPIs +$500M TVL / $10M rev / 25k users

2024-07ARB distribution peaks — grant-gen inflow $140.4M in one month; AG claim TVL $92M→$266M

2024-08–12Window ends → grant cohort headcount 531→217; post-grant suppliers replace it

2025-032025 renewal V3 (#416) rejected on-chain: For 100.3K vs Against 517K

2025-04Reduced V4 ($5.35M) also fails; Compound Foundation formed mid-vote

2025-05–072-month interim ($340K USDC), then the program ends; per-campaign work only after

2026-06DAO deprecates Linea & Mantle comets (2 of 4 AG-launched chains); COMP top-ups paused

Monthly net flow ($M) by generation — grant-era (blue) and post-grant (green); bar length ∝ |net|, red = outflow. ◂ marks the ARB window.

Month	Grant net	Post net	New addrs
2024-01	0.0	0.0	426
2024-02	0.0	0.0	1,266
2024-03	-0.6	-1.4	3,112
2024-04 ◂	-2.8	-1.0	13,492
2024-05 ◂	2.8	-1.0	9,590
2024-06 ◂	4.1	-0.7	5,500
2024-07 ◂	123	1.3	3,338
2024-08	17.7	-0.1	409
2024-09	8.6	-1.1	336
2024-10	5.3	2.6	438
2024-11	8.8	3.6	485
2024-12	11.8	0.0	576
2025-01	15.5	-0.2	455
2025-02	8.4	-0.8	409
2025-03	4.7	6.9	380
2025-04	4.1	6.9	218
2025-05	2.0	-4.6	212
2025-06	0.6	2.6	204
2025-07	4.9	3.5	444
2025-08	0.6	2.5	297
2025-09	1.4	0.9	334
2025-10	-0.2	-0.5	282
2025-11	2.4	6.7	287
2025-12	0.0	2.3	137
2026-01	0.0	-1.2	173
2026-02	0.3	6.0	141
2026-03	0.5	-2.0	293
2026-04	0.0	1.7	143
2026-05	0.4	-0.1	112
2026-06	0.0	-0.3	93
2026-07	0.0	-0.4	36

Apr–Jun 2024: thousands of small wallets enter. Jul 2024: whale money lands (grant-gen net +$122.9M in one month). After the window the grant generation's headcount collapses while its net flow stays positive into 2025 — the handful of whales (top-5 = 92% of the cohort balance) still cycling money. From late 2025 it flatlines to ±0.

comp.xyz governance (grant post 5182, CGP-2024, V4 6563, interim 6677), Tally (#416), Dune (decoded Comet Supply/Withdraw events; generation by first-ever supply date). Raw flows don't reconcile 1:1 with the ≥0-floored net-balance table in Fact 2. Verified 2026-07-09.

Counter-readings above are from AlphaGrowth's published response: alphagrowth.io/cardano-prime-track-record-response \u2197

Caveat carried from source: grants ≠ sole cause (2024–25 was a broad DeFi recovery; Compound is a bluechip with organic demand).

⑤

Cardano's own incentive programs — what the on-chain data shows

The same lifespan lens on 19 Cardano-native protocols (ADA-denominated TVL, so ADA price swings are stripped out; Tx from ADAtool's DBSync). Median post-peak half-life is 55 days (range 4–496). The consistent split is not "did TVL rise" but "did usage outlast the TVL":

Usage outlasted the TVL (durable)

Minswap (Tx persistence ~91%), Djed (~100%), Indigo (~84%), WingRiders (>100%). Real activity stuck even as mercenary TVL left.

The TVL was the product (rented)

SundaeSwap V2 (Tx ~12%, now 1.7% of peak), Splash (~4%), Lenfi (now 0.9%), VyFinance (~20%). Excess kept ≈ 0% — TVL reverted to the pre-incentive baseline once rewards faded.

▸Granular: 12 protocols — half-life, 1yr retention, excess kept, Tx persistence

Protocol	Type	Peak (ADA)	Half-life	1yr ret	Excess kept	Tx persist
Minswap DEX	DEX	190M	496d	78%	39%	91%
Djed Stablecoin	Stablecoin	46M	392d	73%	18%	100%
Indigo	CDP	184M	188d	38%	1%	84%
WingRiders	DEX	116M	7d	28%	18%	122%
Liqwid	Lending	176M	—	69%	60%	30%
SundaeSwap V2	DEX	151M	24d	14%	0%	12%
Splash Protocol	DEX	79M	4d	43%	0%	4%
Lenfi	Lending	74M	4d	6%	—	5%
VyFinance	DEX	49M	110d	19%	0%	20%
Optim Finance	Staking	51M	108d	52%	0%	—
Astarter ISPO	Staking	45M	99d	8%	0%	—
Levvy for Tokens	Lending	10M	58d	35%	0%	—

"Excess kept" = of the incentive-driven rise (peak − pre-incentive baseline), how much remains today. "Tx persist" = transaction rate ~1yr after peak vs near-peak — the clearest tell of whether a real user base formed. 19 protocols analyzed; the 12 with the largest peaks shown.

DefiLlama (ADA-denominated) + ADAtool DBSync (through Jun 2026). Caveat: TVL ≠ a single incentive's effect; directional evidence, not precise attribution.

Facts compiled from Dune (Analyst plan), DefiLlama, CoinMetrics community data, ADAtool's Cardano DBSync, comp.xyz governance, Tally, and the official AlphaGrowth Cardano Prime proposal, June–July 2026. This page states figures and their sources only; it makes no recommendation and is not investment or voting advice.