Sanctum Crypto: What the Infinity Pool Actually Holds (2026)

Sanctum crypto is the liquid staking infrastructure behind more Solana LSTs than any other protocol — and its flagship product, the Infinity pool, is marketed everywhere as a diversified basket of liquid staking tokens. We decoded that pool directly from mainnet on 25 October 2026, at slot 461,301,916 in epoch 1067. It holds 1,929,121 SOL. 94.40% of it is a single token, and 99.99968% of that token's entire supply sits in one account: Infinity's own reserve.

This is a rewrite, not a refresh of marketing copy. Every number below comes from an RPC call you can repeat, a Jupiter quote you can re-run, or Sanctum's own documentation. The previous version of this page quoted "a 6–8% staking APY" as the opportunity cost of parking SOL instead of trading it. That band is above Solana's own ceiling: the network pays 5.3602% APY at a 0% commission validator this epoch, and the engine underneath 94% of Infinity returned 4.8819% annualised last epoch. Here is what the chain actually says, and what it means if you are deciding between passive LST yield and a Solana copy trading bot.

Direct answer: Sanctum is a Solana liquid staking infrastructure company, not a single staking token. It builds the plumbing that lets other teams issue branded liquid staking tokens, and it runs the shared liquidity pool those tokens swap through. Its own documentation describes five distinct products:

• Infinity (INF). A multi-LST liquidity pool. Deposit SOL or a whitelisted LST and receive INF, an LP token whose yield is the weighted average of the staking yields inside the pool plus the swap fees the pool earns. This is the product most people mean when they say "Sanctum crypto".
• Sanctum LSTs. A fork of the Solana Labs stake pool program used to launch branded tokens. Sanctum's docs claim 1,361+ LSTs launched, including jupSOL (Jupiter), bbSOL (Bybit), cdcSOL (crypto.com), dSOL (Drift) and dfdvSOL for the DeFi Development Corp treasury.
• The Reserve. Backstop liquidity for instant unstaking, documented at "over 600k SOL" and explicitly closed to retail deposits. Its fee is dynamic and ranges from 8 to 800 basis points depending on how drained the pool is.
• swSOL. A wrapped-SOL replacement launched in September 2026 that stakes part of the underlying balance instead of leaving it idle.
• Gateway. Transaction delivery infrastructure, unrelated to staking.

The protocol token is SANC, renamed from CLOUD in September 2026 with no migration — the mint address is unchanged at CLoUDKc4Ane7HeQcPpE3YHnznRxhMimJ4MyaUqyHFzAu. On 6 October 2026 Sanctum burned the 250M-token community reserve; the mint reports a supply of 740,669,448 SANC today, down from the original one billion. Sanctum's own governance post gave the reason for the rename plainly: searching "CLOUD" returned Google Cloud and Alibaba Cloud, so the ticker was invisible.

Sanctum is not a trading terminal, a copy trading platform, or a sniper. Active execution still routes through Jupiter, Raydium or Orca, and the tools in our Solana trading platform comparison.

All figures below were read from Solana mainnet on 25 October 2026 during epoch 1067, except where a source is named.

Measurement	Value	How we got it
Infinity pool size	1,929,121.47 SOL	Pool state account, total_sol_value
Largest single holding	94.40% in one token	LST state list, 199 entries decoded
1 INF intrinsic value	1.458977027 SOL	Pool SOL value ÷ INF supply
Measured exit fee	20.58 bps, flat	Jupiter quote, Infinity-only route
Instant SOL exit capacity	≈16,000 INF (23,296 SOL)	Quote ladder, 1 → 100,000 INF
Underlying engine's last-epoch yield	4.8819% APY	Stake pool balance snapshots
Network ceiling at 0% commission	5.3602% APY	Inflation ÷ staked share, compounded
Validators behind 94% of the pool	Two	Stake pool validator list

Nothing here says Sanctum is broken. The engineering is real, the fee is small, and the two validators are honest operators charging nothing. What the data says is that the description of the product in most Sanctum coverage — including the earlier version of this page — does not match what the pool holds.

The Infinity pool's S Controller program is 5ocnV1qiCgaQR8Jb8xWnVbApfaygJ8tNoZfgPwsgx9kx. It owns exactly two accounts: the pool state, which carries the total SOL value, and an LST state list with one 80-byte entry per whitelisted token. Decoding the list gives you the entire composition of the pool in one call, and the per-entry SOL values sum to the pool total to the lamport — so there is no interpretation to argue with.

There are 199 whitelisted slots. 156 carry a non-zero balance, 43 are empty, and 6 have deposits disabled. Six tokens account for 99.71% of the pool:

Token	SOL value	Share of pool
sctmSOL (Sanctum Automated)	1,821,154.83	94.40%
haSOL	28,108.60	1.46%
jupSOL (Jupiter)	24,177.90	1.25%
wSOL (unstaked SOL)	23,388.96	1.21%
bpSOL	17,319.66	0.90%
dSOL (Drift)	9,361.25	0.49%

The three LSTs every Sanctum explainer names as the pool's backbone are, in practice, absent. mSOL holds exactly 0.000 SOL in the pool. bSOL holds 0.039 SOL. jitoSOL holds 175.93 SOL — 0.009% of the pool. The earlier version of this article told you Infinity "aggregates liquidity across Marinade mSOL, Jito jitoSOL, Blaze bSOL". That was wrong, and one RPC call shows it.

The token holding 94.40% of the pool is sctmAy9zFfZznZUWxEMgGtF6PHZ3gwoiCGcfwRua1AZ, which wallets display as "(Sanctum Automated) LSTQq". It came from Sanctum's Automated LST programme — the February 2025 launch that minted an LST for every Solana validator. Sanctum has since deprecated that programme; both announcement posts now carry a notice saying they are kept for historical reference only.

Two facts make it clear this is an internal instrument rather than a token anyone holds:

• Infinity holds essentially all of it. Total sctmSOL supply is 1,626,116.03. The pool's reserve account 6wCsF4qacwRF9Wc7x2nrz4sbfCwbWMen94LAEMQA8mf8 holds 1,626,110.76 of it — 99.99968% of the entire supply.
• Only Sanctum can mint or redeem it with SOL. The stake pool sets both sol_deposit_authority and sol_withdraw_authority to 6utAm5grcWt4jCtLHPNWJQXmakQzC35dYyqMRnHmNr79, which is the same key as Infinity's own rebalance_authority. The pool's manager is also the same key as Infinity's admin.

So the honest description of Infinity is not "a basket of liquid staking tokens". It is a stake pool with a thin skin of partner LSTs on top: roughly 94% directly-managed stake, about 5% third-party LSTs, and 1.21% unstaked SOL for instant exits. That is not a scandal — a concentrated pool with a 0% fee and 0% commission validators can be a perfectly good product. It is just a completely different risk profile from "diversified across dozens of protocols", which is what the marketing implies and what most readers will assume.

Follow the stake one more level down. The sctmSOL stake pool's validator list has room for ten validators, carries three entries, and funds exactly two:

Vote account	Stake from this pool	Commission	Host / location
LSTmLs1…AY6Ys	1,512,519.74 SOL (83.05%)	0%	Latitude.sh, Frankfurt
685Njwv…XFiTd	308,639.61 SOL (16.95%)	0%	OVH, Gravelines

Both run Agave 4.3.2, both are Jito-enabled at 0 basis points, both report 100% uptime, and the larger one shows a 0.2199% skip rate against 0% for the smaller. Stakewiz puts their total APY at 4.95% and 4.99%, which stake-weights to 4.9568%. Running your own client is a different conversation — see our breakdown of what Firedancer adoption changed for validator performance.

Two observations the pool's size makes unavoidable. First, Infinity supplies 82.74% of the larger validator's entire 1,828,123 SOL stake and 99.97% of the smaller one's — these are not independent operators that happen to be chosen, they exist to serve the pool. Second, zero commission is genuinely good for holders: against a network where the median validator charges 5% and the stake-weighted average is 26.77% (65 validators charge 100% and hold 104.5M SOL between them), paying nothing to the validator is the single biggest lever on net yield.

The flip side is correlation. Two validators in two European datacentres is two machines, not a diversified validator set, and a double-signing or prolonged-downtime event at the larger one touches 83% of the stake behind 94% of the pool. Compare that with Marinade's stake auction, which funded 66 validators at its last scoring run.

Start with the ceiling, because it bounds everything else. Solana's inflation rate at epoch 1067 is 3.5873%, applied to a total supply of 636,395,915 SOL. Activated stake across 665 active and 25 delinquent validators is 437,191,006 SOL, so 68.698% of the supply is staked. Stakers therefore split the inflation among themselves at a gross rate of 3.5873 ÷ 0.68698 = 5.2218% APR. Our Solana staking guide walks through why that division is the whole derivation.

Compounding needs an epoch length, and this is where most published figures are stale. Epoch 1066 ran from 23 October 16:31:53 UTC to 24 October 17:59:46 UTC — 91,673 seconds, or 25.465 hours, at 212.21 ms per slot. That is 344.24 epochs per year, not the 160-odd that a two-day epoch implies. Compounded at that frequency:

• 5.3602% APY at a 0% commission validator — the absolute ceiling for staking SOL today.
• 5.0855% APY at the median 5% validator.
• 4.8115% APY at a 10% validator.

Now measure Sanctum's engine. The sctmSOL stake pool stores its previous-epoch balance alongside its current one, so one account read gives you a realised epoch return with no modelling. Supply was unchanged at 1,626,116.029055152 tokens across both snapshots; total lamports went from 1,820,908.591925656 SOL to 1,821,160.736380846 SOL. That is a price move from 1.119790076 to 1.119945135 SOL per token, a one-epoch return of 0.013847%, which annualises to 4.7668% APR and 4.8819% APY. It agrees with Stakewiz's 4.9568% stake-weighted estimate for the same two validators to within seven basis points.

Sanctum's own surfaces publish higher numbers, and both are defensible: the homepage shows 5.26% APY, and the INF page shows 6.04% "Last Epoch's APY" alongside 1.87M SOL staked and 41,573 holders. INF legitimately earns more than its underlying LSTs because it also collects swap fees — Sanctum's docs cite a 26.12% epoch APY during the October 2025 flash crash, when redemption demand spiked. But a single-epoch figure driven by trading volume is not a yield you can plan around, and the docs themselves note the front page uses a 10-epoch rolling mean precisely because individual epochs are noisy.

The conclusion the earlier version of this page got wrong: a "6–8% staking APY" is not available on Solana right now at any commission level. The staking component of INF's return is high-4s. Anything above the 5.3602% ceiling is swap-fee income, which is volatile, mean-reverting, and the first thing to disappear in a quiet market.

Sanctum publishes its fees in two places that do not match, which matters because the gap is a factor of two.

• The Infinity FAQ says: no deposit fee, a flat 10 bps to withdraw liquidity, and a flat 8 bps per swap.
• The Infinity technical documentation says: withdrawing liquidity "incurs a 20bps fee", and trading fees "vary depending on each LST and include both input and output fees".

The chain settles it. Quoting INF → SOL through Infinity's own AMM and nothing else, 100 INF returns 145.597527533 SOL and 10,000 INF returns 14,559.752759 SOL. Against an intrinsic value of 1.458977027 SOL per INF, both are 0.2058% below fair value — the technical doc's 20 bps, not the FAQ's 10. If you sized an exit off the FAQ you budgeted half the real cost.

The rest of the stack, from Sanctum's own documentation:

• Sanctum LST yield fee: 10% of staking yield by default, split equally between Sanctum and the partner who launched the token, plus a 10 bps withdrawal fee. Note that the older Automated LST documentation quotes 5% with half to the validator and a 0.10% withdrawal fee — another pair of numbers that do not agree.
• Launching an LST: 100 SOL one-time, which Sanctum commits back into the new token for its first three months.
• The Reserve: 8 to 800 bps, dynamic with utilisation, and not open to retail.
• Infinity's internal stake pool charges nothing. We decoded the sctmSOL pool's fee fields directly: epoch fee 0/1000, stake deposit fee 0, stake withdrawal fee 0/100, SOL deposit fee 0, SOL withdrawal fee 0. Every fee on the pool holding 94% of Infinity is zero, because Sanctum takes its cut at the Infinity layer instead.
• Network fees. Deposits, swaps and redemptions are ordinary Solana transactions costing fractions of a cent, though priority fees rise exactly when redemption demand does — see Solana transaction fees and Solana MEV.

One pleasant surprise in the other direction: buying INF rather than minting it is currently cheaper than intrinsic. A 10 SOL market buy returns 6.866044 INF against an intrinsic 6.854118 — 0.174% more INF than the pool's own exchange rate, because INF trades at a small persistent discount on secondary venues. Entry and exit are both roughly 20 bps around a market price that sits just under book value.

A flat fee tells you nothing about size. We ran a quote ladder from 1 to 100,000 INF and the result is a cliff, not a curve:

Size	SOL received	Cost vs intrinsic
1,000 INF	1,455.98	0.206%
10,000 INF	14,559.75	0.206%
16,000 INF	23,295.60	0.206%
17,000 INF	24,322.97	1.934%
18,000 INF	24,788.55	5.609%
20,000 INF	24,766.23	15.13%
50,000 INF	25,410.00	65.17%

Read the bottom rows carefully: at 50,000 and 100,000 INF the output barely moves (25,410.00 and 25,410.67 SOL). The route is saturated — past roughly 17,000 INF you are burning tokens for almost no additional SOL. Jupiter's own routing shows the transition: Infinity takes 100% of the fill up to 16,000 INF, then starts peeling slices to Saber at 17,000.

The cause is sitting in the composition table. The pool's unstaked SOL reserve is 23,388.96 SOL, and the exit saturates at 23,296 SOL. Instant exit capacity is the SOL reserve, nothing more. At $101.54 per SOL that is about $2.37M of instant liquidity against a 1.93M SOL pool — 1.21%. Sanctum's partner documentation describes "~620k SOL (and growing) liquidity available for instant withdrawals"; whatever that figure counts, it is not the reserve a redemption actually draws from today.

Practical consequence: INF is fine for retail position sizes and bad for anything institutional that needs same-block exit. Anyone holding more than ~16,000 INF has to exit over several epochs, route through the underlying LSTs, or accept a double-digit haircut. If you use INF as collateral on Kamino, Marginfi or Loopscale, that is the number your liquidation math should assume, not the pool's headline TVL.

Sanctum's documentation says "each epoch (approximately 2 days)" and its LST onboarding page says a few epochs is "approximately 2-3 days". Both were accurate once. Measured across epoch 1066, an epoch is 25.465 hours, and the slot time behind it is 212.21 ms rather than the 400 ms every older document assumes.

This is a consequence of the Alpenglow upgrade, which activated on mainnet in October 2026 and roughly halved effective slot times. It changes three things for anyone evaluating Sanctum crypto:

• Compounding is faster. 344 epochs a year instead of ~180 is why the 0%-commission APY (5.3602%) sits 14 basis points above the simple APR (5.2218%).
• Delayed unstaking is cheap again. Native stake deactivation completes at the next epoch boundary, which is now about a day — not the "2–3 epoch cooldown" language still in circulation. If you are not in a hurry, waiting out an epoch beats paying any instant-exit fee.
• Every APY quoted from a 2-day epoch assumption is understated. Check whether a dashboard annualises with a measured epoch length or a hardcoded constant before you compare two numbers.

All three prices below were derived from the protocols' own on-chain state on the same day, which also exposes a trap: Sanctum's public pricing endpoint is stale.

Dimension	Sanctum INF	Marinade mSOL	Jito jitoSOL
Pool size	1.93M SOL	2.26M SOL backing mSOL	10.42M SOL
Token price today	1.458977 SOL	1.414087 SOL	1.308043 SOL
Validator spread	2 (94% of pool)	66 funded by auction	Large, Jito-enabled set
Exit fee	20.58 bps measured	0.2% delayed, or a Jupiter route	DEX swap or delayed unstake
Instant exit depth	~23,300 SOL	~22,000 mSOL on Jupiter	Deepest of the three
Best for	LST-to-LST routing, swap-fee upside	Widest DeFi integration	Scale plus MEV tip yield

Jito's pool is 5.4× Infinity's. Decoding its stake pool account gives 10,422,566 SOL backing 7,968,062 jitoSOL, a price of 1.308042722 SOL. Marinade's auction-funded validator set and the cost of exiting mSOL are covered in detail in our Marinade review.

The trap worth repeating: Sanctum's public pricing endpoint returns values that are materially wrong. On the day we measured, extra-api.sanctum.so priced INF at 1.407837461 SOL (3.505% below the real 1.458977027) and jitoSOL at 1.275866054 (2.460% below the real 1.308042722), and returned an APY of exactly 0.0 for all six LSTs we queried. Its mSOL figure of 1.376648063 is 2.6% under Marinade's own published 1.4140872403 and byte-identical to the value the same endpoint served three days earlier, so it is frozen rather than merely lagging. Derive LST prices from the stake pool accounts, and treat any dashboard built on that endpoint as decorative. If you track positions programmatically, read the pool state yourself — our Helius API guide covers the RPC side.

Liquid staking is not a savings account. Here is what is actually measurable, ranked by how much it should change your sizing.

• Concentration, not diversification. 94.40% of the pool is one token, 100% of that token is held by the pool itself, and its stake sits with two validators in two European datacentres. Any model that treats INF as spread across dozens of protocols is wrong by construction.
• Exit capacity is 1.21% of the pool. Instant redemption draws on 23,389 SOL. In a market-wide rush for SOL, that reserve empties before most holders reach it, and the remainder waits an epoch.
• The program is upgradeable and unverified. The S Controller was last deployed at slot 411,884,877 — 8 April 2026 — and its upgrade authority 47SND7bGKvNXrqfP1bjsLCbwTgZhFBzAgmZ42QSkRScz is still set, so the program is not frozen. The verifiable-build registry at verify.osec.io reports it as not verified, with the last verification attempt logged 24 April 2026 against github.com/igneous-labs/S. The source is public; no reproducible build matching the deployed bytecode has been published.
• Audit coverage is real but dated. Sanctum's docs list three independent audits of Infinity (Neodyme, Ottersec, Sec3) and five of the stake pool lineage it forked (Halborn, Kudelski, Ottersec, Neodyme, Quantstamp). The stake pool programs sit under an 11-member multisig that Sanctum says includes Jupiter, Jito, Sol Strategies and Solblaze. Nothing public covers the April 2026 S Controller deployment specifically.
• Documentation drift. Two fee figures that differ by 2×, an epoch length that is half what the docs say, a deprecated product holding 94% of the flagship pool, and a pricing API returning zeros. None of these is a loss on its own; together they mean you cannot size a position from the docs alone.
• Slashing and validator risk. With two validators, the usual "diversification reduces slashing exposure" argument does not apply here. Both currently report 100% uptime and low skip rates, which is the mitigation that is actually working.

Now the comparison can be made with a real number instead of a made-up band.

• Sanctum INF targets the staking yield of its underlying pool plus swap fees. The staking component measured 4.8819% APY last epoch and is capped by the network's 5.3602% ceiling; the swap-fee component is genuinely additive but volatile. On a $10,000 stack, 5% is about $500 a year, minus roughly 20 bps to get out. Downside is SOL price movement plus the protocol risks above, with no liquidation unless you borrow against the position.
• Copy trading targets active spot PnL by mirroring wallets whose history is public and auditable. Return depends entirely on wallet selection and market conditions. A well-chosen wallet in a memecoin cycle can beat a year of staking in a session; a bad one loses capital faster than any yield offsets. uwuu.ai charges a performance-based fee — you pay only when you profit — and execution is non-custodial with sub-400ms mirroring.

Treating these as alternatives is the mistake. The correct framing is that $500 a year on $10,000 is the hurdle rate, not the strategy. If a trading allocation cannot clear 5% annually after fees and mistakes, it belongs in an LST. If it can, the staking yield is the floor under the part of your stack you are not deploying.

The split most Solana traders actually run: a core SOL allocation in an LST for baseline yield and collateral optionality on Jupiter Lend or Kamino, plus a separate trading allocation sized to what you can lose. Use the verified on-chain leaderboard to pick wallets with auditable PnL, set slippage limits and token filters, and keep the trading capital liquid — a wallet you are copying does not wait for your epoch boundary. Our copy trading profitability breakdown sets realistic expectations, and how to copy trade on Solana covers the setup.

Use Sanctum if: you hold several LSTs and want cheap routing between them, since 20 bps flat is competitive and the Infinity route is the deepest LST-to-LST venue on Solana; you want swap-fee exposure on top of staking yield and accept that it is lumpy; your position is comfortably under the ~16,000 INF instant-exit ceiling; or you are a team launching a branded LST and the 100 SOL plus 10% yield fee is cheaper than running a stake pool yourself.

Skip Sanctum if: you chose it specifically for diversification, because 94% of the pool is one token delegated to two validators; you need same-day exit on more than ~$2.4M; you were budgeting 6–8% yield, which does not exist on Solana at any commission level today; or your edge is active trading, in which case locked capital is the cost, not the product.

Sanctum crypto is infrastructure — competent, open-source, cheaply priced, and considerably more concentrated than it looks. Pair it with execution tools, a wallet tracker, and a wallet you control (Phantom or Solflare both connect to the app) for a complete Solana workflow.

What is Sanctum crypto?

Sanctum is a Solana liquid staking infrastructure protocol. Its flagship product is Infinity, a shared liquidity pool whose LP token is INF, and it also runs the stake pool infrastructure behind more than 1,300 branded LSTs including jupSOL, bbSOL and dSOL. Its protocol token is SANC, renamed from CLOUD in September 2026 on the same mint address.

What is Sanctum Infinity and what does INF hold?

Infinity is a multi-LST pool; INF is the token you receive for depositing into it. As of epoch 1067 the pool holds 1,929,121 SOL of value, of which 94.40% is a single Sanctum-managed token called sctmSOL, 1.21% is unstaked SOL, and the rest is spread across jupSOL, haSOL, bpSOL, dSOL and roughly 150 smaller tokens. mSOL and bSOL balances are effectively zero.

What is the real Sanctum APY in 2026?

The stake pool holding 94% of Infinity returned 0.013847% in epoch 1066, which annualises to 4.8819% APY. Solana's ceiling at a 0% commission validator is 5.3602% APY this epoch. Sanctum's own pages show 5.26% and 6.04%; the gap above the ceiling is swap-fee income, which varies a lot epoch to epoch.

What fees does Sanctum charge to exit INF?

Measured through Infinity's own route, exiting INF to SOL costs 20.58 basis points, flat, at every size up to about 16,000 INF. That matches the 20 bps in Sanctum's technical documentation rather than the 10 bps quoted in its FAQ. Depositing is free, and buying INF on the open market is currently about 0.17% cheaper than minting.

How much INF can I sell instantly?

Roughly 16,000 INF, about 23,300 SOL or $2.4M, before the route saturates. At 17,000 INF the cost jumps to 1.93%, at 20,000 it is 15.1%, and beyond that extra tokens return almost no extra SOL. The limit is the pool's unstaked SOL reserve of 23,389 SOL, which is 1.21% of its total value.

Is Sanctum staking safe?

Infinity has three published audits (Neodyme, Ottersec, Sec3) and no public exploit. The honest caveats are that the program is upgradeable and reported as unverified by the verifiable-build registry, that 94% of the pool sits with two validators, and that instant exit capacity is about 1.2% of the pool. Size positions accordingly rather than treating an LST as cash.

Can Sanctum replace a Solana copy trading bot?

No — they are different instruments. Sanctum pays a network yield in the high 4% range on SOL you are not using. Copy trading mirrors active spot trades from wallets with auditable on-chain history. The useful way to use both is to treat the staking yield as the hurdle rate your trading allocation has to beat.