What is Solana? How the high-speed chain works, what SOL earns and why institutions use it

Who this is for: Allocators, treasurers, founders and policy staff who need to understand how Solana works, what SOL actually earns, where the chain’s institutional traction is real and where its risks sit, so they can read a validator dashboard, an ETF fact sheet or a tokenized-asset report without taking the headline numbers on faith.

Solana launched its mainnet in March 2020 with a simple and contested thesis: a blockchain can be fast and cheap enough for trading, payments and consumer apps without splitting activity across separate chains. Ethereum chose to keep its base layer conservative and push most activity to rollups, a design explained in our guide to Ethereum layer 2s. Solana chose to make the base layer itself as fast as the hardware allows. That decision explains nearly everything about the network, from why it became the home of memecoin trading and tokenized equities to why it has been criticized for demanding validator hardware and for its outage record.

The question matters more in 2026 than it did in 2021 because the people using Solana have changed. In the first week of October 2026 the Solana Foundation released an open-source delivery-versus-payment settlement program built with advisory input from JPMorgan, the Bank of North Dakota launched a dollar stablecoin on Solana for more than 90 banks and credit unions, and the SEC had just issued a five-year exemptive order that lets tokenized shares trade on permissionless chains, a market where Solana says it hosts more than half of 2026 volume. None of that would have been plausible without the ETF approvals of October 2025 and the two years of uninterrupted uptime behind them.

This guide explains how the chain is built, how SOL issuance and fees flow to validators and stakers, why the network went down in the past and what is being done about it, where institutions actually show up on-chain, and how to evaluate the whole picture. It includes a worked example of a mid-size validator’s income statement and a timeline from the 2020 launch to the Alpenglow testnet.

Solana by the numbers

How Solana is built

Most blockchains process transactions one after another. Solana’s runtime, Sealevel, requires every transaction to declare in advance which accounts it will read and write, so transactions that touch different accounts can run in parallel across many CPU cores. A cryptographic clock called Proof of History timestamps events so validators agree on ordering without constant back-and-forth messaging. Leaders are scheduled ahead of time in four-slot windows and produce a new block every slot. Until September 2026 a slot lasted 400 milliseconds; validators activated 250 millisecond slots at epoch 1037 on September 18, 2026, which shortens each leader’s window to about one second, and a draft proposal, SIMD-0525, would take slots to 200 milliseconds with a smaller 50 million compute unit budget per slot if the network’s skip rate stays acceptable. In the first epoch at 250 milliseconds the network skipped about 0.05 percent of scheduled slots.

Fees and local congestion pricing

Fees are split into a base fee of 5,000 lamports (0.000005 SOL) per signature and an optional priority fee that users attach to jump the queue. Half of every base fee is burned and half goes to the validator that produced the block. Priority fees used to be split the same way; since SIMD-0096 was adopted, the block producer keeps 100 percent of them. Because congestion is priced per account rather than across the whole network, a popular token launch raises costs for the accounts involved rather than for every user. Over the 24 hours to October 6, 2026 the chain collected about $1.04 million in fees, of which roughly $115,000 was burned, according to DefiLlama. On September 22, 2026 Solana’s 24-hour fees of about $1.1 million were roughly 70 percent higher than Ethereum’s, although Ethereum still burned more because it destroys its entire base fee.

The hardware trade-off

The cost of a single fast state is that validators need high-end servers and fast connections. Bare-metal hosting suitable for a validator runs about $350 to $470 a month according to a Helius primer, which is modest, but the voting cost is not: each validator pays roughly 1 SOL a day in vote transaction fees just to participate in consensus. That is the single biggest barrier to small operators and is the main reason Alpenglow, described below, matters for decentralization as well as speed. Geography matters too. Analysis published with the 250 millisecond change found a median first-slot penalty of about 28 milliseconds when consecutive leaders sit within 500 kilometres of each other, rising to 122 milliseconds beyond 8,000 kilometres, which is why faster slots push validators toward a handful of well-connected data centres. For a comparison of how Solana’s approach differs from other chains, see our guide to how blockchain consensus works.

Solana versus the alternatives

Sources: CCS reporting on September 7 and 20, 2026; Helius; Anza Alpenglow paper.

SOL: fees, staking and supply

SOL is used to pay fees and to stake. Validators, and the delegators who stake with them, earn newly issued SOL plus a share of fees. The issuance schedule began at 8 percent a year in early 2021 and declines by 15 percent each “inflation year” until it reaches a 1.5 percent floor. Because an inflation year is measured in epochs rather than calendar days, the rate lags the calendar slightly, but by the second half of 2026 it sits a little under 4 percent. Those new coins are paid to stakers in proportion to stake, so the staking yield before commission is roughly the inflation rate divided by the share of supply staked. With about two thirds of supply staked that works out to a gross yield in the mid-5 percent range, which matches the 5.46 percent net staking reward rate that Bitwise’s BSOL reported on October 4, 2026 after its 0.20 percent sponsor fee. Liquid staking tokens such as JitoSOL add a share of maximal extractable value (MEV) tips on top. Our guide to crypto staking covers the mechanics, and how to read a token supply schedule explains why the issuance curve matters more than the headline rate.

The issuance debate: SIMD-0228, SIMD-0550 and SGP-0002

Whether 4 percent issuance is too much is Solana’s longest-running governance argument. In February 2025 a proposal called SIMD-0228 suggested replacing the fixed schedule with a market-based formula, r(1 minus the square root of s), where s is the staked share of supply, so that issuance would fall when a large share of SOL was already staked and rise if participation dropped toward 33 percent. At the roughly 65 percent staking rate of the time, the proposal’s own arithmetic implied inflation of about 1.4 percent. Validators voted across epochs 753 to 755 in March 2025. Roughly 61 percent of yes-or-no votes were in favour (the later SIMD-0550 discussion cites 38.61 percent opposition), short of the two-thirds supermajority required, so the fixed schedule stayed in place.

The idea came back in a simpler form in 2026. SIMD-0550, titled Double Disinflation Rate, keeps the fixed curve but raises the annual disinflation step from 15 percent to 30 percent, which would reach the 1.5 percent floor years earlier. A companion governance proposal, SGP-0002, has been accepted and is contingent on SIMD-0550 being activated. The maintainer has said the proposal will go to a validator vote, with the main points of contention being the exact exponent arithmetic and whether it should be bundled with changes to validator commissions.

Who gets the fees

Three fee streams reach validators. The validator’s share of base fees (50 percent) and all priority fees go to whoever produced the block, so they accrue to the validator, not to its delegators, unless the operator chooses to share them. Jito MEV tips, paid by searchers who want their bundles included, are distributed through Jito’s tip programs to the validator and its stakers after an MEV commission. Inflation rewards go to stakers after the validator’s commission. The practical result is that a validator’s income depends on two very different things: its stake share, which determines how often it leads and how much commission it earns, and network activity, which determines how much fee and tip revenue each leader slot is worth. The worked example below puts numbers on both.

Reliability and the client question

Solana’s history includes a string of network halts. Helius counts eight full outages: a Turbine propagation bug in December 2020 (about 6 hours), the Grape Protocol launch flood in September 2021 (17 hours, with bots pushing 300,000 transactions a second), the Candy Machine NFT mint spam in May 2022 (8 hours), a durable nonce bug in June 2022 (4.5 hours), a duplicate block bug in September 2022 (8.5 hours), a large-block Turbine failure in February 2023 (about 19 hours) and an infinite recompile loop in the JIT cache in February 2024 (about 5 hours). Each required validators to coordinate a restart. A contributing factor was that almost every validator ran the same software, so a single bug could stop the chain. In August 2024 a coordinated patch fixed an ELF alignment vulnerability without any downtime. Since February 2024 the network has not halted, although on August 12, 2026 an outage at the hosting provider TeraSwitch left 28.83 percent of stake delinquent for 33 minutes, a reminder that the chain can lose a large share of validators without stopping.

Firedancer

Firedancer is an independent validator client written from scratch in C by Jump Crypto so that a bug in the Agave codebase cannot halt the network. Its hybrid form, Frankendancer, which pairs Firedancer’s networking and block production with Agave’s runtime, has run on mainnet since 2025, and the fully independent version is in early production with a gradually growing share of stake. The Solana Foundation reported that about 92 percent of stake ran Agave or Jito-Solana clients in April 2025 and about 7 percent ran Firedancer or Frankendancer. A stake-weighted survey on September 20, 2026 found 87.4 percent of stake on 4.x releases of the Agave family, with the remaining 12.6 percent on 0.x and 26.x builds, the version lines that include Frankendancer and Firedancer. That is better than Ethereum’s execution layer, where Geth runs about 50 percent of clients, but still short of the point where a single Agave bug would be survivable.

Alpenglow

Alpenglow is the largest consensus change in Solana’s history. Announced by Anza on May 19, 2025 and specified in SIMD-0326, it replaces TowerBFT voting and the consensus role of Proof of History with a protocol called Votor, which finalizes a block in a single round of voting when 80 percent of stake responds and in two rounds when only 60 percent does. The design target is median finality of about 150 milliseconds, with about 100 milliseconds possible in the fast path, compared with roughly 12.8 seconds today. Votes move directly between validators and are aggregated into BLS certificates rather than being posted as on-chain transactions, which removes the roughly 1 SOL a day validators currently pay in vote fees. The protocol tolerates up to 20 percent adversarial stake plus a further 20 percent that is offline, the so-called 20+20 model. A second component, Rotor, will replace the Turbine block propagation layer on a separate timeline. Anza published version 1.2 of the white paper on July 29, 2026 and a companion post on July 30, 2026 on how the design holds up against quantum computing.

Alpenglow went live on testnet on September 22, 2026, and Agave 4.3.0 release candidates carrying Alpenglow fixes were recommended for 10 percent mainnet adoption from September 4, 2026. Core developers have said publicly they will not rush it to mainnet; the next mainnet feature activation window opens on November 9, 2026, which is a window, not a confirmed Alpenglow date. The upgrade leaves the virtual machine, transactions, programs and fees untouched; only the consensus layer changes.

Decentralization in numbers

Solana Compass put the Nakamoto coefficient, the number of validators whose combined stake could halt the chain, at 18 on September 6, 2026, up from 12 two years earlier, with 671 staked validators across 46 countries and 439 data centres. Bitcoin and Ethereum each need only three entities to reach the equivalent threshold. The weak points are hosting and software rather than ownership: TeraSwitch alone hosted 22.1 percent of active stake on September 7, 2026, TeraSwitch and Latitude together hosted 45.7 percent, and the largest single validator controlled about 4 percent. Validators also face an exit queue, so stake cannot move instantly if a provider fails.

Where institutions show up

Spot ETFs with staking

US spot Solana ETFs began trading on October 23, 2025 under the SEC’s generic listing standards, led by Bitwise’s BSOL. Unlike the first Ether ETFs, they were able to stake from launch. The CCS ETF tracker follows nine US vehicles from issuers including Bitwise, Grayscale, Fidelity, VanEck, 21Shares and Morgan Stanley. By the end of August 2026 the category held roughly $1.5 billion; BSOL alone reported net assets of $1.34 billion on October 4, 2026, a 0.20 percent sponsor fee, 99 percent of assets staked and a 5.46 percent net staking reward rate. Solana ETFs attracted inflows even on days when bitcoin funds saw outflows in late August 2026. The structure is the same as the spot bitcoin ETFs, with the addition of a staking provider and a reward pass-through.

Stablecoins and payments

Solana held $16.75 billion of stablecoins on October 6, 2026 according to DefiLlama, up 2.4 percent over the prior week. Circle’s USDC is the largest at about $7.3 billion, followed by Tether at about $2.9 billion, with Paxos, BitGo and Anchorage Digital Bank issuing a further $4 billion or so between them, per rwa.xyz on October 5, 2026. The 2026 entrants are banks and payment processors rather than crypto firms: on October 2, 2026 the Bank of North Dakota launched Roughrider Coin on Solana through Fiserv’s banking platform, with VersaBank issuing and Fireblocks tokenizing, for settlement among more than 90 state banks and credit unions. Open Standard launched OUSD on Solana the same week. For how these instruments work, see what is a stablecoin.

Tokenized assets and securities settlement

Beyond stablecoins, rwa.xyz counted $4.34 billion of tokenized assets on Solana on October 5, 2026, from managers including BlackRock, Ondo and Securitize, up from about $4.6 billion in early September by the Solana Foundation’s own count and over $4 billion a month earlier. Tokenized equities are the fastest-growing slice. Across all chains, tokenized stocks held in DeFi protocols rose from $14.66 million in January 2026 to $262.4 million on September 30, 2026, and Solana led with $95.9 million, a 36.6 percent share; Kamino Lend held $54.1 million of that as loan collateral. Forward Industries’ tokenized common stock, FWDI, has traded on the Orca DEX since November 2025 with Superstate as transfer agent, using Solana’s token extensions to enforce an allowlist. In September 2026 the SEC issued a five-year exemptive order permitting tokenized shares that carry the same dividends, votes and liquidation rights as the underlying to trade on permissionless chains, with the Solana Policy Institute, Superstate and Orca among the parties consulted; Solana says it hosted more than half of all tokenized equity trading volume in 2026. On October 6, 2026 the Solana Foundation released Solana DvP, an audited open-source delivery-versus-payment program for atomic settlement of both legs of a trade, built with advisory input from JPMorgan, which had settled a Galaxy Digital commercial paper deal on Solana in December 2025. Our guides to tokenization and tokenized stocks go deeper.

Treasury companies

Public companies have also built SOL treasuries, holding and staking SOL on their balance sheets in the way Strategy holds bitcoin (see bitcoin treasury companies explained). Forward Industries, chaired by Multicoin’s Kyle Samani, reported 8,501,298 SOL worth about $1.0 billion on September 30, 2026, up 948,601 SOL in the period, and offers a liquid staking token, fwdSOL. DeFi Development Corp held about 2.31 million SOL at its August 2026 quarterly report but posted a $27 million net loss on digital assets and closed its Treasury Accelerator program, a reminder that the model is leveraged to the SOL price. Upexi is another holder. These companies are a source of persistent demand for staked SOL, and also a source of forced selling if their shares trade below net asset value for long.

Trading activity

Solana’s spot DEXs processed about $2.06 billion in the 24 hours to October 6, 2026 and $77.9 billion over the trailing 30 days according to DefiLlama, led by PumpSwap, Orca, pump.fun, Meteora and Axiom, with a further $3.56 billion a day in perpetual futures. In the week ending September 13, 2026 Solana DEXs executed about 208 million spot trades, slightly more than the roughly 190 million on the New York Stock Exchange. A large share of that count is memecoin trading through launchpads such as pump.fun, which is cyclical and thin per trade. The chain also hosts prediction markets, perpetuals venues and the largest DePIN networks, including Helium and Render, which moved to Solana in 2023. CCS interviewed Fogo’s Doug Colkitt in July 2025 on high-frequency trading on Solana.

Validator economics in detail

A Solana validator earns in three ways and spends in two. Income is inflation commission (a percentage of the staking rewards its delegators earn), block rewards (its half of base fees plus all priority fees for the blocks it produces) and MEV commission (a percentage of Jito tips routed to its stakers). Costs are vote fees, about 1 SOL a day or 300 to 365 SOL a year, and infrastructure, about $350 to $470 a month for bare metal plus engineering time. Because vote fees are fixed in SOL while income scales with stake, there is a minimum stake below which a validator loses money; the Solana Foundation’s delegation program covers vote costs on a declining schedule (100 percent for the first three months, then 75, 50 and 25 percent) and matches stake one for one up to 100,000 SOL to help new operators cross that line. Jito’s own figures from a tracked week showed about 2 million bundles generating 8,500 SOL in tips to validators, which gives a sense of the MEV layer’s scale relative to issuance.

Two governance changes reshaped the income side. SIMD-0096 gave block producers 100 percent of priority fees instead of 50 percent, which increased validator revenue during busy periods but, as the proposal itself acknowledged, let validators inflate reported priority fees cheaply. SIMD-0228 would have cut issuance, and therefore commission income, by more than half, and failed. Alpenglow will change the cost side by removing vote fees entirely, which Helius estimates would sharply lower the stake needed to break even and is the strongest decentralization argument for the upgrade. A September 2026 attempt to enforce strict fee-priority ordering within blocks, SIMD-0649, closed on September 25, 2026 without the client-developer approvals it needed, so leaders retain discretion over ordering and can keep priority fees paid to themselves.

How we got here: a timeline

Worked example: a mid-size validator’s year

How to evaluate Solana exposure: a checklist

• What is the current issuance rate and how much of your yield is inflation? A staking yield that is mostly new issuance is a transfer from non-stakers, not income. Check the schedule and the live rate; in late 2026 issuance is a little under 4 percent and most of a 5.5 percent gross yield comes from it.
• Who runs the validators you are staked with, and on which client? Favour operators that publish their client, hosting provider and commission, and note whether they run Firedancer or Frankendancer; client diversity is the network’s main unfinished safety work.
• How concentrated is hosting? Two providers hosted 45.7 percent of stake in September 2026. Ask where a validator’s servers sit and what happens to your stake during a provider outage.
• Does the ETF or product actually stake, and what does it keep? BSOL staked 99 percent of assets and passed through 5.46 percent net on October 4, 2026 after a 0.20 percent fee. Compare the staked share, the fee and the staking provider across products rather than the headline ticker.
• Is the on-chain activity you are pricing durable? Separate stablecoin and tokenized-asset balances, which are sticky, from memecoin trade counts, which are not. Trade counts above the NYSE’s say little about fee revenue per trade.
• How much tokenized value is actually used? Of $4.34 billion in tokenized assets, only a fraction is lent or posted as collateral; the $95.9 million of tokenized stocks in Solana DeFi on September 30, 2026 is the usage figure, not the issuance figure.
• What is the upgrade risk calendar? Alpenglow is the biggest consensus change the chain has made. Know whether the November 9, 2026 window is being used and what the rollback plan is before sizing positions around it.
• For treasury companies, what is the premium to net asset value and the debt structure? Forward Industries’ NAV was about $870 million against $1.0 billion of SOL on September 30, 2026; DeFi Development Corp carries 5.5 percent convertible notes due 2030. A discount to NAV plus debt is the failure mode.

Risks and open questions

Four issues deserve scrutiny. First, reliability: Solana has been stable since February 2024, but a consensus change as large as Alpenglow carries execution risk, which is why developers are moving slowly, and the August 2026 TeraSwitch incident shows that infrastructure concentration can take a quarter of the network offline without any bug at all. Faster slots make the network more sensitive to geography, so the push to 200 milliseconds trades inclusiveness for speed.

Second, decentralization: hardware and vote-fee requirements and stake concentration among large operators and liquid staking providers remain higher than on Ethereum, roughly 87 percent of stake still ran one client family in September 2026, and the Nakamoto coefficient of 18, while improving, is a small number for a chain settling bank stablecoins. Third, activity quality: a large share of Solana’s volume has come from memecoin speculation, which is cyclical, and MEV practices such as sandwich trading have drawn criticism; the failure of SIMD-0649 in September 2026 shows how hard it is to constrain leader discretion without consensus among client teams.

Fourth, the issuance question is unresolved. Stakers and treasury companies benefit from 4 percent issuance; non-staking holders and ETF investors in products that do not stake pay for it. SIMD-0550 and SGP-0002 would cut it faster, but the same validators who earn commission on it must vote it through. Much of the tokenized-asset total is held rather than lent or used as collateral, so headline figures overstate economic use. SOL is also a volatile asset whose ETF and treasury demand can reverse, and treasury companies that trade below net asset value can become forced sellers.

What to watch next

Glossary

Slot

The fixed time window in which one scheduled leader produces a block; 250 milliseconds on Solana since September 18, 2026.

Epoch

A period of 432,000 slots during which the leader schedule and stake weights are fixed; staking rewards are paid at epoch boundaries.

Proof of History

A verifiable clock built from a sequential hash chain that lets validators agree on the order of events; Alpenglow removes its consensus role.

TowerBFT

Solana’s current voting protocol, in which validators post vote transactions on-chain and finality takes about 12.8 seconds.

Alpenglow

The consensus overhaul specified in SIMD-0326 that replaces TowerBFT with Votor and Turbine with Rotor, targeting 100 to 150 millisecond finality.

Votor

Alpenglow’s voting component, which finalizes a block in one round with 80 percent of stake responding or two rounds with 60 percent, using off-chain votes aggregated into BLS certificates.

Firedancer and Frankendancer

Jump Crypto’s independent validator client written in C, and the hybrid that pairs its networking with the Agave runtime.

Agave

The reference validator client maintained by Anza; Jito-Solana is a fork of it that adds MEV auctions.

Sealevel

Solana’s parallel runtime, which executes transactions that touch different accounts at the same time.

Priority fee

An optional fee attached to a transaction to improve its chance of inclusion; since SIMD-0096 the block producer keeps all of it.

Jito tips

Payments from searchers to have transaction bundles included in a block, distributed to the validator and its stakers after an MEV commission.

SIMD

Solana Improvement Document, the proposal process for protocol changes; some go to a stake-weighted validator vote.

Nakamoto coefficient

The smallest number of validators whose combined stake could halt or censor the chain; 18 for Solana in September 2026.

Token extensions

Features of the Token-2022 standard, such as transfer allowlists and permanent delegates, that let regulated issuers enforce rules on-chain.

Why it matters

Solana is the clearest test of whether a single fast base layer can win the markets that care most about speed and cost: trading, payments and tokenized securities. Its adoption by ETF issuers, stablecoin issuers, asset managers and, in October 2026, state banks and a settlement tool advised by JPMorgan, shows that institutions are willing to use it. Whether it keeps that position depends less on raw throughput than on the unglamorous work now under way: a second independent client, faster and simpler consensus, a credible path to lower issuance, and a record of staying online.

For readers, the useful discipline is to separate three things that get blurred in Solana coverage: the engineering (slot times, finality, client diversity), the economics (issuance, fees, commission, MEV) and the adoption (stablecoins, tokenized assets, ETFs). Each has its own data sources and its own calendar, and a strong number in one does not substitute for a weak number in another.

Sources

1. Anza: Alpenglow, A New Consensus for Solana, May 19, 2025 (white paper v1.2 July 29, 2026)
2. Anza: Agave releases, v4.3.0-rc.0 mainnet upgrade candidate, September 4, 2026
3. Solana Foundation: SIMD-0096 Reward full priority fee to validator, December 18, 2023
4. Solana Foundation: SIMD-0228 Market-based emission mechanism, February 2025
5. Solana Foundation: SIMD-0550 Double Disinflation Rate, 2026
6. Helius: Solana Outages, Complete History, 2025
7. Helius: Solana Validator Economics, A Primer, 2024
8. DefiLlama: Solana chain overview (TVL, stablecoins, fees, DEX volume), October 6, 2026
9. rwa.xyz: Solana network tokenized assets, October 5, 2026
10. Bitwise: BSOL Solana Staking ETF fund details, October 4, 2026
11. Forward Industries: SOL treasury dashboard, September 30, 2026
12. Crypto Coin Show: Solana deploys Alpenglow consensus upgrade to testnet, September 23, 2026
13. Crypto Coin Show: Solana cuts block time to 250 milliseconds, September 20, 2026
14. Crypto Coin Show: Solana beats Bitcoin on one key metric, but a single software bug could still take down the network, September 7, 2026
15. Crypto Coin Show: SEC approves tokenized stock trading on permissionless blockchains for five years, September 23, 2026
16. Crypto Coin Show: Tokenized stocks in DeFi vaults reach $262.4 million, September 30, 2026
17. Crypto Coin Show: Solana Foundation releases open-source settlement tool advised by JPMorgan, October 6, 2026
18. Crypto Coin Show: Bank of North Dakota launches Roughrider Coin stablecoin on Solana, October 2, 2026
19. Crypto Coin Show: Solana’s transaction ordering proposal (SIMD-0649) stalls, September 27, 2026
20. Crypto Coin Show: DeFi Development Corp shutters SOL accelerator after $27 million quarterly reversal, August 13, 2026

Disclosure: This guide is for education only and is not investment, legal or tax advice.