Solana Ecosystem Guide: Speed, DeFi, NFTs & Memecoins

Introduction

Until 21 October 2026, about 65% of the transactions Solana reported each second were validator votes, not user activity: the network averaged 4,783 transactions per second in the 60 minutes to 16 October 2026, and only 1,675 were non-vote transactions. Alpenglow, the consensus rewrite that went live on 21 October 2026 (Solana Foundation), took votes out of blocks, so the headline figure now counts non-vote transactions alone and averaged 1,284 per second in the hour to 10:51 UTC on 5 November 2026. This guide covers the Solana ecosystem as it stands: the architecture and which upgrades are live, measured speed, decentralized finance (DeFi) with $5.57B in total value locked (TVL), non-fungible tokens (NFTs), memecoin launchpads and the risks to check before using any of them. Every figure carries a date, and every vendor claim is labelled as one.

How Does Solana Work, and Which Upgrades Are Actually Live?

Solana mainnet switched from TowerBFT to Alpenglow consensus on 21 October 2026, replacing the design the network launched with. A mainnet query on 22 October 2026 returned the Alpenglow genesis certificate for slot 459,652,999, which was certified with votes from 446 of the network's 654 validators (Solana Foundation, 2026).

Proof of History, TowerBFT and the switch to Votor

Proof of History (PoH) was the cryptographic clock that ordered events before validators voted on them, and TowerBFT was the voting scheme built on top. Under TowerBFT, validators cast votes as ordinary transactions, and finality took about 12.8 seconds (Solana Foundation, 2026). Under Votor, Alpenglow's voting protocol, validators send votes directly to each other and bundle them into certificates, and PoH no longer paces block production. Block time, the interval between slots, was cut from 400 to 200 milliseconds in four stages completed in October 2026 (Solana Foundation, 2026). The Solana Virtual Machine (SVM) executes the programs, and Alpenglow left the SVM, transaction formats, fees and account model untouched. Solana's decentralized finance (DeFi), non-fungible token (NFT) and memecoin applications therefore kept running unchanged when consensus changed.

Agave, Firedancer and Alpenglow

Alpenglow activated at the start of epoch 1064, at about 13:36 UTC on 21 October 2026, according to the Solana Foundation and Anza. TowerBFT ran for a further 5,000 slots, about 18 minutes, and Alpenglow consensus took over from slot 459,653,000. It replaced vote transactions with direct messages between validators and targets roughly 150 milliseconds to finality, a figure from the Solana Foundation and Anza rather than an independent measurement. On 5 November 2026 a public RPC node reported the finalized slot one slot behind the newest processed slot in 26 of 30 reads and two behind in the other four, roughly 0.2 to 0.4 seconds at the measured slot rate, which is consistent with sub-second finality but is not a direct test of the 150-millisecond target (Solana RPC, 2026). Votor, the voting half, shipped in Agave 4.3. Rotor, a block-propagation redesign, is planned for a later release and not yet scheduled (Solana Foundation, 2026). For the migration itself, validators had to run version 4.3.2 of Agave, Jito-Solana or Harmonic Salsa. Firedancer, the independent client from Jump Crypto, went live on mainnet on 12 December 2025, yet neither it nor the hybrid Frankendancer was supported for the migration (Solana Foundation, 2026). Firedancer releases from v26.10.2 onward run Alpenglow consensus, and validators running them held about 18% of active stake on 5 November 2026 (Firedancer, 2026; Solana RPC, 2026).

Data current as of November 2026.

Consensus status matters less to users than measured speed, and the measured figures differ sharply from those in promotional material.

How Fast Is Solana When Throughput Is Measured Rather Than Claimed?

Solana processed about 1,284 transactions per second on average in the hour to 10:51 UTC on 5 November 2026, all of them non-vote transactions, because Alpenglow took validator votes out of blocks. Before the switch, roughly 65% of the 4,783 per second measured on 16 October 2026 were votes, and non-vote transactions averaged 1,675 per second.

Total versus non-vote throughput

Solana's own performance samples separate total from non-vote transactions. Sixty one-minute samples from the mainnet method getRecentPerformanceSamples that ended at about 11:55 UTC on 16 October 2026, before the switch, averaged 4,783 total transactions per second, with a range of 4,267 to 5,539, while non-vote throughput averaged 1,675, with a range of 1,195 to 2,409 (Solana RPC, 2026). A sample on 23 September 2026 read 4,000 to 4,300 total and 1,400 to 1,800 non-vote (Solscanner, 2026), so the two pre-switch dates agree within the sampling range. Since Alpenglow removed votes from blocks on 21 October 2026, total-throughput dashboards have fallen sharply with no change in real activity (Solana Foundation, 2026). The same query for the hour to 10:51 UTC on 5 November 2026 averaged 1,284 per second, with a range of 1,051 to 1,577, and the total matched the non-vote count in 57 of the 60 samples, differing by one transaction in the other three. Non-vote counts still include failed transactions, and throughput varies with the time of day, so one hour is a snapshot rather than a trend.

Data current as of November 2026.

Outages and uptime since 2024

Solana's last full halt began at 09:53 UTC on 6 February 2024 and lasted about five hours, after a bug in the program cache stalled validators running version 1.17 (Solana Foundation, 2024). The Solana Foundation reports 100% uptime since that date (Solana Foundation, 2026). A routing failure at the network's largest infrastructure provider on 12 August 2026 took nearly 29% of stake offline, yet blocks kept being produced, and the provider recovered in just over 30 minutes (Solana Foundation, 2026). The foundation attributes the earlier outages to the network's first years. The sources for this record are the foundation's own reports, which makes them claims to weigh rather than independent audits.

Speed only matters if applications use it, and DeFi uses it most visibly.

What Does Solana DeFi Look Like in Numbers Today?

Solana DeFi holds $5.57B in total value locked (TVL), the third-largest on any chain behind Ethereum at $45.82B and Base at $6.21B ↗ (DefiLlama, 5 November 2026). That is 58% below the $13.23B peak of 14 September 2025.

Swaps, perpetuals and aggregation

Solana decentralized exchanges processed $2.22B in the 24 hours to 5 November 2026 and $56.24B over 30 days (DefiLlama, 2026). Volume is spread across venues: PumpSwap led the day with $315M, ahead of Orca at $276M, BisonFi at $257M and Raydium at $203M, while Meteora's DLMM pools handled $95M. Raydium's automated market maker holds $1.08B in TVL. Jupiter aggregates swap routes and also runs Jupiter Perpetual Exchange, which holds $698M. Volume measures turnover and TVL measures deposits, so the two figures answer different questions. Aggregators route one swap across several of these venues, which is why a single trade can touch Orca, Raydium and Meteora pools at once.

Lending and liquid staking

Liquid staking and lending lead the Solana ranking by TVL. Sanctum's validator liquid staking tokens hold $1.70B, Kamino Lend $1.25B, Jito Liquid Staking $994M, Binance Staked SOL $1.05B and Jupiter Lend $1.03B (DefiLlama, 2026). Marinade's native staking pool holds a further $348M, and Jito Restaking holds $16M. A liquid staking token (LST) represents staked SOL that keeps earning rewards while it is used elsewhere. Inflation of 3.57% a year goes entirely to validators and their delegators, and 68.0% of supply is staked (Solana RPC, 2026). That ratio implies a gross yield near 5.3% before validator commission.

Stablecoins on Solana

Stablecoins circulating on Solana totalled $14.80B on 5 November 2026, down from $16.81B 30 days earlier (DefiLlama, 2026). USDC accounts for $7.11B and USDT for $2.56B, with USD1 at $1.31B and PayPal USD at $0.69B. Stablecoin supply counts dollars issued on the chain, a separate measure from TVL, which counts assets deposited in protocols. Solana's low fee, 5,000 lamports or about $0.0005 per signature (Solana documentation, 2026), suits payment-sized transfers. DefiLlama's stablecoin dataset also lists BlackRock's BUIDL tokenized fund at $0.58B on Solana.

Data current as of November 2026.

NFTs draw far less capital than DeFi, and their market has shrunk faster.

What Are Solana NFTs and Where Are They Traded?

Solana NFTs trade mainly on Magic Eden and Tensor, and compressed NFTs made minting cheap enough for tens of millions of mints. Trading volume today is a small fraction of the 2024 level.

Marketplaces and compressed NFTs

Magic Eden accounted for about 87% of volume in a seven-day snapshot of roughly 10,400 SOL published in September 2026, with Tensor at about 10%, OpenSea at 2% and Orbis at 1% (Digital Coin Journal, 2026). The snapshot comes from Chadbot, which tracks only those four venues, so the shares do not measure the whole market. Tensor targets active traders with on-chain pool liquidity for buying and selling NFTs (xgram, 2026). A compressed NFT stores ownership data in a compact on-chain structure instead of a full account, which lowers the minting cost. Syndica counted more than 30 million compressed NFTs minted in the first half of 2024 and found they made up over half of all NFT transfers since February 2024 (Syndica, 2024).

Market size today

Weekly trading on Magic Eden's M2 marketplace program fell from $31.5M in the week of 15 April 2024 to $0.57M in the week of 18 May 2026, a 98% decline, after $9.56M in the week of 8 December 2025 (Dune, ilemi dashboard, 2026). NFTs carry no TVL, so volume is the available measure, and it covers one venue rather than all of them. The collapse in trading did not remove the standard: compressed NFTs keep serving ticketing, loyalty and game-item uses that depend on cheap minting rather than resale value. Collectors face a thin market in which exits are hard.

How Do Solana Memecoins and Launchpads Work, and How Often Do They Fail?

Pump.fun created 78% of the 11.5 million new tokens that traded on Solana between 1 November 2025 and 7 October 2026 (Bitquery, 2026). Of its coins, 79% never traded after launch day, and 1 in 58 graduated to a trading pool.

Bonding curves and graduation

A launchpad such as Pump.fun lets anyone create a token in a few clicks and trade it at once against a bonding curve, a formula that raises the price as buyers add SOL. When enough SOL has come in, the token graduates, and the SOL and remaining tokens move into a pool on PumpSwap, Pump.fun's own exchange (Bitquery, 2026). Pump.fun charges roughly 1% on bonding-curve trades, half of its net fees fund PUMP token buybacks, and creators earn up to 0.05% of PumpSwap volume (CryptoBriefing, 2026). Rivals include Meteora's dynamic bonding curve, which made 10% of new tokens, and Raydium's LaunchLab, home of Bonk.fun, at 4% (Bitquery, 2026).

What the failure data shows

Half of Pump.fun's launches had their last trade within about three minutes, and only 1 in 437 still had ten trades in its fifth week (Bitquery, 2026). A separate risk sits in PumpSwap pools opened without a curve: 451,402 such pools appeared up to 5 July 2026, and 99% were later emptied by the developer who opened them, a median of 12 minutes in. Pools created by graduation almost never lose their money that way. PumpSwap handled $315M of DEX volume and pump.fun $126M in the 24 hours to 5 November 2026 (DefiLlama, 2026), so activity stays large even though most coins fail.

Surviving these markets depends less on picking winners than on avoiding the structures that guarantee losses.

How Can a Beginner Start Using the Solana Ecosystem Safely?

SOL carries a market capitalisation of $57.08B (CoinPaprika API, 5 November 2026). Its price sits 67.1% below the all-time high set on 19 January 2025. Safe use begins with a self-custody wallet, small amounts and checks on what each token and pool allow.

Wallet, SOL and staking basics

A Solana wallet such as Phantom or Solflare holds SOL and tokens, and SOL pays every fee. The base fee is 5,000 lamports, or 0.000005 SOL, per signature (Solana documentation, 2026). Delegating SOL to a validator leaves the coins in the holder's own stake account. Stake is concentrated: 693 vote accounts held stake on 5 November 2026, and the 17 largest held 33.3% of it (Solana RPC, 2026). One operator can run several vote accounts, so 17 is an upper bound on the true Nakamoto coefficient ↗. The Solana Foundation's introductory page still says "over 1,000 validators". Unstaked SOL is diluted by the full 3.57% issuance each year, a cost that staking offsets.

Risks to check before buying anything

Four checks cover the main exposures. A token creator can configure a freeze authority that freezes balances (Solana Foundation, 2026). A new PumpSwap pool created without a bonding curve matches the pattern behind most emptied pools in Bitquery's data. Drainer sites and contract bugs cannot be reversed, and a small separate wallet limits what a malicious site can reach (LatestCrypto, 2025). Since 21 October 2026, each validator pays a 0.8 SOL admission fee from its vote account every epoch, and one that leaves the account short does not vote or produce blocks in the next epoch, so delegators check their operator's status (Solana Foundation, 2026). Price is the other exposure: SOL stood 67.1% below its high on 5 November 2026, and this guide offers no forecast.

FAQ

?How much does a Solana transaction cost in fees?

The base fee is 5,000 lamports, or 0.000005 SOL, per signature, about $0.0005 at the SOL price of $96.71 on 5 November 2026 (CoinPaprika API). Priority fees can add to that during congestion, and applications charge their own fees on top, such as roughly 1% on Pump.fun bonding-curve trades. A swap through a decentralized exchange therefore costs far more than the network fee alone.

?How does Solana differ from Ethereum and its layer 2 networks?

Solana runs applications on a single layer 1 ledger instead of splitting activity across rollups ↗. Ethereum scales through layer 2 networks that post data back to it, and L2BEAT, renamed ETHBEAT on 4 November 2026, tracked 83 active layer 2s with about $38.8B in value secured on 5 November 2026. That metric differs from the $5.57B of TVL on Solana, so the two cannot be compared directly. The trade-off falls on node operators: Anza's documentation calls for 256 GB of RAM or more.

?What happens to my staked SOL if my validator drops out of consensus?

Delegated SOL stays in the holder's own stake account, so a validator going offline does not remove the coins. It does stop that validator earning rewards for the epochs it misses, and delegators earn nothing from it until it returns or the stake moves. Since Alpenglow, operators run an Agave-based client or a Firedancer release that supports the new consensus, and a validator without 0.8 SOL in its vote account for the admission fee charged each epoch is not admitted for the next epoch, according to the Solana Foundation.

?Is a liquid staking token like JitoSOL the same as SOL?

A liquid staking token is a separate token that represents SOL staked through a pool, so it is not SOL itself. Its value tracks the staked SOL plus accrued rewards, and it adds the risk of the pool's smart contracts. DefiLlama showed $994M in Jito Liquid Staking and $1.70B in Sanctum's validator tokens on 5 November 2026. Holders can use these tokens in lending or trading while the underlying SOL keeps earning.

?Can Solana go down again after the February 2024 outage?

Any network can halt again. The Solana Foundation lists the 6 February 2024 stall as the last full outage and reports 100% uptime since. On 12 August 2026 a routing failure took nearly 29% of stake offline, yet blocks kept being produced. Anza tested the Alpenglow migration on a community cluster, testnet and devnet before mainnet, where Alpenglow consensus has run since slot 459,653,000 on 21 October 2026. New consensus code has little mainnet history, so its first months deserve watching.

?Does a graduated pump.fun token carry less risk than a new one?

Graduation changes the structure, not the odds that a token holds value. Bitquery found that pools created by graduation almost never lose their money to a developer withdrawal, unlike curve-skip pools, of which 99% were emptied. Only 1 in 58 Pump.fun launches graduated, though, and a third of September 2026 graduations took three trades or fewer. A graduated token answers the rug-pull question, not the value question.

?What is the difference between a vote transaction and a user transaction?

Until 21 October 2026, validators proved they were alive and agreed on blocks by sending vote transactions, which sat in blocks like any other transaction. They carried no user value but counted in the total transactions-per-second figure: the 60-minute sample to 16 October 2026 put votes at roughly 3,108 of 4,783 transactions per second. Alpenglow moved votes off the ledger, so validators now exchange votes directly and every transaction counted in a block is a non-vote one. Total throughput counts dropped at activation without any loss of user activity.

?Can the creator of a Solana token freeze my balance?

Yes, if the creator configured that power. The Solana Foundation notes that some tokens have freeze capabilities when their creators set them. A freeze authority lets its holder lock an individual token account so the balance cannot move. Checking a token's mint details in a block explorer before buying shows whether the authority is still active, and tokens with the authority disabled remove that risk.

References / Sources