Solana traded at around $122 on Friday afternoon, reclaiming the $120 level it had slipped below on Thursday. The question that will decide the coming weeks, however, is not in the chart but
Solana traded at around $122 on Friday afternoon, reclaiming the $120 level it had slipped below on Thursday. The question that will decide the coming weeks, however, is not in the chart but in an announcement from Jump Crypto: the validator client Frankendancer loses its support as soon as the consensus upgrade Alpenglow is switched on on mainnet. Anyone who has staked Solana therefore needs to know which software is running on the validator their own coins are delegated to.
First the hard numbers on the price itself. According to CoinGecko data, SOL stood at $122.09 on October 2 at 2:41 p.m. German time, and the range over the preceding 24 hours ran from $116.75 to $123.55. Kraken showed $122.26 a few minutes later, Coinbase the same value and 108.71 euros. Market capitalisation came to roughly $71.9 billion, which is rank seven. Trading volume over 24 hours was worth about $4.6 billion.
Frankendancer is a hybrid. The client combines the networking layer of Firedancer, developed by Jump Crypto, with the consensus code of Agave, the client maintained by Anza. That mixture has been running on mainnet since September 2024, and it was always meant as an interim step until the full Firedancer was ready. That has been the case since December 2025.
Now Jump Crypto is drawing a line. According to a report by Cryptobriefing, support for Frankendancer ends with the activation of Alpenglow, which the project is targeting for October 2026. The team cites the maintenance and security burden of the hybrid as the reason. Those resources are to go into the full Firedancer instead.
The technical compulsion behind it is easy to explain. Alpenglow replaces the two mechanisms Solana has been built on since 2020: Proof of History and TowerBFT. Votor takes their place. Because Frankendancer borrowed its consensus code from the older Agave version, the client becomes inoperable the moment the switch is flipped. Two clients remain, the full Firedancer from Jump Crypto and Agave from Anza.
The order of magnitude is not small. Cryptobriefing puts the Firedancer family's share of total Solana stake at 14 to 26 percent, and the full Firedancer alone at roughly 11 to 14 percent in more recent snapshots. The difference between those two figures is roughly the share still running on Frankendancer today. Operators who do not move to the full Firedancer in time, or fall back to Agave, drop out of consensus when the switch happens.
For you as a delegator that means: a validator that goes offline at the switch produces no blocks and earns no rewards. Your coins are not lost, because they sit in your own stake account and not with the validator. But the earnings stand still until the operator catches up or you move your delegation.
Votor is the new voting mechanism. Validators exchange their votes directly among themselves, the signatures are aggregated, and agreement runs along two possible paths. If more than 80 percent of the stake agrees in the first round, a block is final after roughly 100 milliseconds. If it is only enough for more than 60 percent, it takes two rounds and roughly 150 milliseconds. Those are the figures the infrastructure provider Alchemy cites in its technical write-up. Today finality takes roughly 12.8 seconds.
Finality is not a marketing word here but the moment from which a transfer can no longer be reversed. As long as it is pending, the recipient carries a residual risk. Solana co-founder Anatoly Yakovenko once boiled this down to the formula that finality really only matters at the checkout. For payments in retail, the difference between 12.8 seconds and a tenth of a second is precisely the difference between unusable and usable.
Rotor is the second building block. It replaces Turbine, the previous mechanism for passing block data around the network. Alchemy states that 1,500 data fragments pass through in 18 milliseconds at one gigabit per second. On top of that comes a changed fault tolerance: Alpenglow is meant to cope with 20 percent malicious stake plus a further 20 percent of stake sitting offline, so 40 percent in total, whereas classic BFT mechanisms end at 33 percent.
The cut-off date for Frankendancer hangs on an event and not on a calendar date, and that is exactly what makes preparing uncomfortable for delegators.
On October 2 at 2:50 p.m. German time we queried the public mainnet node at api.mainnet-beta.solana.com. The answer to getVersion reads solana-core 4.3.0, feature set 3383571666. Agave 4.3 is precisely the version that brings the switch for Alpenglow with it. So the node is already running the software that is meant to carry the upgrade.
This distinction matters. The fact that a node runs version 4.3.0 does not mean Alpenglow is active. Consensus only changes once the associated feature gate is unlocked and enough stake is running the new version. The software is the truck, the feature gate is the key.
The same query returns one more figure that matters directly for your staking. The node reports epoch 1047 at slot 311,009 of 432,000. Three measurement windows of 60 seconds each produced 229, 222 and 230 slots, together 681 slots in 180 seconds. That works out at roughly 264 milliseconds per slot.
An epoch always spans 432,000 slots. At the classic 400 milliseconds it lasted 48 hours. At the measured 264 milliseconds it lasts roughly 31.7 hours. At the time of measurement, the running epoch 1047 was still about 8.9 hours from its end.
With Solana staking, a deactivation does not take effect immediately. You deregister your stake, and the deregistration becomes effective at the end of the running epoch. Only after that can you withdraw the coins. So anyone exiting at the start of an epoch currently waits just under a day and a half instead of the former two days. That is a noticeable improvement, and it is already here without Alpenglow being active: the shortening of the slot time happened in separate steps.
In practical terms for the upcoming client switch: if your validator fails at the changeover and you want to react, moving the delegation costs you roughly a day and a half without earnings. Anyone who wants to avoid that looks at the validator beforehand rather than afterwards. If you would rather not delegate yourself at all, the staking platform comparison lists providers that handle the operation on your behalf.
The query also gives the current inflation rate: 3.62 percent, and according to the node the entire share of it flows to the validators, with the foundation share at zero. The nominal staking reward is fed from that rate, minus the validator's commission. Anyone quoted a yield far above it should ask where it comes from.
The overhaul fundamentally changes the economics of running a validator, and that is the part which feeds through to delegator returns over the medium term. Today a validator pays for its votes as transactions on chain, roughly one SOL a day according to Alchemy's compilation. Over a year that adds up to a three-digit amount of SOL before the first euro of profit appears.
With Alpenglow those vote transactions disappear. In their place comes a so-called Validator Admission Ticket of roughly 0.8 SOL a day, which is burned in full and therefore lowers inflation. Alchemy puts the minimum stake needed for an economically viable operation at roughly 4,850 SOL before and roughly 450 SOL after.
At the same time room frees up. According to the same source, vote transactions occupy roughly 75 percent of block space today, and the vote data per slot shrinks from roughly 500 kilobytes to roughly 1,000 bytes. Our own measurement supports the order of magnitude: in the 60-second window the node counted 283,904 transactions in total and 130,560 excluding vote transactions. So roughly 54 percent of all transactions were votes.
A lower barrier to entry admits more small operators. That spreads the stake more widely and reduces dependence on a few large addresses. But it can also sharpen the competition for delegations and push commissions down. Both are plausible, and so far neither is proven, because the overhaul is not live yet.
Rotor will distribute block data differently across the network in future, and this part of the overhaul sits behind the same switch as the consensus change.
The previous day belongs to any reading of the price picture. As we reported on October 1, the US spot ETFs on Solana recorded a net outflow of $11.1 million on September 30, and the price subsequently slipped below $120. That ended a run of inflows which had carried the price through September.
On October 2 the picture turned. Bitstamp showed a gain of 4.43 percent over 24 hours for the SOL against euro pair, most recently 108.66 euros, within a range of 103.39 to 109.84 euros. Kraken reported a 24-hour range of $116.70 to $123.74 for SOL against the dollar at a weighted average price of $120.38. The recapture of the $120 level is therefore still fresh, and it sits close to the day's average price.
For the coming days the two ends of the 24-hour range are the obvious reference points, because that is where trading actually took place most recently. On the downside the low sits at $116.70. If the price falls below it, the recapture of the $120 level is void and the previous day's picture with the ETF outflow resumes. On the upside the daily high stands at $123.74. Above that, the recent range offers the price no resistance.
The overhaul itself works poorly as a price driver in either direction. It helps if payment and trading applications actually make use of the shorter finality. It hurts if a noticeable share of the stake drops out at the changeover because operators sleep through the client switch. Which case materialises cannot be measured in advance, and nobody should tell you otherwise. What you can measure is the client of your own validator.
Staking rewards are not capital gains in Germany. The tax authorities treat them as other income under Section 22 number 3 of the Income Tax Act and charge them at your personal rate of tax. In return an exemption limit of 256 euros a year applies. Anyone who stays below it pays nothing. Anyone who exceeds the limit pays tax on the full amount and not merely on the excess, because an exemption limit works differently from an allowance.
What counts is the value of the coins received at the moment they are received. With Solana, rewards arrive epoch by epoch, so currently about every 32 hours. Over a year that adds up to several hundred inflows, and every single one needs its price. Doing that by hand is barely feasible.
Whether this legal position stays as it is remains open. A draft bill on crypto taxation is due to go before the cabinet on October 14, and comments were only possible until October 6. What it contains and which deadlines are running, we wrote up on October 2 in the article on crypto tax before the cabinet.
For the sale of the coins themselves, Section 23 of the Income Tax Act still applies. If more than twelve months lie between purchase and sale, the gain stays tax free. Below that, an exemption limit of 1,000 euros a year applies to all private disposals taken together.
For a long time it was disputed whether staking extends the holding period to ten years. The Federal Ministry of Finance clarified in its letter of May 10, 2022 that it does not: the period is one year for staked coins too. For you that means the client switch at your validator has no tax consequences. Moving a delegation is not a sale, and it does not start a new period as long as the coins do not leave your stake account.
Since the EU regulation on markets in crypto-assets, providers addressing German customers need an authorisation as a crypto-asset service provider. For you that is the first point to examine when buying: is the provider listed in the register of the competent supervisory authority, or does it merely advertise with the term? The register is the only reliable source on that; a marketing claim on a home page is not.
On custody the paths diverge. If your SOL sit on an exchange, you are not delegating yourself, and the question of the validator client does not arise for you in the first place. In exchange you carry the custodian's risk. If you hold the coins in your own wallet and delegate from there, you have the choice of validator in your own hands, and with it the responsibility for keeping an eye on its client.
The common validator directories show the running software version for each operator, among them validators.app and Solana Beach. There you can see whether a validator is running Agave, the full Firedancer or still Frankendancer. An operator already on Agave 4.3 or on the full Firedancer has the switch behind it.
Anyone waiting for a date is probably waiting in vain. Alpenglow did not launch on mainnet on September 28, although many market participants had expected exactly that. Anza clearly contradicted the timing expectation at the time. We documented the episode on September 28 in the article Alpenglow is not coming on September 28, including the next window named there on November 9, which is explicitly not a launch date.
That is precisely where the inconvenience for delegators lies. The shutdown of Frankendancer hangs not on a calendar date but on the event of activation. A date you could note down does not exist. Anyone who waits until the switch falls finds out from the missing earnings.
(As of October 2, 2026. This article is not investment advice. Prices and fee structures change; check the terms with the provider before you buy.)
