These are the results from the 2026 EthStaker community staking survey, compared against the 2024 and 2025 surveys where questions matched.
This year's survey had 528 completed responses, down from 740 in 2025 and 1,023 in 2024. Of this year's 528, 382 came from people running validators with their own capital. The decline likely reflects lower promotional reach, survey fatigue, or less engaged stakers. Of the respondents, roughly 78 percent stake the majority of their ETH, most have run validators for years, and their client choices still lean toward minority software.
These threads stood out this year:
The anonymized survey data behind this report is published at github.com/ethstaker/staking-survey-data.
Data files
2026-survey-dataset.csv (n=528)2026-survey-demographics.csv2025-survey-dataset.csv (n=740)2025-survey-demographics.csv2024-meat-potatoes.csv (n=1,023)2024-demographics.csvColor key: 2024 2025 2026
These results come from the 2026 EthStaker staking survey. Treat them as directional: year-over-year shifts under about five percentage points are noise, so trust the larger gaps and the consistent multi-year trends. The data is self-reported, and respondents are a self-selected, engaged subset of stakers that skews toward solo and independent home operators and toward English-speaking, Western regions. Where a question existed in earlier surveys we compare 2024, 2025, and 2026. Otherwise we show the years available. Unless noted, charts filter to respondents who run validators on mainnet with their own capital, matching the methodology of the published 2024 and 2025 reports. Demographic charts use the full respondent base.
A few hardware and operational questions from 2025 aren't shown this year, so those comparisons are absent. Some newer questions were asked only of the respondents they applied to, so a handful of charts have smaller sample sizes, noted on each chart.
The survey was announced and distributed through EthStaker's channels; see the launch announcement. The anonymized survey data is published at github.com/ethstaker/staking-survey-data. The analysis was built in Python (pandas for the data work, Plotly for the charts). Open-ended responses were handled two ways: recurring concerns were grouped into themes with AI assistance and paraphrased, while the browsable comments are shown close to verbatim, edited only to remove names and identifying details.
Year over year on the foundational questions: who is running validators, since when, and on what. Where a question existed in all three surveys we show 2024, 2025, and 2026 side by side. Where it began later, we show the years available.
The respondents skew heavily toward veterans. Genesis and early post-Merge starters remain the largest groups, and the share of brand-new entrants stays thin in every survey. Read alongside the open-ended comments, where long-standing genesis stakers are well represented and several note that few newcomers are arriving, it recurs across the responses.
Small operators dominate and the picture barely moved. In 2026, 59 percent run between one and five validators and another 21 percent run six to fifteen, almost identical to 2025. The long tail of larger operators (100 or more) is small but real at 7 percent. This remains a survey of home and solo-scale operators rather than professional fleets.
Solo validators remain the backbone at about 75 percent of own-capital respondents, with Rocket Pool minipools the most common pooled choice (around 27 percent) and Lido CSM next (around 12 percent). Respondents could select more than one, so these overlap. DVT-based options such as SSV and Obol stay niche in this sample. The multi-node results in Part 7 say more about that.
Counting each respondent's top three reasons, benefiting the network leads in both years (about 74 percent), with minimizing trust assumptions a clear second (59 to 61 percent). Gaining the experience of running a validator and maximizing profitability follow close behind, each named by roughly 45 to 48 percent and little changed year over year. The issuance section returns to how stakers weigh the financial side.
Execution-layer diversity remains healthy in this community. Nethermind leads at about 45 percent, Geth sits near 33 percent, and Besu at 15 percent. Besu has slipped from 21 to 18 to 15 percent over the three years while Geth has edged up, but no single execution client approaches a supermajority among respondents.
Consensus is where the diversity story is less comfortable. Lighthouse is not only the plurality client but has grown its share, from 36 percent in 2024 to 38 percent in 2026, even as Nimbus rose to second place (about 21 percent) and Prysm slipped to 19 percent. Lighthouse's standing as a majority client across the network is why we added a direct follow-up question this year, covered in Part 7.
Conviction is steady and high. Around 78 percent of respondents stake the majority (66 to 100 percent) of their ETH, about the same in all three years. The doubts that surface elsewhere in the survey have not yet pushed people to pull capital out.
About 41 percent hold some liquid staking tokens (roughly 24 percent in meaningful amounts and 17 percent minimally), little changed from 2025. Running your own validators and holding an LST are not mutually exclusive, even though they are often framed as opposites.
How this community reads the threats to Ethereum staking, from centralization to client concentration to whether they feel represented, and how that sentiment has moved over three years.
Counting how often each concern lands in a respondent's top three, rather than only their single biggest, gives a steadier read. In 2026 privacy and exposed IP (33 percent) and lack of tax clarity (32 percent) lead, with supermajority client risk (29 percent) close behind. The shift from 2025 is real: tax climbed about ten points and privacy about seven, while the concerns that led 2025, risk to protocol or price and internet or bandwidth, slipped down the list. The worry is drifting away from keeping a validator online and toward the legal and personal exposure of running one.
Concern about stake centralization has eased slightly but steadily, from a mean of 4.11 in 2024 to 3.95 in 2025 to 3.82 in 2026 on a zero-to-six scale. The softening is modest and should not be read as complacency: as the written responses make clear, many respondents have simply shifted their anxiety from clients toward exchanges, corporate treasuries, and liquid staking providers.
Worry about a client supermajority held essentially flat year over year (3.55 versus 3.57, and the 2024 survey did not ask it on this scale). It remains a live concern rather than a solved one, and it reappears unprompted in the written responses, often paired with the argument that the burden of fixing client diversity should not fall on small home stakers.
Views on what solo staking is worth to Ethereum as a protocol cooled after 2024 and have since plateaued. The share calling it more favorable fell from 48 percent in 2024 to 38 percent in 2025 and stayed at 38 percent in 2026, with most of the rest landing on neutral.
The individual picture is harder. Only 22 percent now see solo staking as more favorable for themselves personally, while 39 percent see it as less favorable, close to the mirror image of how respondents talk about the network benefit. People increasingly believe in solo staking for Ethereum more than they believe it pays for them. The issuance debate is where that tension surfaces.
Stakers feel only moderately heard, and that has not improved. The mean rating for how well their interests are represented in protocol research has sat at roughly 3.1 out of 6 for three straight years (3.18, 3.17, 3.12). Three years flat reads as settled, not a dip.
Sentiment about advocacy was broadly stable, with at most a slight lean toward pessimism that sits within the noise: the share who feel advocates are effective slipped from 29 to 25 percent, and those who feel proposals largely ignore solo stakers nudged from 17 to 19 percent. The largest single group (28 percent) feel advocates exist but are captured or powerless, and very few call advocacy malicious. The prevailing mood is doubt about its effectiveness rather than its intent.
What it actually takes to run a validator day to day: time spent and the network connection behind it.
Day-to-day burden is modest and stable. The median respondent spends about two hours a month on maintenance and troubleshooting, with a mean near four hours pulled up by a small number of heavy cases. Three years of nearly identical figures suggest that, for those who stay, a validator is not especially time-consuming once it is running. The friction shows up at setup and during upgrades, not in routine months.
For reference, EIP-7870 ("Hardware and Bandwidth Recommendations") puts a validator's needs at roughly 50 Mbps down and 25 Mbps up for attesting duties, and about 100 down and 50 up to build and propagate blocks locally. Against that bar, own-capital respondents in both 2025 and 2026 sit well above the requirement on advertised and maximum-available upload and download alike, and the distributions barely moved year over year. Bandwidth is rarely the binding constraint for this group. Where connectivity does come up in the comments, it is usually about upload limits or rural and unreliable service rather than headline speed.
Linux remains overwhelmingly the platform of choice, at 82 percent in 2026 and up slightly from 80 percent in 2025, with DAppNode appliances a distant second near 13 percent and Windows, Avado, and MacOS each marginal.
Fallback nodes remain a minority and may be ticking up, run by 26 percent in 2026 versus 23 percent in 2025 and 22 percent in 2024. Each step is small enough to be noise on its own, but the direction has been consistent for three years, which hints that operators are investing a little more in resilience over time.
Smoothing-pool participation has held steady near 28 percent, but the share unfamiliar with the concept roughly halved, from 13 percent in 2025 to 6 percent in 2026. Awareness appears to be rising even where adoption is flat, which suggests reward-smoothing tooling is reaching more operators.
The money question: are people still adding stake, what yield would make them stop, and how do they think about issuance?
Appetite to add stake softened. No is the single most common answer and rose to 34 percent, undecided climbed to 26 percent, and the share actively planning to add with no plans to stop fell from 22 to 19 percent. Read together with the yield comments below, this is a community holding its position rather than expanding it.
Where respondents name a yield at which they would exit, the median has held near 2 percent across all three years, though the sample is small (72 in 2026) and self-interested, as prior reports cautioned. The more useful signal is in the written responses, where a recurring line is that solo-staking yield has already fallen below inflation and ordinary savings or TradFi alternatives.
The threshold at which respondents would stop adding new stake also centers around 2 percent, on a very small base (20 responses in 2026). We show it for continuity with prior years rather than as a precise estimate.
Issuance was the single largest theme in this year's open-ended responses, and it is contentious enough to deserve its own treatment. We asked directly whether respondents factor ETH supply inflation into their decision to stake, and we read every written comment that touched issuance, yield, or dilution. For a deeper walk through the arguments on each side, see issuance.wtf and EthStaker's ethdebate.com/issuance.
The share who actively factor ETH supply inflation into their staking decision has risen steadily, from 21 percent in 2024 to 35 percent in 2025 and 39 percent in 2026, while the share who have never considered it has fallen. Awareness of the issuance question is clearly growing. Of the roughly six in ten in 2026 who have not actively weighed it, nearly half say they would like tools to understand its effect on their yield, rather than that it does not matter. (The 2024 question was worded slightly differently but offered the same response options.)
When issuance comes up unprompted in the open-ended responses, it is the most polarized topic in the survey and the single largest theme (roughly 50 of the 251 written comments touch issuance, yield, or dilution). About half of those weigh in on the issuance curve itself, and they split sharply:
Counting each respondent's top three onboarding sources, EthStaker's own docs and community are the most widely used (about 60 percent in both years), with Somer Esat and CoinCashew style written guides next (about 46 percent) and Ethereum.org close behind (39 to 41 percent). The written comments add a note of fragility: several respondents worry that key guides have gone stale or that their maintainers have gone quiet, which matters when one of those guides is often the first thing a newcomer follows.
For keeping up with protocol and ecosystem news, no single channel dominates. Twitter or X is the most commonly named source in both years (about 38 to 39 percent), with Reddit and various Discords close behind and roughly interchangeable, Discords edging ahead in 2026. A notable minority say they mostly do not keep up, which lines up with the sizable group who told us they would like clearer, less jargon-heavy communication about where the protocol is heading.
Asked how they keep up with validator updates, respondents lean on GitHub and client release channels alongside community spaces like Discord, and the mix is stable across 2025 and 2026. The 2024 survey asked a narrower version (no GitHub or RSS option), so its bars cover only the shared channels, Discord, Reddit, Twitter, and block explorers, which have held roughly steady since. For client teams, GitHub release notes and Discord are where this audience looks.
Demographics use the full respondent base (not just own-capital validators), and bars show percent of respondents so years with different totals stay comparable. Gender, technical background, and crypto employment were asked in the same form in all three surveys and are shown for 2024, 2025, and 2026. The detailed region breakdown is shown for 2025 and 2026 only, because the 2024 survey used a coarser region list (a single "Europe" bucket, "Australia" rather than "Oceania"); a broad continent-level region chart below includes 2024.
Geography is stable and concentrated. North America (about 35 percent) and Western Europe (about 29 percent) account for roughly two thirds of respondents in both years, with the rest spread thinly across other European regions, Oceania, and beyond. The chart shows the eight largest regions; smaller ones are grouped as Other, with the full list below. (2024 is not shown: that survey used a coarser region list, a single Europe bucket rather than the Western, Northern, Eastern, and Southern split used since 2025, so it is not directly comparable.)
| Region | 2025 | 2026 |
|---|---|---|
| North America | 39.6% | 35.3% |
| Western Europe / UK | 26.3% | 29.0% |
| Eastern Europe | 5.2% | 6.1% |
| Southern Europe | 4.4% | 5.3% |
| Northern Europe | 5.7% | 5.0% |
| Oceania | 4.7% | 3.6% |
| Southeast Asia | 2.0% | 3.4% |
| South America | 2.2% | 2.9% |
| Not sure / I'm not the operator | 1.2% | 2.5% |
| East Asia | 4.2% | 2.1% |
| Western Asia | 1.6% | 1.3% |
| Southern Africa | 0.7% | 0.8% |
| West Africa | 0.1% | 0.6% |
| Central America / Caribbean | 0.6% | 0.6% |
| Central Asia | 0.3% | 0.6% |
| Central Africa | 0.1% | 0.4% |
| East Africa | 0.3% | 0.2% |
| South Asia | 0.7% | 0.2% |
Collapsed to continents, geography is stable across all three years. North America and Europe together make up roughly two thirds to three quarters of respondents, with Asia and Oceania the largest of the rest. This broad view is the only regional cut available for 2024, whose survey used a coarser region list than the detailed 2025 and 2026 breakdown above.
Respondents remain a consistently technical crowd across all three years. The share calling themselves strongly technical has held in the mid-50s (56 percent in 2024, 57 percent in 2025, 54 percent in 2026), with about a third somewhat technical or hobbyist and roughly 10 percent not technical. The barrier-to-entry theme in the comments is consistent with this: even this self-selected, technical group repeatedly describes solo staking as too hard for the average newcomer.
Most respondents do not work in crypto, and the industry-insider share has thinned since 2024. Those employed in crypto full-time fell from about 20 percent in 2024 to 15 percent in 2025 and 17 percent in 2026, and those in a related field declined steadily from 14 to 11 to 9 percent. Across all three years 56 to 60 percent work in an unrelated field. (The 2024 survey had no separate "unemployed" option, folding it into the unrelated-field answer, so that category is comparable only from 2025 on.) This is largely a community of motivated outsiders rather than industry insiders, a background that tends to make them wary of proposals that appear to favor professional operators.
The respondent base is overwhelmingly male across all three years, around 95 percent (95 percent in 2024, 96 percent in 2025, 95 percent in 2026), with women and non-binary respondents together about 4 to 5 percent. It is the least diverse dimension of the survey, and one respondent raised the point unprompted this year.
Questions added or substantially changed for 2026: multi-node setups (DVT, Vero, Vouch), blob throttling, tax handling, Lean Ethereum readiness, AI in core development, and the qualitative "why Lighthouse" follow-up.
Multi-node and distributed-validator setups remain rare in this community. Ninety percent run a single-node setup, 7 percent use a multi-node arrangement such as DVT, Vero, or Vouch, and 3 percent are unsure. The small Yes group is why the client breakdowns that follow have low response counts: they describe a real but narrow slice of operators.
Among the small number of multi-node operators, client choices are spread rather than concentrated. With a base this small (see n on each chart), read these as indicative only.
We asked the 126 respondents who run Lighthouse as their primary consensus client why they continue to do so even though it is a majority client. 108 answered, and three reasons dominate. Loyalty is not really one of them.
The most common reason is the cost and risk of switching. Migrating a validator carries slashing risk, a multi-day re-sync, and the chance of a costly mistake, and many simply have not found the time or the nerve. The second is reliability: Lighthouse has a long, clean track record for these operators, and several say plainly that they will not touch a configuration that has never failed them. The third is history: a large group adopted Lighthouse before it was a majority client, often at or near genesis, and never re-evaluated.
Two smaller but important strands run through the responses. Around a dozen respondents push back on the premise, arguing that diversification is the responsibility of large professional operators rather than small home stakers, especially while no client is near the 67 percent danger line, and a number say they already run a minority client on the execution layer or on backup nodes to offset their choice. A handful did not realize Lighthouse had become a majority client, and said the question itself would prompt them to switch.
That means client-diversity campaigns aimed at this group will succeed or fail on switching friction. Safe, fast migration tooling, clear guidance, and confidence that re-syncing will not cost them will matter more than further arguments about why diversity is important.
Matches all words you type. Use -word to exclude, "a phrase" for an exact phrase, and commas (a, b) to match either.
Scrubbed of personal information and trivial non-answers. 105 of 108 responses shown.
Blob throttling is not on most operators' radar. Sixty-two percent do not throttle blobs in their node configuration, just 3 percent do, and 36 percent do not know, the highest "don't know" rate of any operational question this year.
There is no consensus on how to treat staking income for tax. The largest group reports it as self-employment or business income (39 percent), a third treat it as hobby income (33 percent), and nearly a quarter are not reporting it currently (23 percent). Lack of tax clarity was the second most common practical concern this year, and the split likely reflects that: without settled guidance, stakers make different choices to fit their own circumstances, which vary with stake size, jurisdiction, and how they earn and hold their ETH.
When it comes to tooling, respondents lean on a small set of crypto tax services and the community-built ethstaker.tax, but a meaningful number track income manually or not at all, consistent with the cadence results below.
Most who track do so annually (42 percent), with a long tail of monthly, quarterly, and daily approaches, and about a fifth not tracking at all. Combined with the treatment split above, taxes are clearly an area where better guidance would be welcomed.
Interest in Lean Ethereum is cautious and conditional. Three quarters of respondents say "maybe, it depends on the requirements" when asked about running a prover, with only 6 percent a firm yes and 20 percent a no. Most respondents are waiting to see what running a prover would actually demand before committing either way, and the open-ended comments say the same: several had not heard of Lean Ethereum at all.
Willingness to help test is meaningfully higher than willingness to commit. Twenty-two percent would run a Lean Ethereum testnet outright and another 49 percent might, a far warmer response than for running a prover in production. See the Lean Ethereum roadmap for where those requirements are taking shape.
Attitudes toward AI in core development are supportive but guarded. A combined 91 percent are positive, but most of that (58 percent) is support with concerns rather than unqualified support (33 percent), and 9 percent are opposed.
Two optional free-text questions invited respondents to raise protocol-research areas and any staking concerns we had not covered. Together they drew 251 written responses. We have grouped them by theme below, with approximate counts, and have paraphrased rather than quoted in order to protect respondent privacy. A small number of responses contained personal information and have been excluded entirely. Issuance, the largest theme, is treated separately in Part 4.
The recurring themes, roughly in order of how often they appeared:
Taken together, the written responses describe a committed group that is proud of what it does and increasingly unsure the economics and the roadmap are being built with it in mind.
You can browse the responses yourself below, now spanning the 2024, 2025, and 2026 surveys (use the year buttons to filter). These have been scrubbed of personal information and trivial non-answers.
Matches all words you type. Use -word to exclude, "a phrase" for an exact phrase, and commas (a, b) to match either.
Scrubbed of personal information and trivial non-answers. 434 responses across 2024 to 2026.
Matches all words you type. Use -word to exclude, "a phrase" for an exact phrase, and commas (a, b) to match either.
Scrubbed of personal information and trivial non-answers. 478 responses across 2024 to 2026.
This survey is best read as a directional snapshot of an engaged, self-selected slice of solo and home stakers, not a precise census. In that context, the following points of pressure stand out for where to guide research and advocacy next.
Economics. This year's results show an engaged group that is no longer sure the numbers still work. Yield is perceived to have fallen below inflation and ordinary alternatives, appetite to add stake has softened, and the issuance debate has become the most charged topic in the survey precisely because respondents see it as existential to their continued participation.
Centralization concern is shifting. Concern about a client supermajority remains live but flat, while worry has migrated toward exchanges, corporate treasuries, liquid-staking providers, and a growing pool of unattributed stake. Respondents repeatedly argue that the burden of client diversity should not rest on small home stakers.
Representation. For three years running, this group has felt only moderately heard, and sentiment about advocacy has drifted toward doubt about its effectiveness. The enthusiasm gap between solo staking's value to the network and its value to the individual is widening.
Accessibility and operations. The 32 ETH barrier, non-technical setup, stale guides, re-sync pain, and hardware cost recur throughout the written responses. Even this technical cohort describes the path as too hard for newcomers, and brand-new entrants are few in number among the respondents.
The overall picture is a small, committed group that still believes in solo staking but is no longer sure the economics and the roadmap are being built with it in mind. Keeping that group is a choice the protocol still gets to make. We'll keep running surveys like this one to monitor and help the community adapt its efforts.
Contact: team@ethstaker.org
Show the full breakdown (all sources, share in top 3)