More public companies, funds, and crypto-native firms are putting ETH on the balance sheet. Once the asset is there, the next question is what to do with it.
A report released by Kairos Research with Golem Foundation and Ethereum Institutional says staking is still the most direct answer, but the payoff has weakened as more ether has moved into staking. Lido’s average yield has fallen from 4.5% in 2022 to about 2.2% now, and the main liquid staking tokens are all clustered between 2.1% and 2.3%.
The report maps the holdings and allocation choices of major ETH treasuries, then compares the return and risk profile of staking, restaking, lending, first-loss protection, and covered call strategies. Its central finding is simple: in ETH terms, there are not many ways to beat staking, and the options that do come with clear trade-offs.
Key findings from the report
The report highlights several headline points.
- Staking yield has been cut roughly in half. As staked ETH grew about 3.5x to 45.2 million ETH, Lido’s average yield fell from 4.5% in 2022 to about 2.2%.
- ETH treasuries now account for a meaningful share of supply. Public companies hold 8.03 million ETH, equal to 6.6% of supply, while U.S. spot ETFs hold about 6.10 million ETH, or 5.0%.
- Across the 28 treasuries tracked, about 74% of visible ETH allocations have been staked.
- In ETH terms, lending yields 1.3% and liquidity provision yields 1.2%, both below staking at 2.2%.
- The only strategies with higher reported returns are first-loss protection at 4.8% and covered calls at 7.6%, each with a different risk profile.
- Security is the biggest single concern. There have been 328 hacks so far in 2026, with $8.78 billion stolen, versus 148 hacks and $2.72 billion stolen in all of 2025.
- Holding ETH is also a currency position. Since November 2018, one-year ETH holding returns have ranged from -92% to +2,142%, with a median of +14%. About 45% of one-year holding periods were negative.
All service providers surveyed described staking yield as 「a base, but not enough on its own」. Together with the treasury companies interviewed, those respondents hold or manage 7.10 million ETH worth about $16 billion.
The institutional debate has moved on
The report says ETH has become a balance-sheet asset for a wide range of organizations over the past few years. Public companies, funds, and crypto-native institutions now hold large amounts of ETH, and most of them stake it. The appeal is straightforward: staking offers a native protocol yield that few other assets can match.
But as staking returns have declined, the discussion has changed. The first phase of institutional adoption focused on whether ETH belonged on the balance sheet. The next phase is about how to use that exposure more effectively.
Golem Foundation has managed its own ETH treasury since 2019, and the report draws on that operating experience. Ethereum Institutional said it has spent significant time discussing these issues with banks, asset managers, and companies, along with broader topics such as tokenization and settlement on Ethereum. Public research, it said, helps move those conversations forward.
Where the ETH sits today
To understand the available yield opportunities, the report first looks at supply distribution. Treasury holders now control a meaningful share of ETH supply.
As of Oct. 31, 2026, public companies held 8.03 million ETH, worth about $18.1 billion at an ETH price of $2,253.47, equal to 6.6% of total supply. U.S. spot ETFs held about 6.10 million ETH, or 5.0% of supply. DAOs and foundations tracked in the report held 0.6% of supply.
Roughly 45.2 million ETH is staked across the network, or 37% of supply, including 16.8 million ETH represented by liquid staking tokens. By comparison, only about 3% of ETH is deposited into lending protocols and 2% is used in restaking.
How 28 treasuries are allocating ETH
Among the 28 treasuries tracked, about 74% of visible ETH has been staked. Of that, 51% is in native staking and 23% is held through liquid staking tokens. About 21% is unstaked and about 5% is used in lending.
Another amount equal to roughly 12% of total holdings sits off-chain, mostly tied to Mantle, with no public disclosure on its use. The report notes that the figures are rounded.
Allocation choices vary widely. Lido, Gnosis, and Golem Foundation all stake 95% or more of their ETH. Gnosis has also borrowed about $25 million worth of ETH against staked ETH. Arbitrum stakes 89% of its ETH and ENS stakes 83%.
Mantle has staked almost all of its on-chain ETH, but about 71% of its ETH is held off-chain and its use has not been disclosed. The Ethereum Foundation is the largest holder in this group with about 191,000 ETH: 36% is staked, 13% is lent, and about half is unstaked.
The report gives a simple way to frame the opportunity cost. At a 2.2% yield, every 100,000 idle ETH leaves about 2,200 ETH a year on the table.
Julian Zawistowski, founder and executive director of Golem Foundation, said: 「From day one, we have had strong conviction in Ethereum. Since Golem Foundation was established in 2019, our treasury strategy has been long-term ETH holding, and we have staked ETH and run our own validator nodes for years. Staking has played an important role in our sustainability and has allowed us to provide 5,273 ETH in funding through Octant to users and some of the best teams in Ethereum. I believe the phase of simply holding and staking ETH is over. We need an Ethereum economy denominated in ETH. As staking yields declined, we realized that relying on staking alone would limit what we can do. We are committed to diversifying into lower-risk tools that preserve long ETH exposure, and we are actively exploring how ETH treasuries can be made more effective.」
The report adds that Arkham data captures only a small part of the ETH treasury universe. L2s and exchanges also hold large amounts of ETH. One example is Arbitrum DAO, which holds about 16,300 ETH worth roughly $37 million.
Bitmine and Sharplink show two different treasury models
Among corporate holders, Bitmine holds 6.05 million ETH. Of that, 5.23 million ETH, or 86%, is staked through native validator nodes rather than liquid staking tokens. Most of that staking is done through MAVAN. Bitmine built that validator network after acquiring node operator Pier Two in March 2026.
Bitmine disclosed a staking yield of 2.58%, generating about $318 million in annual staking revenue. Staking income accounted for 98% of its total revenue in the prior quarter. The company did not disclose any restaking, lending, liquidity provision, or options positions.
Sharplink has taken a more diversified route. About two-thirds of its ETH is in native staking and one-third in liquid staking. Since inception, the company has deployed $200 million of ether.fi weETH to Linea and EigenCloud, holds about 106,000 ETH of wstETH through Lido worth about $200 million, and has allocated $100 million of staked ETH to an on-chain yield fund managed by Galaxy. As of Sept. 29, 2026, Sharplink had earned 28,351 ETH in cumulative staking rewards.
Young Kim, Bitmine’s chief operating officer and chief financial officer, said: 「As the world’s largest corporate ETH holder, Bitmine has added to its position every week since launching its treasury strategy in June 2025, and we are approaching our goal of holding 5% of total ETH supply. Through MAVAN, we stake our own ETH alongside client ETH and other digital assets, securing the network while providing critical infrastructure services to the ecosystem. We are also investing in efforts that drive commercialization across Ethereum and expect to create value for shareholders. Those investments include moonshot-style projects such as Beast Industries and Eightco Holdings, as well as Ethlabs, Ethereum Institutional, and EthSystems, which were recently incubated out of the Ethereum Foundation. Their shared goal is to help the network serve institutions at scale.」
Joseph Chalom, chief executive officer of Sharplink, said: 「Over the past year, we have seen stablecoins, tokenized assets, DeFi, and AI Agent activity arrive on Ethereum at the same time, creating an economic explosion that will drive unprecedented growth. As the world’s second-largest corporate ETH holder, we see 2027 as an opportunity to use permanent capital on our balance sheet to accelerate that shift through targeted investments while returning value to shareholders. We stake ETH to help secure the network, provide first liquidity to new protocols, and help fund organizations such as ETHLabs, Ethereum Institutional, and ETH Systems so Ethereum is ready for the next wave of users and institutions.」
Staking yield has fallen, and restaking has not filled the gap
The report treats staking as the default source of yield on Ethereum and the benchmark other strategies need to beat. That benchmark is moving lower.
Since 2022, the amount of ETH staked has grown about 3.5x, from an average of 12.9 million ETH that year to 45.2 million ETH now. As more ETH participates in staking, the network staking yield declines, which means yields could fall further.
All major liquid staking tokens are currently yielding between 2.1% and 2.3%.
Restaking has not become a meaningful source of extra return. About 2.6 million ETH is currently restaked, equal to 2.1% of total supply, almost all of it on EigenLayer. That is down from roughly 5 million ETH in mid-2024. Symbiotic holds about 60,000 ETH.
Liquid restaking tokens yield 2.2% to 2.7%, versus 2.2% for plain staking. The report says that, excluding points and EIGEN incentives, the extra return for taking on additional slashing and liquidity risk is limited.
Much of the earlier restaking premium came from EIGEN airdrops, points programs, and other token incentives rather than fees paid by services using restaked ETH. As those incentives faded, the yield premium over standard staking largely disappeared, and the amount of restaked ETH fell by about half from its 2024 peak.
What can beat staking in ETH terms
For ETH treasuries, the report argues, the right benchmark is whether a strategy increases the amount of ETH held, not just the dollar value of the position.
On that basis, the current menu looks like this:
- Staking: about 2.2%
- Lending: 1.3%
- Liquidity provision: 1.2%
- First-loss protection: 4.8%
- Covered calls: 7.6%
Lending and liquidity provision both trail staking. The only strategies that clearly beat it are first-loss protection and covered calls, and each introduces a different kind of risk.
Basis trades and U.S. Treasury trades generate dollar returns and remove ETH exposure. The report gives one example: borrowing USDC at 4.50% to buy U.S. Treasuries yielding 4.09% results in a loss.
On Aave, lending WETH currently yields 1.3%, with a 2026 average of 1.7%, about 0.9 percentage points below staking the same amount of ETH.
Aave Umbrella offers 4.8% on WETH, but deposited assets can be slashed to cover bad debt. The extra yield is compensation for taking first-loss risk.
Covered calls look stronger on paper. A 42-day call option struck 10% above spot currently collects premium equal to 3.2% of the ETH position. ETH implied volatility is 51%, below the average for every year since 2021, which makes option premiums relatively cheap.
Since 2021, selling a monthly call 10% above spot produced a 17.1% annualized return in dollar terms, versus 8.8% annualized for simply holding ETH. But the strategy lags badly in strong up markets. In 2024, it returned 9%, while holding ETH returned 53%.
The risk side of the ledger
The report says every step beyond staking adds risk somewhere. The right mix depends on how much ETH a treasury may need to sell over the next 12 to 24 months and how much loss it can tolerate.
- Liquid staking brings depeg risk and exit queue risk.
- First-loss protection brings solvency risk and bad-debt risk.
- Covered calls cap upside and, on some venues, add counterparty risk.
- Lending brings smart contract risk, liquidity risk, and bad-debt risk.
- Restaking ties slashing risk to external services and has weaker exit liquidity.
- Basis trades carry exchange counterparty risk. Over the past year, 16% to 18% of hourly funding rates were negative.
Across all of these, security stands out as the biggest single concern.
There have been 328 hacks so far in 2026, versus 148 in all of 2025. Losses reached $8.78 billion in 2026 to date and $2.72 billion in 2025. On Oct. 29, 2026 alone, Binance lost about $6.36 billion in what the report describes as the largest crypto asset theft on record. The attackers had previously compromised its hot wallet signing infrastructure.
Excluding the Binance incident, losses from private key or credential compromise and cross-chain bridge exploits were each about $700 million.
Holding ETH also means carrying currency risk
Most treasuries pay salaries and vendors in dollars, so holding ETH is not just an asset allocation decision. It is also a currency position.
Since November 2018, one-year ETH holding returns have ranged from -92% to +2,142%, with a median of +14%. About 45% of one-year holding periods were negative, and 14% lost more than half their value.
Over the past year, staked ETH earned 2.4% in ETH terms but lost 36.8% in dollar terms. The report says ETH treasuries need a performance framework denominated in ETH.
ETH’s volatility over the past year was 61%, or 1.4x bitcoin, 3.1x the Nasdaq 100, 4.8x the S&P 500, and 2.1x gold. Since 2018, ETH has moved more than 5% in a single day on 19% of trading days. For QQQ and SPY, that figure is about 1%.
At the same time, the report says ETH has become less volatile over time. From 2018 through 2022, annual average volatility was about 96%. Since 2023, it has fallen to about 62%. The share of days with moves greater than 5% also dropped from 17% to 32% a year in the earlier period to 7% to 17% a year more recently.
That does not remove drawdown risk. From the high in August 2025 to the low on June 25, 2026, ETH still fell 68%. In 2025, ETH volatility was 1.8x bitcoin, the widest gap since 2018.
Against other major L1 tokens, ETH’s maximum drawdown since 2021 was 79%, versus 96% for SOL and 71% for BNB. HYPE, listed in November 2024, saw a maximum drawdown of 68%. As of Oct. 20, 2026, ETH was down 50% from its high, SOL 60%, BNB 44%, and HYPE 18% from its September 2026 high.
Over the past year, volatility was 61% for ETH, 64% for SOL, 45% for BNB, and 90% for HYPE. Since 2025, the figures were 69% for ETH, 76% for SOL, 49% for BNB, and 98% for HYPE.
Dollar yields on-chain are also falling
The report uses on-chain dollar yields as a comparison point for ETH strategies. Lending, basis trades, reinsurance, and GPU-backed credit can all produce returns above U.S. Treasuries, but those yields are also moving lower as subsidies fade and more capital enters the trade.
Current on-chain dollar yields range from 3.3% on Aave USDT to 11.0% on OnRe reinsurance on Solana. Tokenized U.S. Treasury products from BlackRock and Ondo yield 3.7% to 3.8%. USDai offers 6.2% on GPU-backed loans on Arbitrum.
Aave USDC yields 3.8% and USDT yields 3.3%, about half their 2024 averages of 7.2% and 6.3%. sUSDe yields 5.1%, but its yield ranged from 0.1% to 5.7% over the past year and unstaking takes seven days.
Basis trade returns have also compressed. As more capital entered the trade, Hyperliquid ETH funding fell from 22% in 2023 and 2024 to 5.9% in 2026, while sUSDe yield dropped from 17.5% to 4.2%. The report notes that part of that decline coincided with the launch of ETH futures ETFs in October 2023 and spot ETH ETFs in July 2024, which created new venues for basis traders.
Over the past year, 16% to 18% of hourly funding rates were negative. Deribit dated futures currently lock in annualized returns of 4.3% to 4.7%.
Why institutions still want ETH exposure
The final section turns from market and on-chain data to the investment logic behind large ETH holders. Treasury companies broadly argue that as more assets move on-chain, demand for ETH should rise with them. Bitmine chairman Tom Lee views Ethereum as the settlement layer for stablecoins and tokenized assets.
The report tests that thesis through three lenses: stablecoins, tokenized stocks and real-world assets, and L2s as distribution channels.
Stablecoin supply has grown from $27 billion in January 2021 to $302 billion now, up about $100 billion since January 2025. Ethereum and its L2s account for 51% of that total.
Tokenized stocks reached $3.52 billion as of Oct. 26, 2026, up 11.6% over 30 days, with 4.43 million holders. bStocks, Ondo, and xStocks lead across multiple chains. Robinhood has issued tokenized products covering 190 stocks, with a market size of $147 million.
Robinhood Chain as a direct source of ETH demand
The report points to Robinhood Chain as a direct example of an application creating demand for ETH.
The chain’s bridge holds 262,000 ETH worth about $590 million. Before its public launch on July 1, 2026, the figure was below 1,000 ETH. It peaked at 313,000 ETH on Sept. 16.
Robinhood Chain uses ETH as its gas token, which means every user trading on-chain needs to hold ETH. Since July, users have paid about 21,900 ETH in fees.
The report says Ethereum currently captures value mainly through ETH being used as gas and held on-chain, rather than through settlement fees. Over the longer term, it says, Ethereum could capture a larger share of value from L2s.
Governance and security remain the main constraints
The report closes by saying that while Ethereum and the broader crypto market are showing several positive trends, staking is no longer a complete answer for investors building around ETH.
Staking yield has been cut in half since 2022, down to 2.2%. Restaking has not replaced that lost return, and in ETH terms both lending and liquidity provision still trail staking. The only strategies that clearly beat the benchmark are first-loss protection and covered calls, but one puts capital in the first line of loss absorption and the other gives up part of the upside.
Surveyed service providers agreed that staking yield is 「a base, but not enough on its own」. They also said internal governance is holding back broader on-chain deployment. Respondents cited smart contract risk, board approval, and the lack of clear standards as the main obstacles. Most said that making boards or regulators more comfortable would do more to unlock deployment than simply offering higher yields.
Some of the largest holders are not staking ETH at all. The report says that may reflect ETF investment rules or the fact that staking remains in a regulatory gray area in places such as South Korea. Given the rise in exploits and hacks so far in 2026, caution is understandable.
Still, participants interviewed for the report pointed to constructive trends. Kam Benbrik, head of on-chain research at Bitwise, said: 「By 2027, people will spend less time asking how to tokenize assets and more time asking what can be done once assets are tokenized. Services such as financing and collateral management, which have traditionally been provided by prime brokers, may become easier to access on-chain. That could lower barriers to entry and open these services to a broader set of investors.」
Joshua Lim, co-head of global markets at FalconX, said: 「Throughout the year, growth in on-chain RWA tools across perpetual and spot markets has been a key trend. Looking ahead to 2027, we expect to expand support for clients in these markets.」
Charles d'Haussy, chief executive officer of dYdX Foundation, said: 「As qualified custodians and traditional financial banks are able to offer more RWA, and as distribution improves, we have been able to increase our allocation to tokenized assets materially.」
The report says regulatory clarity and better on-chain security will still take time. In the meantime, one major question for 2027 is whether the investment case behind holding ETH will be validated. Treasury companies are betting that as assets move on-chain, demand for ETH will rise as well. The report says the market is already seeing several signs: stablecoins continue to grow across different market conditions, Robinhood has launched on-chain stocks through an Ethereum L2, and products such as OnRe, USDai, and Ethena’s tokenized basis trade show demand for on-chain yield opportunities.


