⏱ 2 min read
A Sept. 21 snapshot shows ~$120B staked ETH and ~$40.4B on L2s, while Farside reports >$140M US spot ETH ETF outflows over Sept. 15–18. Here’s why those numbers aren’t additive and what to watch for real demand.
Ethereum’s institutional data hub showed roughly $120 billion in staked ETH and a daily average of $40.4 billion in assets on layer-2 networks in a Sept. 21 snapshot. Over the same week, US-traded spot Ethereum ETFs recorded more than $140 million in net outflows from Sept. 15 to 18, according to Farside Investors.
These figures are not contradictory—and they are not additive. Staked ETH is a stock of coins committed to network security. L2 total value locked (TVL) is a stock of assets held on rollups. Adding them together does not measure fresh ETH demand.
Mechanism: why these numbers don’t equal inflows
Staking involves depositing ETH to activate validators and earn rewards. An address can stake coins it already holds, or it can buy ETH specifically to stake. A rising staking balance therefore reflects participation and price, not necessarily new buying. Expressed in dollars, the balance also moves with ETH’s valuation.
Liquid staking tokens (LSTs) further decouple commitment from liquidity. The underlying ETH remains staked while the exposure can be transferred, which changes who holds the claim without changing the amount staked.
L2 TVL is also a stock. The connection to ETH holders runs through what L2s pay Ethereum for settlement—operator on-chain costs for posting transaction data, blobs, proofs, and state updates. L2BEAT’s on-chain-costs series tracks those operator payments. They are distinct from the fees end-users pay to an L2, and blob spending alone is not the whole settlement bill. Even total operator spending is not identical to ETH burned.
▲ 5.40%
ETF flows are one channel, not the market
Farside’s table shows the ETF channel moved around last week: $121.1 million of inflows on Sept. 14; cumulative outflows of $405.4 million from Sept. 14 to Sept. 17; and $143.7 million of fresh inflows on Sept. 18. Net outflows exceeded $140 million across Sept. 15–18. Large on-chain stocks (staked ETH, L2 assets) can coexist with redemptions in a particular investment wrapper.
What this does and doesn’t tell us
Fact: The snapshot captures big pools of staked ETH and assets on L2s. Fact: ETFs saw net redemptions over several sessions. Analysis: Neither number, on its own or combined, measures marginal ETH demand. A better signal comes from flow measures directly tied to ETH units and settlement usage.
What to watch next
- ETH-denominated staking flows (new stake and withdrawals) rather than dollarized staking balances.
- L2BEAT’s operator on-chain costs trend (calldata, blobs, proofs, overhead) as a settlement demand proxy.
- Ongoing US spot ETH ETF net flows by day and issuer to see whether redemptions persist or reverse.
ETF flows cannot settle the question of total ETH demand, and neither can stock snapshots. Use channel-specific flow data to triangulate where the marginal buyer is—or isn’t.
This content is for informational purposes only and does not constitute financial advice.
🧠 HafidWatch Take
The headline figures—approximately $120 billion in staked ETH and $40.4 billion in L2 assets—are stock measures that indicate adoption and valuation but do not reflect net new buying. These values can coexist with the reported US spot ETH ETF outflows exceeding $140 million, as each metric represents a different channel. More accurate indicators of marginal demand are ETH-denominated staking flows and L2 operator on-chain costs, rather than dollarized pool sizes. Monitoring these ETH-native flows alongside ETF activity will help determine whether buyers are moving outside ETFs or if there is a broader demand slowdown. Current data do not clarify this, so it is essential to focus on flow trends rather than stock snapshots.
Daily crypto intelligence. Before the market opens.
Including the Divergence Index — the sentiment gap no other newsletter tracks. Free, every morning at 7:30am ET.
✓ Free forever · ✓ No spam · ✓ 50+ sources monitored
