On October 21, Grayscale quietly widened in-kind creation to ten of its funds. No press conference. No candle worth screenshotting. The list is the news: BTC and ETH wrappers, staking vehicles for ETH, Avalanche, Solana and Sui, a Hyperliquid product, a Dogecoin trust, a Zcash trust, a Chainlink trust.
Check the chain, ignore the noise. Here the chain is the operational plumbing behind the wrapper, not a token ledger — and that plumbing just changed shape.
I have spent the better part of a year auditing how these wrappers actually move assets, not how they are marketed. The detail that matters is not that Grayscale offers "more choice," which is the issuer's own framing and should be stripped off before analysis. It is that a crypto ETF has been pulled back from a special accommodation into the standard structure that commodity ETFs have used for two decades. GLD never sold gold to raise cash for a creation. It took the gold.
For most of the spot-crypto ETF era, creation meant cash. An investor holding BTC had to sell it on the spot market, convert to dollars, wire the dollars to the trust, and receive shares. That sequence is not neutral. It triggers a taxable event the moment the sale settles. It prints sell pressure into the order book exactly when demand for the wrapper is rising. It adds bank compliance review and cross-border wires to a process that should be measured in blocks, not business days.
In-kind creation removes the sequence. The holder delivers BTC directly to the trust or an authorized participant and receives shares. Four things change at once, and they are not equal in weight.
The first is tax. For a long-term holder sitting on large unrealized gains — a miner, a digital-asset treasury company, a family office, an early fund — selling to create is a realized capital gain, taxed at a rate that can reach 20–37% depending on jurisdiction and holding period. Delivering the asset directly defers that event. The decisive variable in this announcement is tax friction, not trading cost. That single change is what converts a "maybe later" allocation into a "now" allocation for price-insensitive, sticky capital. Short-term traders are unaffected. This product is not built for them.
The second is market impact. Cash creation forges a mechanical link: ETF inflow must be accompanied by spot selling. In-kind breaks that link for the portion of flow that arrives in asset form. The third is friction — no fiat on-ramp, no bank review, no wire, a shorter lag between deciding to allocate and completing the allocation. The fourth is arbitrage efficiency. In-kind creation and redemption let authorized participants move between primary and secondary markets more cleanly, which in theory narrows premiums and discounts and tightens tracking error.
Then there is the part the announcement does not mention at all. At least five of the ten products carry staking: ETH, Avalanche, Solana, Sui, and the Hyperliquid vehicle. Staking is where the mechanism stops being clean.
Unbonding periods create liquidity mismatch. Ethereum's exit queue, Solana's roughly two-epoch cooldown, Avalanche's multi-week unbonding, Sui's own schedule — none of these are interchangeable. If in-kind redemption is permitted, the trust must hold a liquidity buffer through the unbonding window or face redemption delays, and delays produce discounts. There is also the asset-state question: does an investor deliver native unstaked tokens, or a staked position? The first forces the trust to delegate validators itself. The second transfers validator keys and slashing liability across the boundary. If a validator is slashed, the loss lands on shareholders who may not have known that exposure existed when they subscribed.
And delegation concentration is the quiet one. ETF staking routes to a handful of institutional validators. Each staking ETF becomes a new concentration point on the underlying network. Grayscale disclosed none of this. The mechanism complexity of the staking products was flattened into a single word in the product name.
The token-economics layer is asymmetric, and this is where most coverage goes wrong. On the supply side a wrapper is friendly: it locks float and adds marginal demand. On the governance side it is corrosive. Holding a staking ETF share is not holding a vote. The wrapper strips governance and keeps price exposure. For L1s that depend on active governance — Avalanche, Sui — that is a slow hollowing that surfaces over one to three years, not one quarter. Value capture gets financialized rather than protocolized: staking rewards are repackaged and resold, the issuer's fee is an external skim on the asset's cash flow, and none of it flows back to the protocol.
Note also the naming split. Grayscale labels some products "Staking ETF" and others "Trust ETF." That distinction implies real differences in regulatory treatment, yield distribution, and custody arrangement. No explanation was offered, and the omission is itself informative.
The competitive read matters more than the price read. Grayscale cannot win the BTC and ETH fee war — BlackRock's distribution and near-zero fees closed that door. So Grayscale is buying breadth and mechanism flexibility instead: ten products, altcoin long tail, first mover on in-kind. The low-fee Mini line paired with in-kind creation reads as a correction of the GBTC fee legacy, a new product line built to offset share bleed from the old one. This is a landscape event, not a price event. The truth is on-chain, not in the chat — and the chain here says the issuer is trading product count for ecosystem survival.
Now the pushback, because the consensus take is that in-kind creation is bullish for every asset on the list, and I think that is lazy. In-kind changes the source structure of demand; it does not touch the supply curve. For a high-inflation or high-unlock asset — Sui, parts of Solana's release schedule — new ETF demand can be fully absorbed by supply hitting the market. "ETF listed" is not "token economics improved." Treating the two as synonyms is the most common error I see in flow commentary.
There is a second blind spot: liquidity. Hyperliquid, Zcash, Sui products may launch with tiny AUM, wide spreads, and volatile premiums. Having a product is not the same as having liquidity. The institutional executability of that long tail is far lower than the product list implies, and the list is designed to look comprehensive.
The third blind spot is data. Once part of the flow arrives in asset form, "ETF net inflow" no longer maps one-to-one onto "spot net buying." Any model that uses ETF flow for short-term timing needs recalibration. That correlation weakening is the observation worth tracking, not the headline number. And there is a structural dependency nobody prices: custody. The entire ten-product stack rests on a small set of custodians and staking providers. One custody failure is not a product failure — it is a category failure.
The obvious next step is in-kind redemption, and I would expect it is already in regulatory conversation. Creation-only in-kind is a half-loop; the arbitrage closes only when both ends are physical. When redemption goes in-kind, watch two things — whether the staking products finally disclose unbonding and slashing handling, and whether ETF-flow models quietly break. That is where the next real information will surface, and it will not come from a press release.