In under sixty days, Charles Schwab moved crypto from a closed pilot to open enrollment across every eligible retail account on its book. No new chain. No new token standard. No protocol upgrade. A switch flipped and five assets went live for the general client base: BTC, ETH, SOL, LINK, AVAX. In the same news cycle, Deutsche Bank is reportedly weeks away from institutional custody for BTC and ETH, with tokenized deposits, blockchain securities issuance and digital bonds queued behind it.
Two institutions, two clocks. One has shipped. One has announced. Most coverage is folding them into a single narrative โ "TradFi enters crypto, again." That framing erases the only variable that matters to anyone holding a position: delivery certainty.
Schwab's build is worth understanding at the architecture level, because the architecture is the news. The firm is not running its own matching engine. Client assets sit with Schwab Premier Bank as primary custodian; Paxos handles trade execution and secondary custody. Clients do not hold keys. They hold a brokerage line item.
That structure is a deliberate compromise. A bank charter supplies the trust layer and the client relationship. A crypto-native infrastructure provider supplies the operational layer โ liquidity, settlement, order routing. The bank avoids building a private-key management stack from scratch and avoids the associated capital treatment questions. What it does not avoid is liability ambiguity: if assets are lost or a settlement fails, which entity absorbs the loss depends on contract language that has not been made public.
The scale claim attached to this rollout is $12 trillion in client assets. The disclosure line that should be attached to it is smaller and harder to find: crypto held at Schwab is not covered by SIPC.
The pilot-to-full expansion in under two months is itself a data point. Product organizations do not remove eligibility gates that quickly unless retention and support load came in at or below internal thresholds. That inference is worth more than any quote in the release.
Deutsche Bank is playing a different game entirely. Its custody build is in-house. The pipeline includes tokenized deposits, issuance of securities on-chain and digital bonds โ a full-stack DLT program rather than a brokerage front-end. Gerald Podobnik, co-head of the corporate bank, framed it as complementary: an addition to traditional services, not necessarily a replacement for how business is done today. That sentence is the most honest disclosure in the entire story. It describes a parallel ledger, not a rebuild.
Start with the toll booth. Schwab did not build execution. It bought it. Paxos already runs the same function for PayPal, Robinhood and Interactive Brokers. Every additional multi-hundred-billion-dollar broker that outsources rather than builds strengthens the B2B network effect and weakens the negotiating position of the next one. This is the quietest structural fact in the announcement: the middleware layer of regulated crypto is consolidating faster than the retail layer it serves.
Second, the $12 trillion figure is an attention anchor, not a flow model. The single most common error in TradFi-entry coverage is linear extrapolation from client assets to buy orders. Schwab's book skews older, long-horizon, allocation-driven. That demographic does not rotate into high-volatility assets at the penetration rates implied by the headline. Every prior comparable โ Fidelity's crypto launch, Robinhood's initial listings โ showed early conversion in the low single digits at best. The number to watch is not AUM. It is the crypto revenue line the first time it appears in a quarterly filing.
Third, and this is the part the coverage underweights: the incremental signal is not BTC and ETH. It is SOL, LINK and AVAX. BTC and ETH already have a fully compliant retail on-ramp through spot ETFs. Marginal new access adds little. SOL, LINK and AVAX had thin compliant retail distribution through mainstream brokerages. That is genuine new channel capacity.
But do not misread what it means legally. A broker-dealer listing an asset is a commercial judgment backed by an internal legal opinion. It is not a regulatory ruling. SOL was named in SEC litigation against trading platforms in 2023; the characterization softened as cases progressed, but no statute and no final judgment resolved it. Schwab's compliance team is expressing confidence, not receiving certainty. Tail risk of a delisting or trading restriction in a changed enforcement climate is small but non-zero, and it is not priced anywhere.
Fourth: the SIPC gap. Crypto at Schwab sits outside securities investor protection. In a broker insolvency, that asset class does not receive the same treatment as the equity account next to it. Operationally this almost certainly means a separate account structure, separate statements, separate support flows โ a client-experience seam that generates complaints long before it generates losses. Nobody reads this line. It is the most consequential sentence in the disclosure.
On the institutional side, Deutsche Bank arrives behind BNY Mellon, State Street and Zodia Custody, all of which run operating custody businesses with existing client mandates. Being late is not fatal in custody โ trust compounds slowly and relationships move slower โ but the first-mover advantage in regulated crypto custody has already been allocated.
Fifth, Deutsche Bank is the more interesting program and the less certain one. Tokenized deposits touch commercial bank money going on-chain. That is a two-to-three-year structural story, not a two-to-three-week trade. It also carries constraints the announcement does not mention. German crypto custody requires a BaFin authorization under the Kryptoverwahrgeschรคft regime โ long application cycles, meaningful capital requirements. Tokenized securities issuance and settlement in the EU need a pathway through MiCA or the DLT Pilot Regime. Any balance-sheet exposure to crypto assets can trigger a 1,250% risk weight under the Basel Group 2b treatment. Pure custody usually avoids the exposure. The moment a bank provides liquidity or inventory, the capital math changes violently.
And the underlying chain is undisclosed. Public, consortium, or private โ the announcement does not say. That omission is not cosmetic. If the settlement layer is permissioned, the program should not be counted toward Web3 adoption metrics at all, no matter how many digital bonds get issued.
The execution asymmetry deserves its own line. Schwab went from limited pilot to full eligibility in under two months. Deutsche Bank has spent the last decade in successive strategic restructurings, and its crypto commitments use the vocabulary of intention โ preparing, planning, next year, future expansion into Asian financial centers. Multi-jurisdiction custody means multiple licenses: BaFin at home, then MAS, SFC and JFSA for the Asia leg. Each is a capital commitment with a multi-quarter timeline. Discount the announcement accordingly.
Here is the angle nobody is publishing. In this entire structure, the token holder is not the beneficiary. The token holder is the served object.
Value capture happens in three places: Paxos, taking execution and sub-custody fees; Schwab shareholders, taking commission and spread; Deutsche Bank shareholders, taking custody and tokenization fees. The assets themselves are inventory moving through a fee pipeline. That is not a bearish call on BTC. It is a correction to the reflex that maps "bank adopts crypto" onto "token goes up."
The second blind spot is fees. Not one number appears in the coverage. Fee structure is the decisive competitive variable in retail crypto โ the thing that determines whether Schwab's entry actually pressures Coinbase's take rate or just adds a slow-moving distribution channel. A low-spread strategy would be materially threatening. A padded one would not.
Third: Deutsche Bank's stated plan to extend into stablecoin trading. The wording is doing heavy lifting. Trading stablecoins as a brokerage service and issuing or reserve-custodying stablecoins are different businesses under different rules, with different competitive casualties. The first competes with exchanges. The second competes with the reserve banking relationships underneath Circle and Tether. The disclosure does not distinguish them, and the gap between those two readings is where most readers will form the wrong conclusion.
Watch two things, neither of which is a price chart. First, Schwab's fee schedule and the first crypto revenue disclosure in a 10-Q โ that is where the channel-versus-commodity question resolves. Second, whether Deutsche Bank names its settlement layer. A permissioned chain means the program is a bank efficiency project wearing crypto vocabulary, and should be valued accordingly.
Speed is the only currency that doesn't inflate. Right now, only one of these two banks has actually spent any.