Better Packaging, Same Plumbing

Tokenized equities, 23-hour sessions, T+1 settlement underneath.

Two approvals, one assumption

In the span of three weeks, the SEC approved two proposals from Nasdaq. On March 23, it approved tokenized securities trading for certain equity securities and ETFs under a pilot program (SEC File SR-NASDAQ-2025-072). On April 10, it approved 23-hour trading, extending Nasdaq's session from 4:00 AM to 8:00 PM Eastern, five days a week, with the NSCC filing in parallel to transition its clearing operations to a 24x5 schedule by June 28.

These are real milestones. Tokenized stocks on a major exchange. Near-continuous trading approved by regulators. The industry is moving, and the direction is clear: investors want always-on access to markets, and exchanges are building it.

Here is the part the headlines left out: tokenized trades still settle T+1. The token moves instantly on-chain. The actual security settles the next business day through the Depository Trust Company. Same batch cycles. Same cutoff times. The same central-counterparty model that has cleared US equity trades since the early 1970s, refined many times but not rearchitected.

Two settlement realities for one trade. The investor sees instant. The system needs 24 hours.

This is not a minor technical footnote. It is the central tension of modern market structure. The trading layer is being upgraded. The settlement layer has not moved.

The race that proves the demand

Every major exchange in the United States is now competing on trading hours. This matters. Not because every proposal is a solution, but because the race itself is evidence that the demand for always-on markets is real.

Cboe filed with the SEC on March 16 for near-24x5 trading on its EDGX exchange, targeting December 2026. NYSE Arca received SEC approval for 22-hour trading in February 2025 and is targeting launch by year-end. Nasdaq's 23-hour session was approved April 10. Robinhood, Charles Schwab, and Interactive Brokers already offer 24/5 trading to retail clients on 1,100+ stocks and ETFs. Volume in extended hours has grown 590% since February 2022 on Cboe alone. The Securities Information Processors submitted a plan amendment in May 2025 to extend consolidated data feeds to near-24-hour coverage.

Investors are voting with their orders. The actual experience, however, is not what the headlines suggest.

Overnight equity trading still accounts for a small fraction of total daily volume. Effective spreads run 3 to 6 times wider than during regular sessions, and price impact is roughly 6 times larger, according to the 2025 "Nocturnal Trading" study by Eaton, Shkilko, and Werner. The picture is mixed: for the handful of names with active overnight market-making, realized spreads narrow once a small set of competitors quote them aggressively. For the long tail of less-liquid names the wider quoted spreads stand and execution drag is real. Let me put the cost in concrete terms: for a retail investor whose extended-hours flow lands on those less-liquid names, the higher transaction costs add up to an estimated 0.3 to 0.5 percentage points of annual return. That is roughly one to two months of market gains, quietly lost to the plumbing. The average expense ratio of an actively managed equity fund is around 0.40%. Many retail investors left actively managed funds for self-directed trading specifically to escape that 0.40% expense ratio, and regular extended-hours trading can recreate exactly that drag through wider spreads and worse execution.

The regulatory protection gap explains part of this. Regulation NMS Rule 611, the Order Protection Rule, requires trading centers to route orders to the best available price across venues. It applies during regular hours only. Rule 605, which governs execution quality reporting, is tied to the same window. After 4 PM, these protections disappear. The SEC has published no guidance on extending them. Robinhood and Schwab restrict overnight trading to limit orders only. No market orders, no stop-losses. Platform capacity has not kept pace either. On August 5, 2024, during the global sell-off that followed the Bank of Japan rate hike, Robinhood's overnight execution partner Blue Ocean ATS hit its database message ceiling and went down, and the orders riding it were canceled mid-execution. The cause was scaling, not settlement, and that is the point. The new overnight venues are still proving they can handle the busiest nights.

None of this means accounts are at risk. Broker capital buffers and SIPC coverage protect against catastrophic failure. No systemic retail losses have occurred from settlement gaps under normal conditions. The problem is more subtle. The always-on surface works. Underneath, the experience is measurably worse: wider spreads, fewer protections, thinner liquidity, platforms that fail during stress. The faucet looks modern. The pipes haven't changed.

Why settlement cannot keep up

Trading can run 24 hours a day. Settlement cannot. Not because of a lack of ambition. Because of architecture.

In the United States, the NSCC processes trades in discrete cycles with hard cutoffs. Same-day settlement closes at 1:30 PM Eastern. The night cycle runs at 9:00 PM. Even the planned transition to "24x5 clearing" by June 28 extends the batch windows. It does not eliminate them. A trade placed at 2 AM Saturday enters the same queue as a trade placed at 10 AM Monday. The "always-on" trade waits for systems that operate on business hours. Fedwire, the Federal Reserve's payment system, operates 22 hours per day, Monday through Friday. No weekends. No holidays. The Fed announced in October 2025 it will expand to include Sundays and weekday holidays, but implementation is targeted for 2028 or 2029.

In Europe, the pattern is the same. The ECB published a consultation paper in June 2025 on extending T2 operating hours. No implementation timeline has been announced. Europe also faces its own complexity: 27+ member states, multiple central securities depositories, different cutoff times. The EU, UK, and Switzerland are coordinating a move from T+2 to T+1 settlement by October 2027, but even T+1 is not what it sounds like. Industry estimates put the actual working window at 2 to 4 hours. After accounting for time zone differences, FX settlement cutoffs, cross-border coordination, and the holiday calendars of every involved jurisdiction, the effective processing time compresses to a window in which any settlement break must be caught, reconciled, and resolved before cutoff.

The constraint is global. CLS, despite being called "Continuous Linked Settlement," operates in a single 5-hour daily window because it depends on all participating central bank systems being open simultaneously, which means the FX leg of any cross-border tokenized equity trade still has to wait for the same real-time gross settlement choke point. You cannot settle a trade at 3 AM if the capital to fund it is locked in a system that opens at 7 AM. True for Fedwire. True for TARGET2. True for every RTGS platform in operation today.

Two more architectural barriers apply everywhere. Today's infrastructure batches reconciliation, reporting, compliance checks, and software patches during off-hours. 24/7 operation requires a shift from batch to streaming architecture, with zero downtime tolerance. And the regulatory frameworks on both sides of the Atlantic assume a closing bell. In the US, the Order Protection Rule and execution quality reporting do not apply outside regular hours. The SEC has flagged extended-hours trading as a 2026 examination priority but has published no guidance on how existing rules should apply. In Europe, settlement discipline under CSDR is designed around business-day cycles. Neither framework contemplates continuous operation.

The best Blu-ray yet

The instinct behind tokenization is sound: if securities exist as blockchain tokens, they should be able to settle instantly, peer-to-peer, around the clock.

This is not new. The industry has been trying to make tokenized stocks work for years. INX launched the first SEC-registered tokenized IPO in 2021, raising roughly $85 million from over 7,000 investors. Today the token trades at $0.09, down 90% from its offering price, with daily volume under $50,000. The platform works. The liquidity never came. FTX offered tokenized versions of popular stocks through a German partner firm. Its tokenized GME at times traded more than $100 above the actual stock price because the market was too thin and too disconnected for anyone to arbitrage the gap. Binance launched tokenized stocks in April 2021 and shut the offering down three months later under regulatory pressure. The EU's DLT Pilot Regime has authorized exactly 3 platforms in over two years of operation, with total issuance under 513 million euros. ESMA's own evaluation in June 2025 described adoption as early-stage and limited. Always the same wall.

The momentum is real too. Tokenized US Treasuries crossed $10 billion in January. BlackRock, Franklin Templeton, and JPMorgan are all running tokenized funds with billions in assets. The capital is moving. AND each attempt, successful or not, has hit the same architectural constraint: settlement.

Each of these efforts proved two things simultaneously: the demand for tokenized securities is real, and the architecture resists.

Nasdaq's approach is different. It is the most sophisticated implementation yet. Tokenized shares trade on the same order book as traditional shares. Same CUSIP number. Same execution rules. Same market data feeds. No separate venue, no price divergence risk within Nasdaq. TD Securities analyzed this approach as one that works within existing market structure rather than creating parallel infrastructure, reducing the risk of price divergence and regulatory fragmentation while leaving the underlying clearing and settlement plumbing exactly where it was.

AND: tokenization under this model becomes a delivery method. The token is a wrapper around the same settlement process. Tokenized securities on Nasdaq settle T+1 through DTC. The DTC designed it this way deliberately during its 3-year pilot. Tokenized entitlements are explicitly excluded from the delivery-versus-payment process. The security settles through the same infrastructure as every other stock. Better packaging. Not natively digital settlement.

The pattern is familiar from consumer technology. When the home video industry moved from DVD to Blu-ray, it delivered a real engineering achievement: sharper picture, more storage, a better disc by every technical measure. AND it was still a physical disc in a physical player. The fundamental delivery model had not changed. Streaming replaced Blu-ray not because it had better picture quality, but because it eliminated the disc entirely and rethought how content reaches the viewer. Tokenization is the capital markets equivalent: better packaging, running on the same settlement infrastructure that has processed trades for decades.

This does not mean tokenization is a dead end. Every infrastructure transition starts by running on the existing system. The internet ran on telephone lines before it replaced them. The question is whether today's tokenization is the beginning of that transition or a permanent layer on top of unchanged plumbing.

The parallel extends to fragmentation. In the US, Nasdaq is building on one blockchain while NYSE partnered with Securitize in March 2026 to build a separate tokenized platform, with a Q3 2026 pilot. In May 2026, the DTCC announced its own tokenization platform with BlackRock and Goldman Sachs as launch partners, with a July pilot and October launch targeted. Three major US settlement tracks heading for production in the same year, each tied to its own venue and infrastructure stack. Same problem, more vendors. TD Securities warned that Nasdaq's three-track approach "could split trading into two markets" with price divergence and regulatory arbitrage between them. Major market makers like Citadel, Virtu, and Jane Street have invested heavily in crypto infrastructure over the past three years, but have not yet committed to making markets directly in tokenized equities, and without their balance sheets the spreads stay wide and the venues stay shallow.

In Europe, the fragmentation is already visible. Börse Stuttgart's Seturion launched in September 2025, and Nasdaq connected its European trading venues to it in March 2026. Simultaneously, the ECB's Project Pontes is targeting Q3 2026 for a DLT settlement system anchored in central bank money. Two parallel settlement infrastructures, built by different institutions, with different settlement logic. DTCC, Euroclear, and Clearstream issued a joint warning in March 2026: "Interoperability is a prerequisite for digital asset security adoption at scale." It does not exist.

The capital flowing into tokenized assets is real: BlackRock's BUIDL holds $2.47 billion across seven blockchains, Franklin Templeton's BENJI stands at $1.86 billion, tokenized US Treasuries crossed $10 billion in January. AND it sits on different chains, different rails, different settlement models. The industry is not converging toward one settlement layer. It is building better packaging for multiple incompatible settlement systems.

What would actually need to change

The industry has proven it can extend trading hours, tokenize assets, and build new venues. The innovation on the trading layer is real and accelerating. The problem is underneath: central banks, clearing houses, and collateral management systems that still operate on fixed schedules.

In the United States, 24/7 settlement requires Fedwire to operate on weekends. That is two to three years away. No shortcut exists. The NSCC needs to move from batch cycles to continuous clearing. The June 28 transition extends the hours, not the architecture. Investor protections do not automatically follow trading hours. The Order Protection Rule does not apply to extended hours. No SEC guidance exists on bridging this gap.

In Europe, the Eurosystem's T2 platform needs to extend beyond business hours. The ECB is consulting, with no timeline. Central bank money needs to be available around the clock for settlement. The ECB's Piero Cipollone spoke in March about "building the rails for Europe's tokenised financial markets" and positioned central bank money as the required anchor, signalling that the ECB will not delegate the settlement leg of tokenized markets to dollar stablecoins or commercial issuers. The ECB's digital euro targets first issuance in 2029. Project Pontes settled 1.6 billion euros in 2024 trials with 64 participants. These are serious, well-funded efforts. AND "building" and "operational" are different words.

Across both jurisdictions, the infrastructure gaps are structural. Collateral that moves atomically, without batch windows, does not exist. The clearing infrastructure cannot mobilize collateral in under a minute. Not at noon. Not at midnight. Risk management that runs continuously, without overnight resets, does not exist either. Surveillance does not run on a schedule. Neither does interoperability across settlement rails being built in parallel by institutions not coordinating on standards. DTC's pilot, NYSE's Securitize platform, DTCC's tokenization platform, Seturion, Pontes: all building in parallel, none interoperable.

Some of these changes are already in motion. The NSCC is extending clearing hours. The Federal Reserve is adding weekend operations. These are real steps. AND continuous, interoperable, 24/7 settlement is not on a timeline that lands in 2028. Industry analysts at DTCC, EY, and SIFMA put end-to-end 24/7 settlement another 3-5 years past Fedwire's 2028-2029 weekend extension, with no consensus yet on corporate actions, atomic DvP, or collateral interoperability. The incremental work is happening. The full architecture change has not started.

The architecture question

You still cannot trade on a Sunday. Not because the technology is missing. The industry has more trading technology than it has ever had.

The settlement architecture was built for a world with a closing bell. That bell still rings. The industry keeps improving the packaging: tokenized assets, extended hours, new venues. All progress. The question this progress has not yet answered: who rebuilds the pipes?

References

• Eaton, Shkilko, and Werner (2025). "Nocturnal Trading." SSRN. ssrn.com/abstract=5181159
• SEC File SR-NASDAQ-2025-072. Order Approving Proposed Rule Change, March 18, 2026 (published Federal Register March 23). Federal Register
• SEC Approval of Nasdaq 23-Hour Trading Proposal, April 10, 2026. Arnold Porter Advisory
• DTCC (2026). "The Shift to 24x5 Trading." dtcc.com
• Federal Reserve (October 2025). Expansion of Fedwire Operating Days. federalreserve.gov
• SWIFT Institute / ISITC Europe. T+1 Settlement Working Window Analysis. swift.com
• ESMA (2025). DLT Pilot Regime: Report on Functioning and Review. esma.europa.eu
• TD Securities (2026). "Nasdaq Tokenization: From One Order Book to Two Systems." Reid Noch. tdsecurities.com
• Cipollone, P. (2026). "Building the Rails for Europe's Tokenised Financial Markets." ECB/BIS, March 24. bis.org
• DTCC, Euroclear, Clearstream (2026). Joint Statement on Interoperability. CoinDesk
• DTCC and EY (December 2025). "Insights on 24x5 Equity Trading." dtcc.com
• SIFMA (2026). "Extended Trading Hours: Industry Considerations." sifma.org