Every risk model you’ve ever trusted has a secret dependency: it assumes the market eventually stops. Not forever — just long enough, once every 24 hours, for someone to draw a line and say this is where today ends. VaR is measured against that line. Margin resets against it. Swap charges are priced for the silence on the other side of it.
That line is disappearing, in three asset classes, in under a year — and almost nobody is talking about what it was actually holding up.
Nasdaq and the NYSE are pushing equities toward 22-23 hour sessions, SEC blessing already secured. CME is moving crypto futures and options to continuous trading in early 2026. Gold went first: Vantage’s XAUUSD247, STARTRADER’s own version on MT5, and CME’s round-the-clock 1-ounce futures contract — three products, landing within about seven weeks of each other. Everyone’s covering this as a story about access. More hours, more liquidity, happier clients. Fine. That part’s obvious and, frankly, a little boring.
Here’s the part that isn’t: a “day” was never just a block of time on a calendar. It’s a unit of measurement. And most of the machinery that keeps a brokerage solvent is calibrated in that unit — which means removing the close doesn’t just add hours to the trading session. It pulls the ruler out from under the risk desk mid-measurement.
What a “Day” Actually Does for a Living
Value-at-Risk — the number every risk committee lives and dies by — is conventionally a 1-day figure: how much could plausibly be lost between one close and the next. Regulatory capital charges scale that same daily number up to a 10-day window. Margin, in most setups, resets once, at end of day, against the closing price. Overnight swap is priced for exactly what its name says: one closed, quiet, tradeless overnight. Settlement is counted in days — T+1, T+2 — because a day is the smallest unit anyone bothered to build the plumbing around.
Take the close away and none of those numbers stop existing. They just stop meaning anything. What is a “1-day VaR” measuring, precisely, when there’s no longer a moment where one day stops and the next begins? Nobody’s put that on a slide deck yet. It’s worth asking before someone else’s live risk reporting answers it for them.
It’s not only the risk desk that leans on that daily reset, either. Regulatory reporting runs on a daily cycle. Trade surveillance and compliance reviews assume a defined window to check. Reconciliation between a broker’s own books and its liquidity providers’ records happens once the day is done, not mid-stream. Take the reset away, and every one of those functions loses the boundary it was built to check against.
Where It Actually Snaps
Gold shows the fracture cleanest, because the fix one provider reached for gives the problem away. Vantage’s XAUUSD247 replaced the traditional overnight swap with a funding rate, recalculated every four hours instead of charged once at a defined rollover moment. That’s not a UX tweak. That’s an engineering team quietly admitting “overnight” doesn’t exist anymore, and building around the fact instead of the fiction. STARTRADER’s own XAUUSD247, launched weeks later on MT5, hasn’t published an equivalent mechanism — which is its own small data point: even brokers moving fast on 24/7 access aren’t necessarily solving the financing question the same way, or as visibly.
Crypto’s move to continuous futures snaps the same joint one level up, in margin. Conventional futures margin assumes an end-of-day moment where gains and losses crystallize and accounts get topped up. Remove the end of day and margin either recalculates on a rolling cycle short enough to matter, or it’s quietly measuring risk against a boundary that isn’t there anymore. CME solves this at the exchange level. Every broker sitting on top of that exchange inherits the same unresolved question for its own book.
Equities have it worst, because so much sits downstream of “the close” — clearing, settlement, corporate actions — that stretching the session without rebuilding what’s underneath just relocates the bottleneck. DTCC’s NSCC went live with 24×5 clearing on 29 June 2026, months before the exchanges it clears for had caught up to the same hours. That sequencing wasn’t a nice-to-have. It was an admission that you can’t stretch the front of the pipe and leave the back of it running on banker’s hours.
Crypto Wasn’t Ahead. It Was Never in the Race.
The lazy version of this story says crypto simply got a head start on round-the-clock trading and everyone else is catching up. That’s not quite it, and the distinction matters. Perpetual futures — crypto’s dominant derivative — settle funding every few hours, not once a day, because there was never a daily close to anchor a once-a-day charge to in the first place. Spot crypto never had an opening bell. Nothing in its risk architecture was ever built assuming one would ring.
That’s a different starting line than gold or equities are running from. A broker bolting 24/7 access onto an existing gold or equities book isn’t extending a system that already speaks in rolling time. It’s translating a system built entirely in daily snapshots into a language it was never designed to speak — live, on a production book, usually under competitive pressure to ship before the translation is finished. Crypto didn’t do that work faster. It never had to do it at all.
The Instinct to Automate Everything Is the Wrong One
Once a desk accepts that nobody can watch a market that never sleeps, the obvious next move is to stop trying — hand the whole decision to an algorithm, let thresholds trigger action without a person in the loop, and call the problem solved. It’s the instinct almost everyone reaches for, and it’s the wrong one.
Crypto is the proof, not the counterexample. It’s the market that automated risk decisions most completely, earliest, and it’s also the market with the longest track record of liquidation cascades — one automated close triggering the price move that forces the next one, and the next, in a spiral that runs faster than any human could have interrupted it, precisely because no human was positioned to interrupt it. The mechanism worked exactly as designed. That was the problem. A threshold doesn’t know the difference between a position that deserves to be closed and a temporary air pocket in weekend liquidity that would have recovered in ten minutes if anyone had been in a position to pause and look.
The rolling risk math this shift demands is necessary. It is not sufficient, and it was never supposed to be. What continuous markets actually need is continuous visibility paired with a human who still makes the call — not continuous autonomy that removes the human from exactly the moments judgment matters most. The version of this transition that goes well isn’t the one with the smartest liquidation engine. It’s the one where a rolling risk figure updates in real time and a person is still the one deciding whether to act on it, widen it, or override it.
What Actually Has to Get Rebuilt
Not a night shift. Not a wider stop-loss, and not a smarter algorithm sitting alone in the loop either. The daily-anchored numbers — VaR, margin thresholds, exposure limits — need rolling equivalents that don’t assume a close is coming to draw the line for them. That part is a genuine engineering problem, and it’s harder than “monitor for more hours” because it changes what the number represents, not just how often someone glances at it.
But the number is only half the fix. The other half is making sure a person can actually see it — exposure, PnL, and hedge coverage pulled into one real-time view, instead of a snapshot that was only ever built to refresh once a day — and is still the one deciding what happens next. Not a threshold, closing positions alone at 3am because nobody wired a human into that hour. A rolling number, in front of someone who can weigh context a formula can’t.
Eighteen months out, round-the-clock access will be table stakes across gold, the major indices, the biggest crypto pairs, and a meaningful slice of US equities. By then the interesting split won’t be who offers continuous trading. It’ll be who rebuilt what “a day” means in their own risk math before it cost them something, and who’s still quietly measuring a market that never closes against a line that isn’t there anymore. If your desk is already asking that question, book a walkthrough of Brokerpilot.
Every risk model you’ve ever trusted has a secret dependency: it assumes the market eventually stops. Not forever — just long enough, once every 24 hours, for someone to draw a line and say this is where today ends. VaR is measured against that line. Margin resets against it. Swap charges are priced for the silence on the other side of it.
That line is disappearing, in three asset classes, in under a year — and almost nobody is talking about what it was actually holding up.
Nasdaq and the NYSE are pushing equities toward 22-23 hour sessions, SEC blessing already secured. CME is moving crypto futures and options to continuous trading in early 2026. Gold went first: Vantage’s XAUUSD247, STARTRADER’s own version on MT5, and CME’s round-the-clock 1-ounce futures contract — three products, landing within about seven weeks of each other. Everyone’s covering this as a story about access. More hours, more liquidity, happier clients. Fine. That part’s obvious and, frankly, a little boring.
Here’s the part that isn’t: a “day” was never just a block of time on a calendar. It’s a unit of measurement. And most of the machinery that keeps a brokerage solvent is calibrated in that unit — which means removing the close doesn’t just add hours to the trading session. It pulls the ruler out from under the risk desk mid-measurement.
What a “Day” Actually Does for a Living
Value-at-Risk — the number every risk committee lives and dies by — is conventionally a 1-day figure: how much could plausibly be lost between one close and the next. Regulatory capital charges scale that same daily number up to a 10-day window. Margin, in most setups, resets once, at end of day, against the closing price. Overnight swap is priced for exactly what its name says: one closed, quiet, tradeless overnight. Settlement is counted in days — T+1, T+2 — because a day is the smallest unit anyone bothered to build the plumbing around.
Take the close away and none of those numbers stop existing. They just stop meaning anything. What is a “1-day VaR” measuring, precisely, when there’s no longer a moment where one day stops and the next begins? Nobody’s put that on a slide deck yet. It’s worth asking before someone else’s live risk reporting answers it for them.
It’s not only the risk desk that leans on that daily reset, either. Regulatory reporting runs on a daily cycle. Trade surveillance and compliance reviews assume a defined window to check. Reconciliation between a broker’s own books and its liquidity providers’ records happens once the day is done, not mid-stream. Take the reset away, and every one of those functions loses the boundary it was built to check against.
Where It Actually Snaps
Gold shows the fracture cleanest, because the fix one provider reached for gives the problem away. Vantage’s XAUUSD247 replaced the traditional overnight swap with a funding rate, recalculated every four hours instead of charged once at a defined rollover moment. That’s not a UX tweak. That’s an engineering team quietly admitting “overnight” doesn’t exist anymore, and building around the fact instead of the fiction. STARTRADER’s own XAUUSD247, launched weeks later on MT5, hasn’t published an equivalent mechanism — which is its own small data point: even brokers moving fast on 24/7 access aren’t necessarily solving the financing question the same way, or as visibly.
Crypto’s move to continuous futures snaps the same joint one level up, in margin. Conventional futures margin assumes an end-of-day moment where gains and losses crystallize and accounts get topped up. Remove the end of day and margin either recalculates on a rolling cycle short enough to matter, or it’s quietly measuring risk against a boundary that isn’t there anymore. CME solves this at the exchange level. Every broker sitting on top of that exchange inherits the same unresolved question for its own book.
Equities have it worst, because so much sits downstream of “the close” — clearing, settlement, corporate actions — that stretching the session without rebuilding what’s underneath just relocates the bottleneck. DTCC’s NSCC went live with 24×5 clearing on 29 June 2026, months before the exchanges it clears for had caught up to the same hours. That sequencing wasn’t a nice-to-have. It was an admission that you can’t stretch the front of the pipe and leave the back of it running on banker’s hours.
Crypto Wasn’t Ahead. It Was Never in the Race.
The lazy version of this story says crypto simply got a head start on round-the-clock trading and everyone else is catching up. That’s not quite it, and the distinction matters. Perpetual futures — crypto’s dominant derivative — settle funding every few hours, not once a day, because there was never a daily close to anchor a once-a-day charge to in the first place. Spot crypto never had an opening bell. Nothing in its risk architecture was ever built assuming one would ring.
That’s a different starting line than gold or equities are running from. A broker bolting 24/7 access onto an existing gold or equities book isn’t extending a system that already speaks in rolling time. It’s translating a system built entirely in daily snapshots into a language it was never designed to speak — live, on a production book, usually under competitive pressure to ship before the translation is finished. Crypto didn’t do that work faster. It never had to do it at all.
The Instinct to Automate Everything Is the Wrong One
Once a desk accepts that nobody can watch a market that never sleeps, the obvious next move is to stop trying — hand the whole decision to an algorithm, let thresholds trigger action without a person in the loop, and call the problem solved. It’s the instinct almost everyone reaches for, and it’s the wrong one.
Crypto is the proof, not the counterexample. It’s the market that automated risk decisions most completely, earliest, and it’s also the market with the longest track record of liquidation cascades — one automated close triggering the price move that forces the next one, and the next, in a spiral that runs faster than any human could have interrupted it, precisely because no human was positioned to interrupt it. The mechanism worked exactly as designed. That was the problem. A threshold doesn’t know the difference between a position that deserves to be closed and a temporary air pocket in weekend liquidity that would have recovered in ten minutes if anyone had been in a position to pause and look.
The rolling risk math this shift demands is necessary. It is not sufficient, and it was never supposed to be. What continuous markets actually need is continuous visibility paired with a human who still makes the call — not continuous autonomy that removes the human from exactly the moments judgment matters most. The version of this transition that goes well isn’t the one with the smartest liquidation engine. It’s the one where a rolling risk figure updates in real time and a person is still the one deciding whether to act on it, widen it, or override it.
What Actually Has to Get Rebuilt
Not a night shift. Not a wider stop-loss, and not a smarter algorithm sitting alone in the loop either. The daily-anchored numbers — VaR, margin thresholds, exposure limits — need rolling equivalents that don’t assume a close is coming to draw the line for them. That part is a genuine engineering problem, and it’s harder than “monitor for more hours” because it changes what the number represents, not just how often someone glances at it.
But the number is only half the fix. The other half is making sure a person can actually see it — exposure, PnL, and hedge coverage pulled into one real-time view, instead of a snapshot that was only ever built to refresh once a day — and is still the one deciding what happens next. Not a threshold, closing positions alone at 3am because nobody wired a human into that hour. A rolling number, in front of someone who can weigh context a formula can’t.
Eighteen months out, round-the-clock access will be table stakes across gold, the major indices, the biggest crypto pairs, and a meaningful slice of US equities. By then the interesting split won’t be who offers continuous trading. It’ll be who rebuilt what “a day” means in their own risk math before it cost them something, and who’s still quietly measuring a market that never closes against a line that isn’t there anymore. If your desk is already asking that question, book a walkthrough of Brokerpilot.
