Spotware opened its new Spotware Talks series with a panel discussion “Gold gone wild: brokers’ survival guide – managing risk under extreme market conditions”, examining how brokers can stay afloat during volatility spikes without compromising execution quality. Hosted by David Kimberley, the panel brought together Angus Walker, Global Head of Trading at IC Markets and Drew Niv, Chief Strategy Officer at ATFX.
How many “dead men walking” are there in brokerage right now – and did January save some of them?
Drew notes that the brokers he knows best are the larger ones and, in his view, they come out fortunate. If conditions persist for another six months, he expects a far worse outcome, with the correction effectively “saving a lot of people” and underlining how serious the episode becomes.
David then draws on conversations from iFX Expo a couple of weeks earlier, where the mood suggests that even a few more weeks of the same conditions could wipe out some brokers. With no major closures surfacing publicly, he questions whether the danger is being overstated. Drew takes the opposite view – for smaller and mid-sized brokers, the runway can be as short as another month. He also highlights how trouble often shows up first through broken commercial commitments: IB payments get delayed, revenue-share deals start being renegotiated or quietly pulled back. Survival timelines, he adds, depending on the regulatory environment and how quickly the regulatory response comes. Stretch the conditions out by a few more months, and many participants do not make it through.
Angus then grounds the discussion in what the week looks like inside a trading operation. Wednesday, 28 January, is when his phone starts ringing, and the following day turns into a major down day for brokers. At IC Markets, exposure peaks on Thursday. He recalls calls coming in from brokers and industry contacts checking on each other, while rumours about who is in trouble start circulating. Keep the environment in place for another couple of weeks, and far more firms face severe stress. Retail participation is heaviest on Wednesday and Thursday, when IC Markets sees its highest exposure. In his view, their book acts as a useful proxy for broader market behaviour – and if brokers are forced to carry that kind of exposure for another consecutive week, the damage is substantial.
Was this a once-in-a-decade / SNB-style event – and how do you manage risk when flow is one-way and price keeps rising?
Angus points to a pattern behind the firms he hears are struggling: heavy concentration in Chinese business. Even with activity elsewhere, those brokers often lack meaningful offsets across the rest of the book, leaving them exposed to outsized metals losses with little compensation from other products. That, he suggests, matches the industry chatter. Diversification becomes the separator, and he positions IC Markets as an example of a broader mix by geography and product, with meaningful crypto and FX volume so it is not running a “super concentrated metals book”.
Drew widens the lens beyond any single region. Drawing on bridging-provider data from PrimeXM and oneZero, he argues that concentration is the norm across the industry. Strip out true multi-asset firms such as IG and CMC with large single-share business, plus outliers like eToro and crypto-heavy players, and most CFD brokers still sit at 70%+ concentration in gold. In his view, this is not simply China-linked – it shows up across brokers with primarily non-US client bases. Angus agrees that metals reach 80%–90% of turnover during the month, while still stressing that FX activity and the rolling of large FX positions provide meaningful balance.
David then tests whether turnover share is even the right way to read the risk. Angus pushes it aside as a blunt metric and shifts the focus to what a broker actually carries: the exposure that stays open and has to be rolled day after day. Clients can churn in and out intraday, generating high volume and strong transactional revenue even if positions close by the session’s end, and that revenue can cushion market-risk losses. The profile turns more acute when clients start holding and rolling positions, forcing the broker to carry exposure through the regime rather than simply process flow.
With imperfect tools, how should brokers actually manage incoming risk?
David shifts to what brokers should be watching, conceding that most firms do not have a perfect toolkit for managing incoming risk.
Drew starts with revenue composition. For 99% of the industry, he stresses, spread is the main income source, not client losses, yet many firms lose sight of that. Spread revenue comes largely from small clients trading frequently. B-book brokers, in his framing, tend to do one thing well: they build a large base of small, daily traders, which delivers a steadier revenue profile than open-position P&L. He points to ATFX’s base of around 150,000 small clients as a clear example.
He then draws a contrast with more concentrated businesses. Prime-of-primes and smaller shops, he suggests, face a tougher challenge because a larger share of results depends on open exposure P&L. In that setup, maximising spread revenue becomes less of a preference and more of a stabiliser, instead of leaning on directional outcomes.
From there, he moves to leverage and a failure pattern he sees repeating. Higher leverage can lift turnover for small accounts, yet the danger shows up when the same leverage is extended to large clients without tight monitoring. He challenges the practice of offering 400:1 or similar leverage to $1 million or $2 million accounts on the same basis as a $5,000 client. Small accounts can be managed under those conditions; large accounts can, as he puts it, “sink the ship”. SNB sits at the centre of his example set, with FXCM’s losses concentrated in that profile and the same pattern recurring in other disasters.
Drew also pushes back on the idea that broker near-death is rare. One-way positioning may be unusual, yet close calls are not. He points to January–August 2023, when gold trades range-bound. In that regime, brokers sit “on the ropes” and, in his estimate, can be four or five months away from half the industry being destroyed. The reason is structural: high gold concentration combined with a range market lets the best clients make money consistently, while broad retail losses are not large enough to cover fixed overheads. Market regimes change the business, he notes, and competitive pressure for deposits accelerates risk-taking. The commercial mistake, in his view, is treating a million-dollar client as if they should receive the same terms as everyone else.
From there, his sequence stays consistent: start with segmentation and leverage policy, then move to formal risk limits. In banking and wider financial services, limits are standard – in this industry, they often remain optional. Limits may reduce returns on peak days, he acknowledges, yet they protect against existential outcomes. He also points out the credibility gap: it is hard to talk about limits seriously while offering 500:1 to clients with $2 million accounts.
He underlines a market-structure issue that sits underneath metals pricing. Some brokers offer retail spreads in gold and silver that are better than what exists in the underlying market. In EUR, brokers can obtain bank pricing tighter than what they quote to retail; in gold, he suggests, that logic breaks. Gold’s “real market” liquidity is small relative to the synthetic volume created by B-book flow and if brokers had to hedge all of it, spreads would become extremely wide.
Angus adds that commercial pressure can overwhelm frameworks. Broker CEOs and owners talk, competitiveness turns into boasting about daily gains and that atmosphere rewards simplistic thinking – adopting a “risk framework” because it makes money on a given day, even if it loses on the way up and wins on the way down. He also notes the market drifting towards conditions like 1,000x leverage, which sits uneasily with traditional controls.
Do IB payout and pricing terms force brokers to rely on B-book client losses – and does that block proper risk management?
Drew outlines that many brokers measure the business incorrectly. He describes a common narrative where a CEO sees $100 million of client deposits, complains the broker “only made $80 million” and concludes a prime-of-prime “took” the missing $20 million – then decides the broker should never hedge. In his view, the correct approach is to calculate revenue composition properly: spread revenue, flow revenue and swaps. From analytics his firm runs for brokers, revenue is usually dominated by spread, swaps and other non-risk income. He adds that for many brokers, a $1 million risk limit per instrument would still preserve about 70%–75% of revenue in normal conditions.
He carves out two high-risk categories: brokers overly dependent on a small number of large clients and prime-of-primes that cater heavily to large clients. These firms, he says, lack stable transactional business and often rely on “leftovers” from brokers hedging what they consider higher-quality risk. They feel compelled to offer unrealistic terms to compete and those same terms create the risk they cannot escape, leaving them effectively hoping for repeated crisis-type windfalls.
Angus frames the solution as a more mature, bank-style model – tighter segmentation, clearer profit attribution and a more scientific approach, and Drew backs that direction, stressing that “blind risk” is never viable, while pointing out that effective controls depend not only on limits and segmentation but also on recognising how often risk is taken upfront, almost unconsciously. Brokers, in his view, offer their tightest spreads and most unrealistic terms in the instruments where clients lose the most money – especially gold and silver. Where clients can make more money, such as certain cross-currencies, spreads are often uncompetitive. He cites AUD/NZD as a range-bound cross where many brokers quote poorly. He stresses that diversification requires pulling business beyond the concentrated segment. Pointing to IG as a more diversified benchmark and eToro as a crypto-led outlier, he warns that an 80% gold business without limits and serious analytics leaves brokers operating on anecdotes rather than structured analysis.
Angus offers a straightforward trader’s-eye explanation of why concentration happens. Right before the webinar, he says, he opens a gold trade because gold is the only market moving. Bitcoin is boring. FX has some movement, but meaningful trading needs volatility, and gold is providing it. He says he would prefer to see more volume elsewhere, noting that gold is not their highest revenue-per-million product even if it is where activity clusters.
Gold, Drew suggests, pulls traders in on its own – it is not something the industry has had to push. He adds that at larger brokers, fewer than 10% of clients are EA users, yet they can generate 40%–50% of all transactions. EA clients are more spread-sensitive, which makes them easier to shift into other instruments. Manual traders will always gravitate to wherever the market is moving, he notes, while warning that many EAs are built around mean reversion and can struggle badly in sustained trends like gold. There is no single lever to pull – brokers need to take four or five concrete steps, or the damage will be greater the longer these moves persist.
After a correction, how do dealing/risk teams convince senior management to take risk limits seriously (and segment leverage properly)?
Angus frames it as proactive work inside a mature framework. At IC Markets, leverage on silver is cut early, as it is at some other brokers. He also stresses that the biggest losses he hears about are in silver rather than gold, with “horror stories” of nine-figure outcomes at some firms – nothing on that scale at IC Markets. By month-end there are plenty of profitable clients, and plenty of losing clients over the final days. The takeaway is simple: strong teams read the trend as it forms, rather than reacting once it is already too late.
He points to the broader backdrop: a very strong September, an average October, then a fresh wave of firms caught out in November, December and January as the regime shifts. Adaptation, in his view, is far from guaranteed. Even at IC Markets, the warning signs are there – platinum and palladium suddenly start moving, creating short-term concentration risk. Without the right team and framework, it is easy to get “run over”, leaving a broker scrambling to hedge platinum, palladium, silver or gold at the worst possible moment, just as a client is running towards nine-figure profit.
Drew agrees with the “pre-crisis” mindset and says hedging during a crisis is when brokers tend to pay the most. With many CFD instruments benchmarked to futures venues such as CME and Eurex, he urges desks to keep those markets in view. He points to CME’s SPAN margin as a dynamic system that shifts requirements with volatility, using 30, 60 and 90-day measures, and suggests brokers should mirror that discipline by tracking margins and spreads rather than realising too late that conditions have moved. What surprises him is how many desks still operate without a futures screen, relying instead on synthetic prime-of-prime feeds. That blind spot, he suggests, sits behind a wider issue: too little analytics against the true underlying markets. The answer, in his framing, starts before stress hits – teams agreeing in advance how to spot regime change and when to de-risk, instead of improvising under pressure.
Will this trigger an industry reset in risk management – or will brokers revert to old habits?
Drew leans towards the second outcome – a return to old habits. Many owners, in his view, will simply “sweat it out” and hope conditions settle. Incentives shape behaviour: in brokerage, the biggest money is typically made by running the operation, while in technology, founders often make their largest gains by selling businesses at higher valuation multiples. He points to FastMatch as a case in point – sold to Euronext for $180 million while earnings are still developing, with around $16 million in revenue and roughly $5 million EBITDA at the time. He also references a failed attempt to sell FastMatch to another exchange for $240 million that falls apart over management compensation. For Drew, those numbers illustrate what a clean, transactional technology business can be worth.
In Drew’s reading, the brokerage industry simply does not see enough deals to nudge owners towards building stable, saleable earnings. Volatility puts buyers off, fewer buyers means less reason to reshape the business, and many M&A conversations drift into distressed territory instead of staying strategic. On top of that, the buyer pool is small and, as he puts it, not geared to evaluating these businesses in the way technology buyers typically are.
He also flags competitive pressure coming from crypto. Over the past two years, he makes the point that the largest crypto firms have effectively unlimited “ammo”, driven by their valuations and access to public-market capital. The top five or six, in his view, can raise financing at a scale no FX broker can match because of valuation constraints, while higher crypto fees support very different unit economics. His bottom line is that a pure A-book model is not commercially viable for the industry, but a workable balance exists – and the firms that strike it are better positioned, even if many never make the shift.
What happened with CME silver margins – and what’s the risk gap between tight CFD metals pricing and real on-exchange liquidity?
At the time, Angus notes, IC Markets can be up to 20 times tighter than futures, quoting gold at 5¢ to 12¢ while futures spreads widen materially over recent months. That effectively puts them at a serious discount to the futures market. For him, this is where the real danger sits: not in the headline conversation about exposure and hedging, but in the retail prices brokers choose to show. Retail traders, he suggests, are broadly the same audience across firms, looking to monetise gaps between retail spot pricing and the primary market, including futures. The risk is mispricing. Closing that gap through hedging often means paying a premium depending on the venue, which makes pricing quality and controls decisive. Without strong liquidity, the broker can end up trading at a guaranteed loss. He also points out how wide the market can get under stress. During volatility, some brokers widen to 50¢ or a dollar, reflecting primary-market conditions, while tier-one venues are wider still – he cites spreads around a dollar for 100 ounces. Against that backdrop, being able to quote inside 20¢ and still hedge is a meaningful advantage. The takeaway he draws is simple: brokers can lose far more, far faster, by showing the wrong price than by rolling large positions.
Drew draws a clear distinction between brokers with broad, small-client flow and those exposed to size. Where a broker has many small accounts, two-way activity can be internalised more safely because natural offsets appear. The danger starts when brokers treat futures liquidity as deep enough to absorb large tickets. Top-of-book spreads may look tight, yet the depth behind them is thin and hedging costs climb quickly with size. He points to a familiar mistake: allowing a $100,000-deposit client to trade gold with very large leverage and trade sizes, then discovering that hedging is not a 10¢ exercise but can cost $1–$2 even in normal conditions. Liquidity, in his view, needs to be tiered by size. Tight pricing can be offered, but only for small sizes to small clients. Letting large clients trade 5,000 or 10,000 ounces at top-of-book spreads is reckless because hedging can cost multiple dollars, leaving the broker to “lose their shirt”. Based on his analytics, he puts roughly 80% of broker losses down to these pricing and size mismatches, with the biggest hits typically tied to large clients rather than the mass retail base. He also notes that around 30 January some large clients lose substantial money and the market narrative becomes “the whale lost”, yet those same large accounts are also the ones that can steadily bleed a broker when conditions suit them.
Drew also warns that client sophistication will keep rising as more traders reference futures while trading spot. Brokers can become victims of their own success: as a balance sheet grows, clients feel comfortable depositing $30,000 or $40,000, and the more capable cohort starts arbitraging brokers against futures pricing. AI accelerates that shift – not by making undisciplined traders disciplined, but by cutting the cost of technical capability for the disciplined minority. In his view, the 1% of knowledgeable, disciplined clients gain low-cost tooling through AI and can do real damage if brokers continue quoting unrealistic prices, which only raises the stakes for proper segmentation. He also links some of the stress seen in November, December and January to parts of the industry failing to update margins in line with SPAN changes at CME.
What role do LPs play in managing “price risk” – and how many are truly set up to handle it?
For Angus, the problem is not simply “price risk” – it is the cost of being even slightly off-market, which can run to several dollars per million. One of his sharper observations is that brokers can take more damage in consolidating markets than in trending ones. The gold rally looks like it could have been a train wreck, yet a persistent low-volatility regime may be worse. He shares numbers suggesting it is nearly a record month for IC Markets – about 50% above their previous record, with unusually high losers as well. Range markets create a slow bleed: fewer big winners and fewer big losers, spreads steadily draining P&L, rebates compounding the drag, then mean reversion wiping out apparent gains. The danger is that this can persist, and mispricing remains expensive in that environment too.
Drew takes the same line, describing range markets as the industry’s worst enemy. He points to 2005–2007 as a stretch when FX is effectively “dead”, before a regime shift in August 2007 sets up an exceptionally strong period for brokers through 2008–2011. The hard part, he notes, is that nobody can time those shifts with confidence – markets can change quickly and catch firms off guard.
From Drew’s perspective, LPs can have a tougher job than brokers because many do not benefit from diversified retail flow. When brokers that usually keep risk internally suddenly hedge aggressively in one-sided markets, that risk can land on the LP’s book in a way that is disproportionate to the P&L the LP earns, eventually making the business less attractive.
Angus looks at it from a different angle. In some respects, the job becomes easier because client expectations shift. Traders once complain about minor slippage around events like NFP, particularly with tick-scalping strategies. Now many tolerate far more slippage because the priority is getting filled. On gold, 50¢ can go through without triggering complaints, either because clients recognise their own flow is poor or because strained liquidity becomes part of the deal. That shift gives LPs more room to manage toxic flow: quote tight spreads, apply aggressive slippage where necessary, then hold or delay softer flow – something he implies is common when upstream controls fail to filter poor flow.
Post-SNB: Did broker risk thinking actually change?
Drew admits that it does – SNB pushes many firms away from over-hedging and away from pure A-book models – yet the industry still has a habit of fighting the last crisis. In the aftermath, many conclude hedging and limits are “bad” because of credit-risk mismatches: clients on extreme leverage can lose more than their deposits, leaving brokers liable to counterparties who may pursue recovery. That leaves a familiar dilemma. Hedge too much and you reopen SNB-style risk; hedge too little and you walk straight into the other failures the panel has been describing. No model eliminates risk, Drew suggests – it simply moves it between market risk, credit risk and other forms. The goal is measured risk in several places rather than an extreme stance in one. Pulling that off requires difficult conversations across dealing, ownership and sales because product design and commercial competitiveness set the boundaries of what is realistic.
In a range-bound “slow bleed,” can you spot “picking up pennies” behavior before it blows up?
David pushes back on the idea that brokers sit around hoping clients lose. He leans on the revenue mix Drew has already laid out and backs it up with an example from XTB’s head of dealing: roughly 60% of revenue comes from spreads, 20% from swaps and 20% from market making. Under that structure, the usual caricature of B-book economics falls apart – the engine is spread revenue far more than clients “being wrong”, and relying mainly on client losses makes for a punishing business model.
Drew picks up the same theme and suggests the industry often describes the numbers in the wrong way. Brokers see “client losses” and treat them as the firm’s edge, when a large share of that loss is simply the cost of spread. That misunderstanding, in his view, keeps reviving schemes like reverse-hedging a broker’s “worst clients” to manufacture profit. Those clients are usually “worst” for a simpler reason: they overtrade with leverage and bleed to spread, rather than being consistently wrong on direction.
He sharpens the warning on range markets: the real threat is not the mass retail base but the disciplined minority. He points to an analytics model that sorts clients by knowledge level. Knowledge level zero – the typical retail client – accounts for about 89% to 91% of clients at most large brokers. Levels one, two and three sit above that, with level three at around 1%. In range-bound conditions, levels one to three form the cohort that tends to make money, while the 90% neither loses dramatically nor generates enough directional loss to offset the drag of spread. Those more disciplined traders, he stresses, “murder” brokers in these regimes.
In practical terms, he estimates that in a range-bound environment the top 9% of clients – and especially the top 5% – can extract enormous value from a broker. If that environment persists for too long, it can bankrupt the firm. He points to 2023, when range-bound conditions last roughly seven to eight months and the market then sees a spike in “contract tear-ups”, with revenue-share arrangements suddenly declared “null and void”. These regimes typically run for three to four months, he notes, but once they stretch to a year or two, the industry lands in serious trouble because that small group steadily chews through a broker. Many of those traders fail to stay profitable when markets turn chaotic, he adds, but what they can earn during quiet regimes is substantial – and heavy leverage for large clients only compounds the risk.
Angus offers a real-time illustration of that rising sophistication. He describes mean-reversion traders getting run over in September – a huge month for many brokers – then recovering in October. The bigger surprise, for him, comes next: in November, some traders switch models, lean aggressively into trend behaviour and perform exceptionally well through the gold rally. The move from mean reversion to momentum is not perfect, he concedes, but it works far better than expected, underscoring how quickly sophistication is advancing.
Is a range-bound slow bleed effectively a tail-risk strategy – small losses for a long time, then a January-style payoff?
Drew frames the real issue as something other than waiting for the crash – it is the speed at which client sophistication improves. He illustrates it with two clients running the same EA: Client A sits at knowledge level zero, Client B at knowledge level three, with identical code but radically different risk outcomes because leverage and portfolio construction do the heavy lifting. The level-zero trader runs a mean-reverting algorithm on gold at extreme leverage. The level-three trader spreads the same idea across eight uncorrelated mean-reverting pairs – he cites sterling–swissie and aussie–kiwi – and treats it as a portfolio at modest leverage. Losses still happen, he notes, because no strategy always wins, but the odds of that portfolio blowing up sit far lower. Betting against that trader is not a sensible plan. By contrast, the highly levered mean-reversion gold trader can look brilliant for months because gold mean-reverts – until volatility breaks out, at which point “he’s done”. For Drew, those two traders represent fundamentally different risk profiles, even if the EA looks identical on paper.
He then widens the lens. As more clients blend strategies, bring futures and correlation data into their process and operate with more discipline, they start to resemble organised teams. He describes setups where multiple trading teams – some 600, others 400 – work on different approaches, while a separate top-level desk sits above them and dynamically de-/re-leverages exposure without the underlying teams even knowing: pulling capital, putting it back, changing portfolio weights based on how each strategy behaves. That top-layer risk allocation, he argues, is where the real skill sits.
As clients learn to operate in that way, the challenge for brokers grows. It remains a small fraction of the client base, but the combination of capital brokers allow those clients to deploy and the technical support they now get means they do not need to join a firm like Point72 to run a serious operation. Drew closes on the same warning: the industry has to face that shift – brokers do not have infinite balance sheets, nor infinite time.
Watch the full "Gold gone wild" panel discussion on Spotware's YouTube channel to find out how to secure stable operations when markets become increasingly unpredictable!
