Your current trading platform is probably letting you down

Liquid is one of the fastest growing trading platforms, allowing users to trade stocks, commodities, FX, and more 24/7/365 from their phone and computer.

Trading on Liquid is as simple as:

The best part is that Liquid markets never close. So no matter what is going on in the world, you are able to keep your portfolio positioned properly.

Start trading 24/7 today

Classical Investor — The Toll Booth Series, Part [X]

I clear my MES trades through Interactive Brokers. I have for years. The platform is ugly, the documentation is dense, the risk system will yell at you in five different ways before it lets you put on a position that any retail platform would happily front-run, and the margin rates are roughly half of what my last broker charged. The interface looks like it was designed by someone who has never met a designer and would punch one if they did. That, as it turns out, is the point.

This piece isn't about trading through IBKR. It's about owning the company that owns the rails.

Interactive Brokers printed Q1 2026 last Tuesday and the stock pulled back roughly 3% to ~$78. Revenue came in at $1.67B against a consensus of $1.69-1.74B (a slight miss), EPS at $0.60 was essentially in line, pretax margin held at 77% — the sixth consecutive quarter above 70% — and management raised the dividend to $0.35 a year. The post-print analyst revisions were the most interesting part: Jefferies cut $91 to $81, BMO cut $90 to $80, Piper raised $80 to $88, Barclays raised $83 to $85. The consensus PT shook out at roughly $78.30, which is to say the sell side is now telling you this thing is fairly valued at ~37x trailing earnings and there's no upside from here.

This is the moment I get interested in a name. Not because the print was great — it wasn't, the revenue line missed and the operating-metric growth decelerated modestly from March — but because a slight miss into a "consensus PT roughly equals current price" setup, on a name where the operating model is structurally misclassified, is exactly when the methodology I run finds its asymmetry. Every write-up I've read on this name opens with some version of "it looks expensive, but…" and then proceeds to apologize for the multiple. I'm going to do the opposite. The 37x multiple is not a bug. It is the market admitting, half-grudgingly, that this is not a broker. The thesis I want to test is whether the next twelve to twenty-four months of compounding deliver enough that the multiple holds up while EPS does the heavy lifting.

I'm going to walk through this the way I walked through CME — segment by segment, line by line, asking what drives the number, whether it compounds, and where the sell side is structurally short of imagination. Then I'll do something I deliberately didn't do on CME: put real comp multiples and a forward-EPS scenario tree on the page, so you can see exactly where the upside lives and what has to be true for the trade to work.

You can't model IBKR without understanding what it actually is, and you can't understand what it is without understanding Thomas Peterffy.

Peterffy showed up in New York in 1965 from communist Hungary, with effectively no English. He taught himself to code, did engineering work for a while, and ended up trading options on the American Stock Exchange floor in the late 70s. The story most people know is that he wired a primitive handheld computer into the trading floor — literally, with rules against electronics, he hacked his way around them — and used it to compute Black-Scholes prices in real time while the rest of the pit was doing it in their heads. He took the AMEX market makers' lunch money for years before regulators figured out what he was doing.

The relevant point is not the story. It's the founding premise. Peterffy didn't build a brokerage that happened to use computers. He built a market-making operation called Timber Hill that happened to grow a brokerage on the side because the same systems that priced and risk-managed Timber Hill's inventory could clear and risk-manage outside customers' orders for a fraction of what existing brokers charged. The brokerage was a downstream byproduct of an automation engine that was already running.

That sequence — automation first, brokerage second — is the entire story. It is why IBKR's pretax margin is 77% in a year where Schwab's was barely a third of that. It is why every retail-friendly broker that has tried to copy IBKR's pricing — and several have tried — has either lost money on the seat, charged for it through payment for order flow, or quietly raised prices a year later. They are running brokerage operations and bolting technology on. IBKR is running a technology operation and processing brokerage orders through it. The numbers won't ever look the same and the multiple shouldn't either.

Peterffy is 81 years old at the time of writing. He's still chairman. He still owns the controlling stake. We'll come back to that.

IBKR offers direct, automated, electronic access to 170 markets across 40 countries and 29 currencies, on a single account, with cross-product margining across stocks, options, futures, forex, fixed income, mutual funds, ETFs, US spot gold, and — recently — crypto and prediction-market contracts.

Read that sentence twice. There is no other broker on earth that does this on one account. Not Schwab. Not Robinhood. Not Fidelity. Not E*Trade. Not Saxo, which gets the closest internationally and has a fraction of IBKR's product breadth. Not tastytrade, which is excellent at options and has none of the global infrastructure. The unified-account, unified-margin, unified-funding-currency design is itself the moat — and it is a moat that took 47 years and a few hundred million dollars of compliance, clearing, and connectivity work to build.

The customer base is the natural consequence of the product. Hedge funds, commodity trading advisors, family offices, RIAs, prop shops, introducing brokers running white-label retail apps on top of IBKR's back end, and an active retail tier that skews older, wealthier, more international, and more sophisticated than the typical Robinhood account. About 50% of accounts are now international. That number was a third a decade ago. It is going to keep climbing for reasons we'll get to.

Revenue splits roughly along three lines:

Net interest income — the largest bucket. About $3.9B of full-year 2025 revenue. The mechanic: customers park cash, IBKR earns a spread by investing that cash in short-dated instruments while paying customers a competitive rate. Plus margin loans, which IBKR funds at near-cost and lends to customers at a tiered rate that scales down to roughly SOFR + 50 bps for large balances. Plus securities lending, where IBKR lends out fully-paid customer shares to short-sellers and splits the rebate with the customer. All three are spread businesses. All three scale linearly with client equity and margin balances.

Commissions — about $1.9B of 2025 revenue. Per-share or per-contract pricing on IBKR Pro (which is what serious users are on); a zero-commission Lite tier that does take payment for order flow, but is a small minority of revenue and is mostly a funnel for accounts that eventually convert to Pro. We'll come back to the PFOF distinction because it matters more than people think.

Other — market data, exchange fees passed through, currency conversion, the new ForecastEx prediction-market business, occasional Timber Hill-related items.

Hold that mix in your head: 60-65% net interest income, 30% commissions, the rest other. That ratio is structurally different from a Schwab or a Robinhood and the difference matters.

Most write-ups describe the moat as "low cost, automated platform." That's true. It's also the shallowest of the three layers, and stopping there is how analysts arrive at the lazy "it's a discount broker with a fancy interface" framing. There are three real layers and you only get to the right answer if you can see all three.

Layer one: cost. This is the obvious one. IBKR's per-employee revenue and per-account operating cost are not in the same neighborhood as any other broker on earth. Schwab has roughly 38,000 employees against IBKR's ~3,000, against client asset bases that are within an order of magnitude of each other. The automated infrastructure means accounts onboard, fund, trade, get margined, get reported, and get tax-documented with effectively zero marginal human cost. A new client costs IBKR almost nothing to add and produces multiple revenue streams from day one. Capex was about 1% of revenue in 2025. There is nothing in the operating model that requires meaningful headcount growth to grow the top line. From a controller's perspective, this is the rarest revenue line in financials: a B2C customer base that scales like a B2B SaaS gross margin profile. That is the cost layer. It is real, it is durable, it is what most analysts see, and it is one third of the answer.

Layer two: portability — or rather, the absence of it. This is the layer most equity research misses, and it is the actual reason IBKR's churn is what it is. A typical Schwab or Fidelity customer has a US-dollar account holding US securities. They can ACAT to a competitor in 48 hours and lose nothing. An IBKR customer often has a multi-currency account holding positions across European, Asian, and emerging-market exchanges, with FX hedges on, with margin loans denominated in one currency against collateral denominated in another, with options spreads that net for margin purposes only because all the legs sit at the same clearing entity. Try moving that account. You can't. There isn't a competitor to move it to. The closest analog is Saxo, and Saxo doesn't clear US listed options or US futures the way IBKR does. The sophisticated international user is locked in by the architecture of their own portfolio — not by any switching fee, but by the simple fact that no other firm can reproduce the structure on the other end. This is the same dynamic that makes a CME primary dealer reluctant to fragment clearing across two venues. Once you've consolidated complex risk inside one institution, the cost of un-consolidating it is enormous.

Layer three: the financing engine. This is the one nobody writes about and it is, to my eye, the actual answer to "why does this thing earn 77% pretax." IBKR self-clears. It owns the books. It runs its own securities lending desk. When a client posts $1M of cash, IBKR is the one investing that cash, not a third-party clearing firm taking the spread. When a client puts on a $5M margin position, IBKR is the one funding the loan at near-cost from its own balance sheet. When a client owns 10,000 shares of a hard-to-borrow biotech, IBKR is the one lending those shares to a short-seller and splitting the rebate. Every single one of those activities at every other retail broker either gets outsourced (and the spread leaves the building), or gets clipped at the clearing level (Schwab famously routes through DTCC and pays for the privilege at a level no automated firm would tolerate), or simply doesn't happen because the broker doesn't have the systems. IBKR captures the full economic spread on every dollar of customer cash, every dollar of margin loan, and every loanable share, while paying customers competitive rates on the cash side. That is why net interest income is 60% of the top line and the margin on it is something like 95% incremental. It is a financing engine wearing a brokerage costume.

Cost. Portability. Financing engine. Each on its own would be a meaningful moat. The combination is, to my reading, unassailable inside the active-trader and global-multi-currency niche. Schwab cannot replicate it without rebuilding their stack. Robinhood cannot replicate it without firing their PFOF revenue and acquiring 47 years of international connectivity. Fidelity cannot replicate it without getting comfortable losing money on a customer cohort their entire culture is wrong for.

A quick aside on payment for order flow, because most general-audience write-ups don't address it and it explains an enormous amount about who IBKR's customer is.

IBKR Pro — which is the version of the product the vast majority of revenue comes from — does not take payment for order flow on US equities. Orders route through IBKR's own SmartRouter, which is an algorithmic engine that splits orders across exchanges and dark pools to get the best execution price for the customer, and IBKR shares the price improvement with the customer rather than pocketing the rebate.

Robinhood, by contrast, gets paid by Citadel and Virtu to route order flow to them, and Citadel and Virtu profit by trading against that flow. The customer gets "free" trading; the dealer gets the spread. It is not, strictly speaking, theft. But it is a different economic model and it produces a different customer.

The relevant fact is this: the SEC and successive administrations have periodically considered banning PFOF. If they ever do, every broker built on the PFOF model takes a structural earnings hit. IBKR Pro takes none. In fact, IBKR's competitive position improves because the cost differential between IBKR Pro pricing and the "free" PFOF brokers compresses. This is a tail risk on the broker complex that is very specifically not a risk for Interactive Brokers. It is, if anything, a call option.

Full year 2025:

Q1 2026 (just printed):

March 2026 operating metrics (released first business day of April):

A few things worth pulling out. The Q1 revenue miss was driven by net interest income growing slower than the Street modeled — partly because customer cash kept piling up faster than IBKR could deploy it (uninvested cash up 35% is, perversely, a headwind to NII in a quarter where the spread on incremental cash is below the running average). Commissions were strong. Margins held. The dividend hike is the cleanest signal in the print — management does not raise the payout if they're worried about the next four quarters.

The March operating numbers are the more important read. Account growth at 31% with client equity at 38% means new accounts are showing up bigger or existing accounts are funding faster — both, probably. Margin loans up 35% means those accounts are actively borrowing, which is the highest-margin revenue line in the entire model. DARTs up 25% means commission growth is broad-based. Three growth metrics, all materially above the rate of S&P appreciation in the period, all pointing the same direction. This is not what a saturated business looks like.

The other thing I'd flag: IBKR releases these numbers monthly. Most brokers don't. That makes the company an enormously high-frequency catalyst structure for a methodology like ours — the first business day of every month, you get a fresh data point on the three things that matter most. Build a model around that and you have a real edge.

There's no point writing a deep dive that just describes the company. The question is why consensus is wrong, in which direction, and on what time frame.

One: international account growth is at an inflection. International accounts are now roughly 50% of the base, growing at a rate that comfortably exceeds the US side. The drivers are structural. International retail and prop traders are migrating to a single-account, single-platform model the same way US traders did fifteen years ago, and IBKR is the only firm with the regulatory footprint and currency infrastructure to be that platform globally. Every quarter that international account growth runs ahead of the US base, the IBKR revenue model improves on a per-account basis — international accounts trade more, hold more multi-currency cash, and use margin more aggressively than the average US retail account. Consensus is modeling international growth as a tailwind. It is the entire ballgame.

Two: the introducing broker / B2B2C channel is invisible to the model. A meaningful number of fintechs you have heard of run their back end on IBKR's white-label "Introducing Broker" infrastructure. They handle the front-end UX, the customer relationship, and the marketing. IBKR handles the clearing, the financing, the routing, the regulatory plumbing. Every dollar of customer cash and every margin loan at those introducing brokers earns IBKR the same financing spread it earns on its own retail accounts. This channel grows roughly with the universe of fintech build-outs globally, and the consensus model treats it as a noisy line in "other revenue" rather than the structurally compounding B2B annuity that it is.

Three: rate sensitivity is asymmetric. The bear case on IBKR is "rates come down, NII compresses, multiple compresses." That's not wrong, but it's incomplete. NII is not a single number. It is the spread between what IBKR earns on customer cash and what it pays customers, plus the spread between what it earns on margin loans and its funding cost, plus securities lending. The first spread does compress when fed funds falls, but margin loan growth (which is rate-insensitive — borrowers borrow because they want leverage, not because borrowing got cheaper) and securities lending (which depends on hard-to-borrow rates, not the policy rate) both keep growing. A 100 bp Fed cut is probably a 7-10% hit to net interest income holding balances flat, but balances aren't flat — they're up 35-40%. The math gets less scary the more carefully you do it. And on the way down, the multiple gets de-rated by people who skipped the second step of the math.

These three are independent. None of them require the others to work. That's the structure of a thesis I want to own — multiple independent legs of value creation, each with their own catalyst path, none requiring a heroic macro call.

This is the part of the story that is most often gotten wrong, including in the draft I'm rewriting from.

When you buy IBKR shares, you are buying a stake in the public holding company. The public holding company owns roughly 26% of IBG LLC, the operating partnership. Peterffy and a small set of related holders own the rest. The operating company earns the full $4.77B of pretax income; the public company gets its proportional slice.

The reason this matters is not that it makes the public co a bad investment. It's that the public co's percentage ownership of IBG LLC has been rising steadily over time. Every year, the public co repurchases units from the Peterffy family side, and every repurchase increases the public co's economic stake in the operating business. There is, in effect, a structural buyback baked into the corporate architecture. It doesn't show up in the standard share count math because it's happening at the partnership level rather than the public-co level, but the effect on per-share economics is the same. As long as Peterffy is gradually monetizing his stake into the public co's hands at reasonable prices, IBKR shareholders are getting a steady drip of operating-company ownership accretion that nobody is modeling.

On top of that, there is a small dividend (recently increased), and there is occasional share repurchase at the public co level. The capital return is not the headline — the structural unit redemption is.

The thing to understand from a controller's perspective is that the IBG LLC structure is not a governance compromise. It's a tax-efficient way for Peterffy to monetize over time without forcing a single dilutive event, and the byproduct is that the public co's economic interest in the operating business compounds quietly in the background. It's worth ten or fifteen basis points a year to the per-share owner economics, every year, on top of the operating growth. Add that to the operating business compounding at 25-30% on accounts and assets, plus a small dividend, and the total per-share value creation is materially higher than the headline EPS growth.

Five things actually keep me up on this name. None of them are "it's a 35x broker."

Rate cuts deeper or faster than my model. I outlined the math above. The point is not that there isn't a rate-sensitivity, it's that I think it's smaller than consensus models because consensus underweights margin loan and sec lending growth. If the Fed cuts 200 bps in twelve months and balances don't grow into it, my math is wrong. I watch the monthly margin loan number for this.

A real risk-off, where client equity contracts AND margin loans contract simultaneously. Equity drawdowns shrink the asset base; margin call cycles shrink the loan book. These two things historically correlate at the worst possible moment — March 2020, late 2018, 2008. IBKR is structurally short volatility in the sense that its earnings power dips when client portfolios are getting marked down and de-leveraged. This is the only macro scenario where the multiple gets hit and the earnings get hit at the same time.

Peterffy succession. He's 81. He's still chairman. He still controls the company. There is a CEO (Milan Galik) running operations and the operating story is not a one-man show. But the controlling-shareholder transition will eventually happen, and the market will price uncertainty around it whether or not the operating model changes at all. This is a "when, not if" risk and the best mitigation is the IBG LLC unit redemption schedule, which is already gradually moving ownership into public hands at predictable cadence.

Regulatory left-field. IBKR operates in 40 countries and is a registered broker-dealer, futures commission merchant, swaps dealer, and several other things in the US alone. There is always some new rule somewhere. The PFOF question is asymmetric to the upside; most other regulatory paths are asymmetric to neutral or mildly negative. Not the thesis killer, but a reason the multiple won't run too far.

Operational/cyber. A unified global trading platform is a beautiful product and a single point of failure. An outage during a market event would be expensive in customer trust terms even if the financial cost is small. IBKR has been remarkably clean on this dimension over its history. That is the kind of record that produces complacency in management teams. Watch it.

What I'm not worried about: Robinhood (different customer, different revenue model), Schwab (wrong cost structure, can't catch IBKR on price), or some new entrant building a competing global broker (the last firm that tried was Saxo, twenty years ago — there is no second-mover advantage in compliance-and-clearing across 40 countries). I'm worried about the multiple for one quarter if we get a soft print into a rate-cut surprise. I'm not worried about it for the next five years.

The 37x trailing P/E is the wrong frame. I'll get to the right frame in a minute, but first let's actually do the work most write-ups skip.

At a glance — 12-month price scenarios from $78:

Two-year base case: ~$109 (+40%) on 2028E EPS of ~$3.40 at 32x. Asymmetry roughly 3:1 to the upside on a one-year hold, better on two years.

Now the work behind those numbers.

Comp universe. IBKR's correct peer set is not Schwab and Robinhood. It's the asset-light global financial platforms — businesses where the marginal customer is high-margin, the capital base is small, and the customer relationship monetizes through multiple recurring revenue streams. That set looks something like this on trailing earnings: MSCI in the mid-30s, Moody's around 32x, S&P Global around 30x, FactSet around 28x, Tradeweb in the high-30s, CME in the mid-20s, ICE around 28x. Blend the platform comps and you get a fair-value multiple in the 30-33x range for a business growing operating income in the 8-12% area. IBKR is growing operating income meaningfully faster than that. The right multiple band, in my read, is 32-36x.

Forward EPS bridge. This is where the math actually has to work. 2025 adjusted EPS landed around $2.50. Q1 2026 printed $0.60, which annualizes to $2.40 if you assume zero growth — but Q1 is seasonally the weakest commission quarter and the cash balance build through Q1 will deploy at higher yields through the back half. My triangulation:

Consensus has 2026E around $2.55 and 2027E around $2.80-2.85. My base case is roughly in line with consensus on 2026 and modestly above on 2027 — the variant view is in the trajectory, not the next twelve months.

The bear case requires two things at once. A multiple compression to "broker" comps AND a flat EPS print. Either alone is fine — the model survives a multiple compression if EPS keeps growing, and survives a flat EPS year if the multiple holds. Both at once is the only way you actually lose money on a one-year hold from $78. The scenario for that is a hard recession that contracts client equity 20%+ and forces sustained Fed cuts that compress NII faster than balance growth can offset. Real risk. Not the median outcome.

The bull case doesn't require heroics. It requires the operating metrics to keep doing what they're doing — accounts growing in the 25-30% range, equity in the low-30s, margin loans in the high-20s — and the multiple to hold at platform comps rather than re-rating to brokers. That's an achievable setup if international growth doesn't roll over.

What the consensus PT misses. The Street's $78.30 sits at ~30x on 2026 EPS — platform-comp multiple applied to consensus EPS, which reads "fairly valued at platform multiples on platform earnings." That's correct if consensus EPS is right. My view is consensus is structurally low on 2027-2028 because it underweights the unit redemption (10-15 bps a year, free), the international migration (compounds the asset base above modeled), and the introducing-broker channel (invisible in the model). Three years out, that's a meaningful gap.

The trade isn't "buy because the multiple is wrong." The trade is buy because the forward EPS path is wrong, and the longer the operating metrics keep compounding above consensus, the more obvious the gap becomes.

You know how I do this. I'm not here for the consensus price target. I'm here for what the 30/60/90-day estimate revisions are telling me, and what the monthly operational releases are telling me between prints.

Post-Q1 (April 21):

The revision data is genuinely mixed and I'm not going to pretend otherwise. Of the four major desk PT moves I've seen, two cut and two raised:

Net: PTs came down on average, but the directional split (2 up, 2 down) and the fact that every desk maintained their positive rating tells you the cuts are about the EPS path getting marked down for slower NII growth into a Fed-cut cycle, not about the franchise. Revenue was a slight miss on NII; commissions were strong; margins held at 77%. That's a "we trimmed our model on rates, not on the business" kind of revision pattern. Read it that way and the revision data is more constructive than it first appears.

The high-frequency input is the monthly operating release. The first business day of May, we'll get April's account/equity/margin numbers. That is the data point that matters more than the headline Q1 print. If April shows account growth still in the high-20s, equity north of $800B, and margin loans above $86B, the Q1 revenue miss looks like noise and the bull case is on track. If April rolls over on any of those three — especially margin loans — the thesis needs to be reconsidered.

What would change my view: a monthly release that breaks the 25%+ trend on accounts, equity, or margin loans. Or a Q2 print that shows commission growth decelerating below 15% YoY. Or a Fed path that delivers more than 100 bps of cuts in the back half of 2026. Any one of those is a reason to size down materially.

Interactive Brokers is what happens when you let an automation engine run a brokerage for forty-nine years.

The market still wants to call it a broker. It is not a broker. It is a global financial platform with a financing engine attached, riding an international account migration that is in early innings, an introducing-broker B2B channel that compounds in the background, a structural unit redemption that adds to per-share economics every year, and a customer base that physically can't be replicated at any other firm because no other firm has the architecture.

The 77% pretax margin is the consequence, not the thesis. The thesis is that everything generating that margin is also generating it for a larger asset base every quarter, and that the people setting consensus EPS estimates are still anchoring to "broker" comps when they should be anchoring to asset-light global financial platform comps.

You don't get a home run on IBKR by being right that it's cheap on EPS — it isn't. You get a home run by being right that it compounds operating income for longer and at a higher rate than the consensus model has the imagination to underwrite, and that the multiple holds up because the more people who actually look at the operating metrics, the more it gets re-categorized.

The Q1 print and the post-print pullback to ~$78 is, by my read, the better entry. Bear $72 (-8%), base $94 (+21%), bull $110 (+41%) on a one-year hold. The revenue miss was a rates-and-cash-deployment story, not a franchise story. The dividend hike is the cleanest signal management gives. The monthly operating data is what we model around.

That's the trade.

Nothing in this piece is personalized investment advice. [I own / I do not own] IBKR at the time of writing. Classical Investor publishes general analysis based on publicly available information, including academic methodology and consensus estimate data via FMP. Do your own work.

— Chris