The thinkorswim playbook, 20 years in the making — the same trade, earlier, in a market of 660 million people.
There is a film on YouTube. Official channel, @RiskMathicsFI. The title translates to a fact, not a slogan: twenty years leading financial training in Mexico and Latin America. No renders. No actors. Lecture halls full of risk officers. Trading floors. John Hull at the front of the room. The badges in the audience belong to the institutions that run a continent's markets — banks, afores, brokerages, exchanges, regulators.
The tape starts in 2005. Since then: 65,000 professionals trained — paid seats, not newsletter sign-ups — and 100,000 trained directly and indirectly once conference floors and corporate cohorts are counted. One position held for twenty years without interruption: the #1 derivatives, risk and trading institute in Latin America. The institute where John Hull teaches.
Under the film, a balance sheet. 2023 revenue of MXN 46.5M — about US$2.6M — at an 18.5% EBITDA margin. Zero financial debt in every year on record. Dividends paid in 2022 and 2023. That is not the pitch. That is the floor the pitch stands on.
Behind the film there is a machine most visitors never see. One operating core — RM·OS — orchestrating a 198-node technology stack across 13 infrastructure layers. Twenty-six business units run on one spine: courses, conferences, certifications, media, placement, publishing. Not twenty-six bets — one flywheel with twenty-six blades.
The content engine is sovereign. A 96GB GPU (RTX PRO 6000) runs nine open-source LLMs on our own hardware — no tokens, no vendor, no per-word tax. It has already produced 720+ production assets across 84 courses and cut content cost 80-95% against the way incumbents build it. Twenty years of proprietary curriculum sits behind it as retrieval: the model speaks with our syllabus, not the internet's.
Distribution is a graph, not a department: 73 platforms in three rings, hub-and-spoke, one input reformatted by AI into every channel, five bots running 24/7 on a single data spine from lead to placement. Twelve employees hold all of it — down from thirty; sovereign AI absorbed the difference — at US$2.6M of revenue and positive EBITDA. That is what operating leverage looks like before capital arrives.
Moat one is written into law. CONSAR-CHSF certification is legally mandatory for Mexico's pension system, and RiskMathics is the incumbent accredited certifier — roughly ten afore relationships, corporate contracts at US$75-80K, inbound demand at near-zero acquisition cost. We volunteer the risk before anyone finds it: one regulator is concentration, so the target is under 20% of group gross profit by 2027. An anchor, not a monopoly. And the certification franchise is widening: the GARP alliance makes RiskMathics the expansion arm of the world's largest association of risk professionals across the Spanish-speaking world, on revenue share — a global certification brand now rides our rails.
Moat two is the data under the AI. The GPU is a cost lever; the asset is twenty years of curriculum, exam banks and market case studies no competitor can scrape — plus the 73-platform distribution graph and the 65,000-name alumni network attached to it.
Moat three is the brand. The institute where John Hull teaches. Nobel laureates on faculty. A CBOE partnership carrying the John Hull World Cup. In a region flooded with motivational trading courses, RiskMathics is the serious one — the exact position tastytrade took against financial punditry. It compounds the same way: slowly, then all at once.
Twice, the same thesis, US$1.6B in exits. thinkorswim taught traders and became their platform — sold to TD Ameritrade for US$606M in 2009. tastytrade gave away eight hours of live math television a day to viewers in 190 countries, proved the audience wanted to trade, built tastyworks, and sold to IG for US$1B in 2021. Content builds community. Community becomes platform. Platform captures flow.
Then, on our call, you named the unfinished business yourself: "This was my issue with Tasty. I couldn't get them to open brokerage in Mexico." We repeat that with respect, because it is the most valuable sentence anyone has said about this market: the playbook worked twice, and the one border it never crossed is the one we have spent twenty years on the other side of. RiskMathics is that movie at the 2015 mark. The content exists — twenty years of it. The community exists — the desks of a continent, plus a retail book already at 30% of revenue and compounding. The religion is identical: volatility as opportunity, mathematics over opinion. Our brand mark is literally a volatility surface. What does not exist yet is the platform layer that turns audience into flow. That is not the flaw in the story. That is the entry price.
And here is where we are not tasty, said before anyone else says it: we have no broker and capture no flow — we charge fees per course, tasty earned on every trade. Our audience is institutional-first and Spanish-speaking, not mass retail in English. Our free content is not yet a daily eight-hour habit. And LATAM charges an FX and regulatory premium the US never billed you. So the ask is precise: capital, and the playbook for the one layer we have never built. You already ran this trade twice. This is the same trade, earlier — except this time, Mexico is the asset instead of the blocker.
Actuals first, on the Alvarez & Marsal diligence base. Revenue: MXN 24.7M in 2020, 39.7M in 2021 (+61%), 39.3M in 2022, 46.5M in 2023 — roughly US$2.6M. On the call I gave you memory numbers: about three million in sales, EBITDA in the mid-thirties. Here is the reconciliation before your team asks for it: the owner's 2023 EBITDA ran 33.4%; A&M normalized payroll, legal and marketing to market rates and printed 18.5%. Underwrite the 18.5%. And the debt you asked about: zero financial debt in twenty years — the US$1.5M is supplier payables, and it is retired at the first close. Dividends paid 2022-23. Volume, not vibes: 1,476 paid tickets in 2023 at a blended ARPU of US$1,440, plus 115 certifications issued.
Then the forward book — after we red-teamed it ourselves. The internal committee rejected management's first model and cut 2030 revenue by 44%. What survived is what we underwrite: US$6.9M in 2026, then 9.8, 13.8, 19.1, reaching US$24.7M in 2030 across seven education and certification verticals — about US$29.7M with media and merchandising on top. Every certification line is sized against a named census of certifiable professionals, not against hope. Management's US$77.8M case sits above it on every page — labeled as upside, never as the book.
And the gates are dated. A 2025 revenue acid test straight from tax-invoice records: the print must show at least US$4.5M, or the plan re-dates and we say so out loud. Full 2023-25 statements rebuilt in 45 days. A stage-weighted B2B pipeline, or the US$35.5M gross figure leaves the deck. We would rather show you a haircut than wear one later.
Four verbs, one graph. LEARN: 84 courses in derivatives, risk, quant finance and AI at US$3,200 average, 72% gross margin, 18× LTV-to-CAC. CERTIFY: eight certifications anchored by the mandatory CONSAR-CHSF and now the GARP rails; a corporate certification contract runs US$80K at 80% gross margin — 51× LTV-to-CAC, because the law does the marketing. ENGAGE: RM TV plus 73 distribution platforms converting free content into demand; memberships run at 47×. HIRE: Talent Hub places the professionals we certify at a 12% take-rate and 88% margin — 17×.
Each layer loads the next. Alumni become certification candidates. Certified professionals become subscribers, then placements. Employers who hire one graduate come back to train entire teams. The 65,000-name alumni graph is the flywheel's bearing — proprietary demand data nobody can rent.
Discipline is part of the design. Consumer feeder programs — kids, first-time investors, next-gen wealth — stay capped at about US$1M combined until they earn more. And quantum: you told us what you think of quantum education, and we heard you — it lives as a capped division and one profitable conference, not as the thesis. This round funds the three or four engines that already work, with board gates on everything else.
Exactly what I told you on the call: we start with five. The round is US$10M, priced — US$30M pre-money, US$40M post, 25% fully drawn — set deliberately below the US$45-50M post-money our own red team supports. No SAFE, no cap gymnastics: you asked direct questions, you get a direct price on RiskMathics Inc., the Delaware holding company, founder-owned 100% — bought back whole after selling half to the Mexican Stock Exchange years ago.
First close: US$5.0M at signing — operators, clean books, the RM TV pilot, and the balance sheet: the US$1.5M in supplier payables retired the day the wire lands, so zero financial debt becomes zero debt of any kind. Second close: US$5.0M at the same price when the evidence lands — the 2025 revenue print of at least US$4.5M third-party verified, the named certification census delivered, the CFO seated. If the gates don't clear, the second five never moves. The full US$10M works eight lines over 24 months: US$2.5M executive team — CFO is hire one, then COO and VP Enterprise Sales · US$1.5M supplier payables cleared · US$1.5M B2B go-to-market · US$1.2M platform and sovereign AI · US$1.0M certifications sized to named censuses only · US$0.8M RM TV daily desk, spend-capped by covenant and measured on attributed demand · US$1.0M working capital · US$0.5M governance and audit-grade books.
Board seat for the lead. Key-man insurance, founder vesting, IP assigned to the company, zero dividends during deployment — the covenant is explicit because the dividend history is real. This is priced off actuals and a haircut model, not off the dream. The dream is in the deck too — clearly labeled, and free.
Your question, verbatim, from the call — the same one that ended the last attempt at this market. Here is the answer, in three movements.
First, the base already exists. One hundred thousand professionals trained, directly and indirectly, across twenty years — the risk desks, the pension funds, the private banks, the regulators, and increasingly their clients. Born B2B, the book is already 30% retail and growing, with self-paid digital products in market. The audience nobody could acquire for a US platform has been assembling here, one paid seat at a time, since 2005. Second, certification becomes an arm, not a course: the GARP alliance makes RiskMathics the expansion vehicle of the world's largest risk-certification body across Spanish-speaking markets, on revenue share. You called it yourself on the call — it is one of our strengths, and it is exactly where we press. Highest margin, lowest CAC, now riding a global brand.
Third — the funnel. You do not monetize Mexico by selling courses; you monetize it by owning the machine that manufactures qualified investors in Spanish. Every platform that ever looked south hit the same wall: no acquisition machine, no trust layer, no local muscle. That machine now exists, profitable on its own P&L. So the day one of your platforms — the next one, whichever it turns out to be — is ready to cross into Spanish-speaking America, its customer base will not need to be built. It will need to be called. The adjacency is live today: our Talent Hub places the professionals we certify — your data-for-talent thesis, LATAM edition. And the exit math holds at the top of this playbook: IG paid US$1B for an educated audience wired to a funnel, and the buyers of the Spanish-speaking audience will find exactly one asset to bid on. I said it to you live, and it is still the whole thesis in two sentences: we train how to invest. The knowledge factory is ours; the investment factory is yours.
Content built thinkorswim. Community built tastytrade. Both exits are on the tape — US$606M and US$1.05B. Latin America is where that trade hasn't been put on yet. We've spent 20 years building the book.
Talk to the founder