Setup

StoneX Group ($9B) is a financial intermediary: broker-dealer, FCM, physical precious metals dealer, and correspondent clearing provider for regional broker-dealers. The May 26 8-K is a second 3-for-2 stock split announcement — noise. The substance is Q2 FY2026 earnings (May 6), the second consecutive record quarter for the precious metals franchise and the first full quarter with RJO integration.

What the Filing Says

Q2 FY2026 (quarter ended March 31, 2026): Net income $174.3M (+143% YoY), ROE 37% on tangible book. H1 net income $313.3M (+100% YoY). Physical contracts net operating revenue: $164.6M (+239% YoY). H1 physical contracts total: $300M — already exceeding the FY2025 full-year baseline in six months.

The mechanism: SNEX is the largest nonbank FCM for metals futures, a wholesale/retail physical bullion dealer, and operates vault custody and London Good Delivery refining. When gold and silver prices rise, physical contract notional scales directly with price. Exchange ADV counts contracts; SNEX's physical revenue counts dollars. This is why SNEX's +239% outpaces CME's metals ADV +130% — price-level amplification on top of volume growth.

RJO integration quantified: 37.8M listed derivative contracts (39% of Q2 total), $6.4B average client equity contributed. CEO on record: synergies "on track to be substantially completed during the course of this fiscal year."

One monitoring flag: Q2 Commercial bad debt $12.4M vs. $0.1M prior year (124× increase). Origin unknown from the press release; Q2 10-Q footnote pending.

Second stock split: management's second 3-for-2 in four months (effective July 20, 2026). Management: "make stock ownership more accessible to employees and investors." Two splits in four months signals confidence in continued price appreciation.

Independent corroboration — CME Q1 2026: Record ADV 36.2M contracts, first time in CME history all six asset classes hit records simultaneously. Metals ADV +130% (gold +133%, silver from 102K to 405K contracts). CEO Terry Duffy: silver went from $50 to $118. CFO: most significant fee changes were in "the metals complex, particularly precious metals micro complex." ICE: +45% total ADV YoY, record monthly volume March 2026. These are independent data sources. They confirm the regime before SNEX's physical amplification layer.

What the Market Thinks

Price $113.91, up 94.8% over one year. Two analysts covering; no options market. Forward P/E 17.13× — implied Street NTM EPS: $6.65, against H1 FY2026 actuals of $3.74. The Street's Q3 FY2026 estimate: $1.23 (Jefferies, sole estimate) against Q2 actual of $2.07. The Street is pricing a 41% sequential EPS decline.

Beat pattern: Q1 +26.3%, Q2 +32.0% — accelerating. The Street has had two quarters of this data. They still model normalization.

Back-solving from the current price using a three-state scenario: market-implied P(metals regime sustained) ≈ 43%. Our view: 60%. Edge: 17 percentage points on the state that drives most of the return.

Why the Gap Exists

The stock doesn't move with gold prices on a daily basis. OLS regression over the trailing 90 days: GLD β = 0.11 (p = 0.64, statistically indistinguishable from zero). SLV β = −0.005 (p = 0.97). SNEX's daily price does not co-move with gold or silver. The metals revenue materializes at quarterly earnings; between events the market prices SNEX as a financials stock (XLF β = 0.80, p = 0.04). Quant models and factor-based strategies see a financial intermediary. The metals exposure sits in the alpha intercept, not in a systematic factor loading — which is exactly why it hasn't been arbitraged away.

Two analysts, slow model update cycles. The $1.23 Q3 estimate was built before CME and ICE's independently corroborating exchange data was available and synthesized. Institutional investors requiring research coverage or options liquidity to size a position cannot participate cleanly. No implied vol surface exists to read the market's tail distribution.

The market priced two record quarters and stopped updating. At 43% implied P(sustained), the current price already reflects some regime persistence. It doesn't reflect CME/ICE cross-exchange confirmation that the regime is structural across the entire exchange and OTC ecosystem simultaneously.

Risks

1. Metals regime reversal. Gold and silver prices are the primary driver. Physical inventory $1,038.8M on balance sheet at March 31. A sustained decline >30% compresses Commercial segment revenue, increases bad debt risk, and erodes the earnings multiple. This is the tail that ends the thesis.

2. Insider selling — Form 4 codes unknown. $19.7M from four executives (O'Connor, Lyon, Maurer, Perkins) in a single week, 4-5 days post-Q1 record earnings, at $124-127. If these were discretionary sales — not 10b5-1 plans — the people with the most information sold at the peak after seeing Q1 results and while Q2 was underway. The Q2 bad debt spike ($12.4M) was accumulating during this period. This is the open question with the most asymmetric impact on conviction.

3. Commercial bad debt origin. $12.4M Q2 vs. $0.1M prior year. Cross-check of IBKR and BNY Mellon Q1 2026 shows no sector-wide stress — zero bad debt at IBKR, net credit recovery at BNY — suggesting a SNEX-specific counterparty event, most likely one client caught on the wrong side of silver's $118→$86 reversal. Q2 10-Q footnote resolves this.

4. BTIG litigation. $200M claim (≈8% of market cap), DOJ and SEC subpoenas active, StoneX suing own insurers. Professional fees $11.3M in Q2. Outcome unknowable.

5. Material weakness. KPMG adverse opinion on internal controls (FY2025). Repo/reverse repo netting error. Remediation disclosed as employee training only — no system overhaul. Restatement risk is not closed.

Catalysts

August 4, 2026 (68 days): Q3 FY2026 earnings. Street at $1.23. A print above $1.65 is the third consecutive large beat against a consensus that has modeled two consecutive misses. Three-beat confirmation at this scale forces coverage expansion from 2 to 3+ analysts and institutional model updates. Sub-$1.30 validates the normalization case.

Q2 10-Q filing (2-4 weeks): bad debt footnote detail. Single isolated counterparty closes the risk entirely. Multiple counterparties or structural language opens a new bear case.

Form 4 disclosures (ongoing): 10b5-1 vs. discretionary determination on February 2026 insider sales. This is binary for conviction.

July 20: second stock split takes effect. No intrinsic value change; retail/employee accessibility improves.

What Would Change Our Mind

• Q3 physical contracts < $80M: regime reverting; exit
• Q3 bad debt > $8M for a second consecutive quarter: structural credit stress confirmed; downgrade factor scenario to normalizes
• Form 4 codes confirm discretionary selling by O'Connor/Lyon/Maurer: reduce to minimum sizing
• Gold sustained below $4,000 for two consecutive quarters: update factor scenario, reduce sustained probability
• BTIG judgment or settlement > $150M: material capital event, reassess

Forward EV (12-month, pre-second-split basis)

At $113.91: standalone Sharpe 0.24 (below 0.5; contributes positively as idiosyncratic portfolio position, %idio variance 75.3%). At $97 (50-day MA, options max pain): Sharpe 0.66, edge widens to 39 percentage points on key state. Alpha decay cost of waiting: 0.19%/day (365-day half-life).

Memo LR: 1.5

Two consecutive record quarters corroborated by independent exchange data (CME/ICE) against a market pricing 43% regime-sustained probability and a Street model expecting 41% sequential EPS decline. Real bears: $19.7M insider selling with unknown 10b5-1 status, BTIG tail, material weakness. The edge is in probability differential and information asymmetry, not narrative confidence.

Evidence