2026-08-17 · EN
SBC — SBC Medical Group Holdings Inc
MonitorOpen this week — full figures
SBC Medical Group Holdings Incorporated (NASDAQ: SBC) — deep-value analysis
Analysis date: August 17, 2026 · Reference price: $3.66 (close 08.14.2026, yfinance) · Market cap: $375.4M · Shares: 102,576,943 · Reporting currency: USD (operations in JPY) · Auditor: MaloneBailey LLP, Tokyo (since 2023)
Primary sources: 10-K FY2025 filed 03.27.2026 · 10-Q Q1 2026 filed 05.14.2026 · 10-Q Q2 2026 filed 08.13.2026 · 10-K FY2024 filed 03.28.2025 · DEF 14A filed 05.28.2026 · 424B7 filed 04.17.2026 · 8-K/EX-99.1 of 08.13.2026 · data pack data-pack-SBC-20260817.md · research brief research-SBC-20260817.md · Monte Carlo mc-SBC-20260817.json.
Executive summary (1 page: thesis, estimated value, verdict)
SBC Medical Group Holdings is a management services organization (MSO) that collects franchise, procurement, management-services and equipment-leasing fees from a group of Japanese medical corporations operating aesthetic clinics under the “Shonan Beauty Clinic” brand. The figure that defines the company is neither margin nor growth: 91.5% of FY2025 revenue ($158,860,970 of $173,607,489) came from affiliated parties — medical corporations whose “members” (the holders of voting rights) are relatives of CEO Dr. Yoshiyuki Aikawa, who in turn controls 81.7% of SBC’s votes (10-K FY2025, Note 20 and the income statement, p. F-5; DEF 14A 05.28.2026). Service pricing is set, on both sides of the table, by the same person.
It’s not an abstraction. In April 2025 management revised the fee structure toward affiliated clinics; the effect was a 15.48% drop in consolidated revenue, from $205,415,542 to $173,607,489, with franchise revenue falling 24.72% and management-services revenue 44.22% (10-K FY2025, Item 7). In August 2026, the same management announced fee increases that “would be expected to add approximately $15 million annually” (release 08.13.2026, EX-99.1). Therefore, this company’s top line is not a market outcome; it’s an internal allocation decision between the pockets of the same owner. This is the core of the thesis and of the risk alike.
Beneath this layer, the underlying asset is healthy and cheap. The network has reached 287 locations (+34 y/y), 6.9 million visits over the trailing 12 months (+10%) and an average spend per visit of $287 (+9%) (EX-99.1, 08.13.2026). LTM operating margin is 37.3%, ROIC (after-tax EBIT / capital employed) is approximately 29.5%, and the balance sheet has $184.3M cash against $37.9M interest-bearing debt as of 06.30.2026 — $146.7M net cash, i.e. 39% of market cap. Enterprise value is therefore ~$229M for an asset generating ~$65M annual EBIT: 3.5× EV/EBIT. On the surface, absurdly cheap.
My FCF bridge (detail in the valuation chapter) starts from CFO and subtracts real capex (including advances for fixed assets, which the automated bridge doesn’t see) and lease principal, then normalizes for non-recurring gains and FX. Result: normalized owner earnings of $39.0 million — below the $60.6M of gross trailing-12-month FCFE (inflated by an $11.2M tax deferral in H1 2026 and by the end of the affiliated-party float drain) and above the gross average of the last three years ($26.6M, depressed by the same drain). The 10.4% yield on market cap is consistent with what brokers publish (2026 consensus: EPS $0.45, i.e. ~$46M net profit — my figure is 15% more conservative).
The triangulated valuation gives a range of $3.39–$5.97 per share, with a central point around $4.05–4.40. Monte Carlo over the mechanical DCF (20,000 scenarios, OE $39M, g1 6%, r 12%, gt 2%) returns a median intrinsic value of $5.97 and a 98.5% probability the stock is undervalued. This probability must be read correctly: it measures flow uncertainty, not the risk that the flows never reach the minority shareholder. The four models incorporating this latter risk (bear, EPV, governance-adjusted, own multiples) cluster between −7% and +26% MOS, with a median of +11%.
Verdict: SPECULATIVE-MONITOR. My calculated GBL score (the tracker doesn’t yet have one — see the final chapter) is 50.8%, in the MONITOR band. The cheapness is real and measurable; the discount is, however, earned, not mistaken. Documented in the filings, in just 18 months: two consecutive fiscal years with unremediated material weaknesses in internal control, exactly around the related-party transaction approval process; a $12,000,000 salary for the CEO in 2025 (23.5% of net profit); a $1.6M salary for the CEO’s mother, not timely identified as a related-party transaction; a plane sold to a CEO entity, later marked up $10.35M, booked to equity, not to profit; a subsidiary guarantee on the CEO’s personal debt; and — most tellingly — the company paid ~$1.3–1.4 million of the costs for the secondary offering through which the CEO sold 3.1 million of his shares in April 2026, while the company’s own $20M buyback program, authorized in December 2025, bought back zero shares in the first six months of 2026.
Position sizing must reflect that the thesis depends on the behavior of a single person, not a business model. A small position, with a clear tracking thesis: the first buyback actually executed from the $20M program would be the signal that validates the thesis; a new downward “fee structure revision” would be the signal that breaks it.
The business and the moat (how it makes money, competitive advantage, durability)
Revenue mechanics. SBC doesn’t own clinics. Japanese law (the Medical Care Act) prohibits for-profit corporate ownership of medical corporations; subsidiary SBC Medical Group Co., Ltd. (“SBC Japan”) is registered as a medical service corporation and provides services to medical corporations (MCs) through franchise and services contracts. As of 12.31.2025 the contractual network comprised seven franchised MCs (Shobikai, Kowakai, Nasukai, Aikeikai, Jukeikai, Ritz Cosmetic Surgery and, from June 2025, Furinkai) plus service contracts with Junikai, Misakikai and Miotokai (10-K FY2025, Item 7). SBC holds “mochibun” — the right to residual assets on dissolution — in six of them, booked at cost ($17,837,293 as of 12.31.2025), with no voting rights and no right to profit distributions.
The five 2025 revenue streams (10-K FY2025, Item 7):
| Stream | FY2025 | FY2024 | Δ% | What it is |
|---|---|---|---|---|
| Franchise | $45,943,241 | $61,033,032 | −24.72% | Brand licensing, patents, trademarks, consulting |
| Procurement | $56,053,171 | $54,814,399 | +2.26% | Purchasing advertising and medical materials for MCs |
| Management services | $29,628,534 | $53,113,155 | −44.22% | Loyalty program, staff secondment, accounting/HR |
| Equipment leasing | $23,032,651 | $16,141,714 | +42.69% | Operating and sales-type leasing, laser hair-removal machines |
| Other | $18,949,892 | $20,313,242 | −6.71% | Fit-outs, real estate brokerage, beauty services |
| Total | $173,607,489 | $205,415,542 | −15.48% |
Concentration is brutal. From Note 20: Shobikai $40,953,913 (23.6% of total revenue), Nasukai $39,559,694 (22.8%), Kowakai $37,101,866 (21.4%). The top three medical corporations = 67.7% of revenue; the top six = 86.1%. All have the CEO’s relatives as “members.” An industrial company with 23% customer concentration would be a top-tier risk factor; here there are three, and none negotiates at arm’s length.
What the moat actually is. There are three candidates and none is what it seems.
The brand. “Shonan Beauty Clinic” has 26 years of operation and, per company materials, ~31% of locations among mid-size and large clinic groups in the Japanese aesthetics market (a figure picked up in the brief from an aggregated search result, not opened at the primary source — I treat it as indicative, not verified). But the brand doesn’t belong to SBC’s revenue stream except indirectly: SBC licenses it to the MCs. If the relationship breaks, the brand stays with SBC, but the clinics, doctors and patients stay with the MCs.
Procurement scale effect. Real and verifiable: with 287 locations, SBC negotiates advertising and consumables at a price a standalone clinic can’t reach. Consolidated gross margin of 73.3% in FY2025 suggests the intermediation margin is substantial. This is the model’s most authentic advantage and is replicable only by a similarly sized competitor.
The contractual relationship. Not a moat, a dependency. Contracts can be unilaterally repriced — and were, in both directions, over 16 months. A moat protects you from someone outside; here the risk is from within.
Durability. The Japanese aesthetics market is cited in the brief at ¥631 billion in 2024, +6.2% y/y, with CAGR projections above 12% for cosmetic surgery through 2033 — figures from aggregated sources, not verified at the primary source, so I use them only as an order of magnitude. The company’s own data is more convincing: visits grew 10% LTM, and spend per visit 9% — so volume and price rise simultaneously at the end-clinic level. That’s real pricing power in the consumer market, and SBC captures a slice through franchise fees.
Durability counter-arguments: (1) Japanese demography — the population is shrinking, and the stated target demographic (30–40 year-olds, extending toward men) is a strategy to offset base contraction; (2) South Korean competition entering the Japanese market with aggressive pricing (mentioned in the brief, unquantified); (3) regulatory risk — the entire MSO arrangement exists because Japanese law doesn’t allow corporate ownership of clinics; a tightening of interpretation would hit the model directly; (4) FX — at ¥158/$ in H1 2026 versus ¥148/$ in H1 2025, yen depreciation cost the company $5,715,671 in reported revenue. At constant rates, core-segment revenue grew ~14% in Q2 2026, not the 3% the dollar figure shows (10-Q Q2 2026, Item 2).
Moat verdict: partial. There’s genuine economy of scale in procurement and a well-known brand; there’s no mechanism preventing the owner from moving value between the entities he controls. A minority investor buys a fee stream set discretionarily by the majority shareholder.
Management and capital allocation (track record, buybacks/dividends/acquisitions, skin in the game)
Operational track record. Dr. Yoshiyuki Aikawa has led the group for 26 years and built, from nothing, Japan’s largest network of aesthetic clinics. Operating results have been remarkably stable: EBIT $70,660,066 (FY2023), $70,303,710 (FY2024), $67,486,398 (FY2025), $65.4M (LTM June 2026) — four consecutive periods between $65M and $71M, despite revenue varying from $193.5M to $205.4M and back to $173.6M. That’s real cost discipline.
Going public. The company went public on 09.17.2024 through a merger with SPAC Pono Capital Two, which renamed itself SBC Medical Group Holdings. The stock hit $16.00 in August 2024 and $2.87 in May 2026 — −82% from the high. In cash terms, the listing brought $11.7M net (10-K FY2024, financing activities), and the accounting cost was a $13,022,692 warrant expense to a provider who supported the listing process (FY2024). It wasn’t a listing for capital; it was one for liquidity and visibility.
Dividends: zero. The company has never paid a dividend. The 10-K FY2025 says it “is evaluating alternative capital allocation methods, including dividends” — unchanged wording for two years.
Buybacks — the complete history and why it matters. In 2025, 1,034,308 shares were repurchased for $5,049,997, i.e. an average price of $4.88 (statement of changes in equity, 10-K FY2025). Of which 512,809 shares for $2,415,262 in Q2 2025 (average price $4.71). On 12.29.2025 the board authorized an additional program of up to $20 million, effective 12.31.2025 – 12.31.2026, and on 12.30.2025 the S-3 registration became effective; the 12.31.2025 release says use of excess cash and future free cash flow was anticipated.
Fact-check as of 06.30.2026: treasury shares are unchanged — 1,304,308 units at a cost of $7,749,997, identical to 12.31.2025 (10-Q Q2 2026, balance sheet and statement of changes in equity). Zero shares repurchased in six months, out of a $20M program, with the stock trading between $2.87 and $4.53. The company bought at $4.88 and bought nothing at $3.00.
What it did instead with the money over the same period. In April 2026, the CEO sold 3,100,000 shares at $3.25, through a secondary offering underwritten by Maxim Group (sole bookrunner) and Roth Capital, netting $9,369,750 net of commissions (424B7 of 04.17.2026). The closing price on 04.16.2026 had been $4.42 — the placement was done at a 26.5% discount. And the prospectus states, verbatim: “The expenses of the offering, not including the underwriting discounts and commissions for which the selling stockholder is responsible, are estimated at $1.3 million … and are payable by us.” The 10-Q for Q2 2026 confirms the accounting effect: consulting fees rose by $1,247,116 for the quarter, “mainly due to $1.4 million transaction costs incurred by the Company in connection with the secondary public offering of common stock sold by the Company’s CEO.”
So: minority shareholders paid ~$1.3–1.4 million so the majority shareholder could monetize part of his stake, below the market price, while the company’s buyback authorization sat untouched. This isn’t an interpretation — it’s quoted from the prospectus and the 10-Q.
Acquisitions. Three in 20 months, all paid in cash:
- Aesthetic Healthcare Holdings (Singapore), November 2024, $4.2M net of cash acquired. Consolidated with a 3-month reporting lag.
- MB career lounge Co., Ltd., 07.17.2025, ¥2,040,000,001 ($14,150,453) plus $975,410 of capitalized transaction costs. Treated as an asset acquisition, not a business combination — so no goodwill; $22,597,646 allocated to intangibles. Brings “JUN CLINIC” (Misakikai/Miotokai) into the network.
- Waqoo, Inc. (TSE Growth: 4937), December 2025. A tender offer at ¥1,900/share, 575,052 shares bought; additionally, the CEO — Waqoo’s largest shareholder — transferred the rest of his shares to SBC Japan through an off-market transaction, outside the tender offer. Total consideration $19,365,196, identifiable net assets $23,666,370, minority interests $14,883,778, goodwill $10,582,604 and intangibles $25,876,011 (amortized over 11 and 16 years), with a related deferred tax liability of $9,164,822 (10-K FY2025, Note 4).
So the third acquisition was, partly, a purchase from the CEO himself, at a price whose off-market component isn’t disclosed separately. Waqoo represented 13% of total assets as of 12.31.2025 and was excluded from the 2025 internal control assessment (Item 9A).
- OT Midco / OrangeTwist, 12.29.2025: $20,000,000 cash for ~18.2% of the voting rights of a US med-spa operator, accounted for via the equity method, plus an additional $5.0M commitment due December 2026. In the first full quarter, the investment produced a loss of $568,652 (Q2 2026).
Capital allocation balance sheet 2025–H1 2026: ~$53.5M into acquisitions and stakes (MB career lounge $14.2M + Waqoo $19.4M + OrangeTwist $20.0M), $5.0M into buybacks, $0 dividends, and cash still rose from $125.0M to $184.3M. Simultaneously, bank debt rose from $6.6M to $42.8M (12.31.2025) — $34,753,946 was borrowed in 2025 at an average cost under 1% per year (interest paid $160,583). This isn’t a criticism: cheap Japanese debt at 0.65% against cash yielding nothing is neutral; what matters is that the debt funded M&A, not dividends or buybacks — a good accounting signal.
Skin in the game. Dr. Aikawa owns 83,839,460 shares = 81.7% of votes as of 05.20.2026 (78,839,460 direct + 5,000,000 through Aikawa Investment Co., Ltd.). Separately, Aikawa Equity Management Co., Ltd. reports 5,284,500 shares (5.2%) — so 86.9% of capital is in family-linked hands, and the actual free float is on the order of 13–19 million shares, ~$50–68M at the current price. Average daily volume over 3 months is 123,619 shares ($452k/day): a $100,000 position is 22% of one day’s trading volume.
“Skin in the game” isn’t the problem. The problem is that, at 81.7%, there’s no market mechanism to correct a bad or self-interested decision, and the company is a “controlled company” under Nasdaq rules — it can forgo an independent board majority and fully independent committees. The 2026 proxy says the company doesn’t intend to use these exemptions, but reserves the right.
The board. Reduced from 5 to 4 members right before the 2026 meeting (Mike Sayama not standing for re-election), 3 of 4 qualified as independent. Total compensation of the three non-executive directors in 2025: $160,256 — for overseeing a company with 24 disclosed related parties and two consecutive fiscal years of material control weaknesses.
What changed over the last 4 quarters (balance sheet item by item from the data pack, margins, cash conversion — explaining EVERY large variation)
Comparison 06.30.2025 → 06.30.2026, with an explanation for every major move. Figures from the 10-Q and 10-K balance sheets; “Δ 4Q” is the variation from the comparable quarter a year earlier, with noted exceptions.
| Item | 2025-06 | 2025-12 | 2026-03 | 2026-06 | Δ 4Q | Explanation verified in the filing |
|---|---|---|---|---|---|---|
| Total assets | $315.3M | $380.4M | $388.0M | $406.7M | +29.0% | Consolidation of Waqoo (13% of assets as of 12.31.2025) + the OrangeTwist investment |
| Cash | $152.7M | $163.8M | $167.3M | $184.3M | +20.7% | H1 2026 CFO $31.7M; no dividend, no buyback |
| Receivables – affiliated parties | $48.9M | $27.5M | $33.8M | $40.7M | −16.7% | Seasonal pattern: builds up in H1, collected in H2. At 06.30.2025 it was $48,920,843 — so year/year it DECREASED |
| Receivables – third parties | ~$2.2M | $2.4M | $3.0M | $2.5M | +13% | Insignificant (0.6% of assets) |
| Inventory | $1.7M | $2.8M | $2.3M | $3.0M | +74.7% | A tiny base ($3.0M = 1.7% of revenue); includes Waqoo inventory (D2C products) |
| Financing lease receivables – affiliates | $14.4M | $26.6M | — | $25.4M | +76% | SBC finances MC equipment; consumed $12.7M of CFO in 2025 |
| Intangibles, net | ~$1.5M | $47.7M | — | $45.7M | +2,900% | $25.9M Waqoo (customer relationships + technology) + $22.6M MB career lounge |
| Goodwill | ~$4.6M | $15.4M | — | $15.2M | +230% | $10.6M from Waqoo |
| Equity-method investments | $0 | $20.3M | — | $19.7M | n/a | OrangeTwist; already at a $0.57M loss |
| Total liabilities | $70.6M | $117.1M | $117.6M | $126.7M | +79.4% | Bank loans $34.8M taken in 2025 + $9.2M deferred tax liability from Waqoo |
| Bank loans | ~$7.1M | $42.8M | — | $37.8M | +432% | M&A financing; average cost <1.2%/yr; maturity 2026 $9.1M, 2027 $14.0M |
| Tax payable | ~$12.7M | $8.8M | — | $19.3M | +52% | +$11.2M in H1 2026 = payment deferral, not new profit; paid in H2 |
| Equity (SBC) | $244.6M | $248.3M | $255.3M | $263.8M | +7.9% | H1 profit $22.0M, minus $9.3M FX translation loss |
| Cumulative FX translation loss | −$35.9M | −$57.3M | −$61.5M | −$66.6M | −85% | Yen depreciation; $66.6M is 25% of equity |
| Minority interests | $0.1M | $15.0M | $15.0M | $16.1M | n/a | 100% Waqoo |
| Shares outstanding | 103.5M | 102.6M | 102.6M | 102.6M | −0.9% | Buybacks in 2025; zero in H1 2026 |
Margins, by quarter. Gross margin: 69.2% (Q2 2025) → 73.1% (Q2 2026); over the half, however, 74.7% → 71.9%, so compression. The reason is in the mix: franchise and procurement revenue — the high-margin ones — fell 27.7% and 14.4% respectively in H1, while growth came from management services (+78.7%) and the Waqoo consolidation, which brings product costs. Operating margin: 42.7% (H1 2025) → 39.8% (H1 2026).
The variation that matters most and that the market misreads. The 08.13.2026 release headlines “Net Income Attributable to SBC Medical Up 335%.” Attributable net profit indeed went from $2,458,240 to $10,692,822. But the cause isn’t operational: in Q2 2025, the effective tax rate was 81.98% ($11,100,509 of tax on pretax profit of $13,540,361), because the markup on the plane sale to a CEO entity was treated as taxable income in Japan with no corresponding accounting revenue (10-Q Q2 2026, Item 2, explicit). In Q2 2026 the rate was 40.01%. The 335% net-profit increase is, mostly, an artifact of a base of comparison distorted by last year’s related-party transaction.
The clean line is operating income: $14,554,192 → $18,960,550, +30.28%. And here too one must go a step further: of the $4,406,358 increase, $3,756,172 (85%) comes from the consolidation of the Waqoo segment, while the core segment grew only $689,942 (+4.74%). Sequentially by quarter, the core segment’s operating income actually fell: $17.72M in Q1 2026 (excluding Waqoo) → $15.24M in Q2 2026.
The corrective nuance, in the company’s favor: at constant exchange rates, core-segment revenue grew ~14% in Q2 (the yen depreciated from ¥144.03 to ¥159.33 per dollar; the unfavorable impact quantified by the company is $4,747,098 per quarter). So the real growth is much better than what’s reported in dollars, and the “reacceleration” isn’t false — but its source isn’t the one the headline suggests.
Cash conversion — the best news in the report. CFO: $−6,411,168 in H1 2025 → $+31,741,248 in H1 2026. Over the trailing 12 months, CFO reaches $62.8M against net profit of $49.6M — a ratio of 127%, after two years in which the ratio was 44% (2024) and 48% (2025). The explanation is in the next chapter and is structural, not cosmetic; but $11.2M of the $31.7M is a tax deferral, so the adjusted figure is ~$20.5M for the half.
Balance sheet analysis — Quality of Earnings (Thornton O’Glove method)
O’Glove’s premise: profit is an opinion, the balance sheet is closer to a fact. At a company where 91.5% of revenue comes from related parties, opinion counts doubly — because the customer is also an opinion.
1. Receivables vs. revenue (% AR growth vs. % revenue growth) and 3-year DSO
| Period | Total receivables | Revenue (annual/annualized) | DSO | Δ% AR | Δ% Revenue | Spread |
|---|---|---|---|---|---|---|
| 12.31.2023 | $35,113,749 | $193,542,423 | 66.2 days | — | — | — |
| 12.31.2024 | $30,260,113 | $205,415,542 | 53.8 days | −13.8% | +6.1% | −19.9pp (favorable) |
| 12.31.2025 | $29,899,751 | $173,607,489 | 62.9 days | −1.2% | −15.5% | +14.3pp (unfavorable) |
| 06.30.2025 | ~$51,100,000 | $181.4M annualized | ~102.8 days | — | — | — |
| 06.30.2026 | $43,237,261 | $184.5M annualized | ~85.5 days | −15.4% | +1.7% | −17.1pp (favorable) |
Three observations that partly contradict each other and must be read together.
First: over fiscal 2025, receivables fell only 1.2% while revenue fell 15.5%. That’s the classic O’Glove signal of relative deterioration: DSO rose from 53.8 to 62.9 days. The likely cause, however, isn’t “fake” revenue, but that MCs were paying lower fees on a growing activity base (287 clinics) — the receivable per clinic stayed, the fee fell.
Second: the year/year comparison at June 30 is favorable — related-party receivables fell from $48,920,843 to $40,748,686, and DSO from ~103 to ~86 days. The data pack doesn’t see this, because it doesn’t have the 06.30.2025 balance sheet in its 8-quarter series; I read it directly in the 08.13.2025 10-Q. The correct conclusion is that, year/year, receivable quality has improved, not degraded.
Third, and most important: there’s a violent, structural seasonal pattern. Receivables from affiliates grow massively in H1 (+$14,570,601 in H1 2026, +$17,039,113 in H1 2025, per the cash flow statements) and are collected in H2. The 2026 trajectory: $27.5M (Dec) → $33.8M (Mar) → $40.7M (Jun). Effectively, SBC finances the working capital of the CEO’s family clinics for free, six months of every year, at a volume of $13–18 million. At a normal company, credit exposure to the largest customer would be negotiated; here the payment term is set by the same person on both sides.
The credit-loss allowance on related-party receivables was $2,762,999 as of 12.31.2025 (9.1% of the gross balance), down from $2,836,013 — with a new provision of $0 and a reversal of $73,014 in 2025. A zero provision in a year when revenue from the same customers fell 18.6% is an accounting choice, not a finding.
2. Inventory vs. COGS, DIO and structure
| FY2023 | FY2024 | FY2025 | LTM Jun. 2026 | |
|---|---|---|---|---|
| Inventory | $3,090,923 | $1,494,891 | $2,792,617 | $2,979,818 |
| Cost of sales | $56,238,385 | $49,365,035 | $46,323,767 | $48,305,000 |
| DIO | 20.1 days | 11.1 days | 22.0 days | 22.5 days |
The data pack flags “inventory +74.7% vs. COGS −4.8%” as a negative signal. I don’t validate it. Inventory represents 1.7% of revenue at a management-services company that manufactures nothing; the $1.3M increase is fully explained by the Waqoo consolidation (D2C cosmetics products, $830,765 acquired at purchase) and the rebuilding of a base that had fallen abnormally in 2024. Finished-goods structure isn’t disclosed separately in the filings — nor would materiality make sense at this scale. The inventory signal is a false positive at SBC; I remove it from the verdict.
3. Debt — gross/net, maturity schedule, cost, covenants, 3-year evolution
| 12.31.2023 | 12.31.2024 | 12.31.2025 | 06.30.2026 | |
|---|---|---|---|---|
| Bank loans | $0.16M | $6.60M | $42.83M | $37.75M |
| Notes payable – affiliates | $3.37M | $0.03M | $0 | $0 |
| Finance lease | — | $0 | $0.25M | $0.18M |
| Gross interest-bearing debt | $3.53M | $6.63M | $43.08M | $37.93M |
| Cash + short-term investments | $103.02M | $125.04M | $164.09M | $184.62M |
| Net cash | $99.49M | $118.41M | $121.01M | $146.69M |
| Operating leases (off-balance-sheet as economic debt) | $9.81M | $5.58M | $8.55M | $10.30M |
Maturity schedule (10-K FY2025, Item 7): 2026 $9,099,046 · 2027 $13,954,055 · 2028 $7,437,818 · 2029 $6,906,623 · 2030 $5,435,942 · total $42,833,484. The next 24 months require $23.05M — covered 8× just from cash.
Cost: interest paid $160,583 for FY2025 on average debt of ~$24.7M = 0.65%/yr; in H1 2026, $243,435 annualized on ~$40M = ~1.2%. Cheap Japanese debt, with near-zero refinancing risk at this liquidity level.
Covenants: not disclosed in the 10-K. The absence is itself partial information — at a debt/EBITDA ratio of 0.6× it’s plausible there are no strict financial restrictions, but I can’t confirm.
Debt used for dividends/buybacks? No. The $34,753,946 borrowed in 2025 corresponds, in order of magnitude and timing, to $22,941,701 paid for subsidiary acquisitions and $20,062,642 for OrangeTwist. Buybacks ($5.0M) were smaller than CFO ($24.7M). This test passes clean.
4. Abnormal discretionary expenses and accounting estimate changes
Comparison FY2024 → FY2025 (10-K FY2025, Item 7), with revenue falling 15.5%:
| Expense | FY2024 | FY2025 | Δ | % of 2025 revenue | Reading |
|---|---|---|---|---|---|
| Salaries and benefits | $26,843,524 | $26,472,154 | −1.4% | 15.2% (from 13.1%) | Not cut; grew as a share |
| Consulting and professional fees | $14,555,087 | $17,022,128 | +17.0% | 9.8% | Growth on falling revenue |
| Advertising | $2,782,944 | $3,122,660 | +12.2% | 1.8% | Maintained |
| Recruiting | $1,570,299 | $703,213 | −55.2% | 0.4% | Discretionary cut |
| Amortization | $2,258,364 | $1,738,010 | −23.0% | 1.0% | Mechanical, post-2024 write-off |
| Impairment of intangibles | $15,058,965 | $0 | −100% | — | Patent right fully written down in 2024 |
| Stock compensation | $13,022,692 | $0 | −100% | — | Warrants issued at listing, non-recurring |
And H1 2026 vs. H1 2025: advertising $1,656,099 → $1,034,952 (−37.5%), local taxes and fees $837,279 → $565,098 (−32.5%), amortization $933,168 → $657,075 (−29.6%), while consulting rose from $7,176,118 to $8,599,741 (+19.8%).
O’Glove’s verdict on this table would be ambivalent. On one hand, the 39.8% operating margin in H1 2026 is partly supported by cuts: own advertising −37.5% and recruiting −55.2% aren’t sustainable savings at a company selling marketing and staff-recruitment services. On the other hand, the expense growing the most — consulting, at 9.8% of revenue — isn’t a business investment: $1.4M of the Q2 2026 increase is the cost of the offering through which the CEO sold his shares, and the rest is “legal, tax and market research associated with public-company compliance” (10-K FY2025). A company spending $17M on consultants and $3.1M on its own advertising, on $173.6M of revenue, has a holding-company cost structure, not an operator’s.
Accounting estimate changes: I identified no policy or estimate changes with a material impact disclosed in FY2025 or H1 2026, beyond the application of the “multi-period excess earnings” method to valuing the Waqoo intangibles — a new estimate, but one related to an acquisition, not a rewrite of an existing estimate. One item, however, deserves flagging as a policy change with a direct revenue effect: from June 2025, the loyalty program was revised so that point redemptions (which reduced management revenue) fell — which alone explains the 78.7% increase in management-services revenue in H1 2026. Similarly, “prior to June 2025, any over-collection resulting from such [vendor] discounts was returned to MCs; since June 2025, it is retained for future transactions” (10-K FY2025, procurement revenue recognition policy). Two policy changes, both in SBC’s favor and to the MCs’ detriment, both decided by the same person who also controls the MCs.
5. Summary: net profit vs. CFO divergence (accruals) over 3 years
| Fiscal year | Net profit (total) | CFO | CFO/NP | Sloan accruals (NP−CFO)/average assets |
|---|---|---|---|---|
| FY2023 | $38,560,606 | $50,670,322 | 131% | negative (favorable) |
| FY2024 | $46,689,892 | $20,582,933 | 44% | +9.95% |
| FY2025 | $51,045,023 | $24,668,496 | 48% | +8.16% |
| Cumulative 2023–2025 | $136,295,521 | $95,921,751 | 70.4% | — |
| LTM Jun. 2026 | $49,604,000 | $62,821,000 | 127% | −3.66% (favorable) |
The figure to remember: over three years, only 70 cents of every dollar of reported profit turned into operating cash. That’s exactly the kind of divergence O’Glove puts on page one.
But the diagnosis matters more than the symptom. I tracked where the missing $40.4M went, in the cash flow statements:
- Notes payable to related parties: −$34,756,754 (2024) −$14,252,502 (2025) = −$49.0M. SBC had an interest-free trade payable to the MCs and paid it down.
- Advances from customers – related parties: −$9,144,031 (2024) −$6,693,127 (2025) = −$15.8M. A second float draining down (balance: $23.06M at end-2023 → $5.36M at end-2025 → $3.99M as of 06.30.2026).
- Financing lease receivables – related parties: −$5,991,486 (2024) −$12,746,857 (2025) = −$18.7M. SBC started financing the MCs’ equipment.
- Partly offset by declining customer credits (+$18.5M in 2024, +$15.8M in 2025).
The analytical conclusion, which differs from the mechanical one: the 2024–2025 cash gap wasn’t fictitious revenue, but the unwinding of a working-capital float the MCs had historically extended to the company. That float is now nearly exhausted — from ~$26.4M at end-2023 to ~$4.0M today — so it can no longer consume cash. Confirmation comes from the LTM: CFO 127% of profit, negative accruals. The counterpart is that a new float now flows the opposite direction: receivables from affiliates and financing lease receivables, together $66.1M as of 06.30.2026, versus $37.9M two years earlier. SBC went from being financed by its family-customers to financing them.
Quality of Earnings verdict: MEDIUM
Not WEAK, because: (a) there’s no evidence of revenue recognized without substance; (b) DSO improved year/year; (c) LTM cash conversion is 127%; (d) the auditor issued no qualification; (e) the 2024–25 accrual gap has an identifiable, closed structural explanation; (f) maintenance capex is negligible (~$2.1M LTM against $3.2M of amortization) — nothing is hidden in capitalizations.
Not HIGH, because: (a) 91.5% of revenue is administratively set by the controlling party; (b) two policy changes in 2025, both increasing SBC’s revenue at the MCs’ expense; (c) $15.5M of gains from related-party transactions bypassed the income statement and went directly to equity in 2024–H1 2026; (d) zero provision on receivables from customers whose volume fell 18.6%; (e) material control weaknesses exactly around the related-party transaction approval process, unremediated two years running.
Signal to watch at the Q3 2026 report (expected ~11.12.2026): related-party receivables. The pattern requires that from $40.7M (06.30) it falls toward $28–32M by 12.31. If as of 09.30.2026 they exceed $45M, the seasonal pattern has broken and it must be assumed the MCs can’t afford the $15M/year fee increase announced in August. Second, subordinate signal: whether a new “fee structure revision” gets announced.
CEO profile — Outsider traits (William Thorndike method)
Dr. Yoshiyuki Aikawa, board chairman and CEO, founder, physician by training, 26 years at the helm. Thorndike’s grid, point by point, with evidence from the filings.
1. Capital allocation as priority #1 — NO. The releases and the CEO’s letter of 08.13.2026 talk about an “AI-enabled MSO platform,” “three dimensions of growth,” the “Longevity market,” a “second growth engine.” There’s no capital-allocation indicator in the communication: no return on invested capital for acquisitions, no hurdle rate, no explicit comparison between buyback and investment. The only mention of share value is aspirational: “achieving fair equity valuation for our shares in the capital markets.” An Outsider doesn’t hope for a fair valuation; he exploits it.
2. Contrarian buyback — NO, the reverse. The timeline is unambiguous. Repurchased 1,034,308 shares at an average price of $4.88 in 2025 (including 512,809 at $4.71 in Q2 2025, when the stock was at local highs). Authorized a $20M program on 12.29.2025. Bought zero shares in H1 2026, with the stock between $2.87 and $4.53, verifiable by the fact that treasury shares are identical at 12.31.2025 and 06.30.2026 (1,304,308 units, $7,749,997). In the same window, the CEO sold 3.1 million shares at $3.25. He bought expensive, stopped when cheap, and personally sold cheap. The exact opposite of the definition.
3. Decentralization and frugality — NO. The company owned a plane, sold to a CEO-controlled entity. It spends $17.0M/year on consultants on $173.6M of revenue. It has offices in Irvine, California, and Tokyo, for a business operating in Japan. The CEO’s salary — $12,000,000 in 2025, $14,506,032 in 2024 — is 75 times the CFO’s ($416,072 in 2025) and represents 23.5% of attributable net profit and 6.9% of total revenue. For context: it’s 8.6× the company’s annual capex and 2.4× everything ever spent on buybacks. Frugality isn’t a trait here.
4. Independence from Wall Street — PARTIALLY YES, but for the wrong reasons. The company doesn’t offer quarterly guidance and doesn’t track adjusted EPS short-term in communications — on paper, an Outsider-type plus. But the independence is a byproduct of the fact that ownership is 81.7% in one person’s hands: there’s no market pressure because there’s no market. And the communication isn’t honest in the way Thorndike required: the Q2 2026 release headlines “+335% net income” without mentioning that the base was distorted by an 82% tax rate caused by a transaction with the CEO’s own affiliated entity; and headlines “Restructuring Complete, Growth Reaccelerates” without saying 85% of the operating-income increase comes from a consolidation, not the core business. The explanations exist — but in the 10-Q, not the release.
5. Skin in the game and long-term incentives — MIXED, tilted negative. 81.7% ownership is maximal. But the incentives are the reverse of what the grid requires: the CEO’s compensation is 100% cash salary, zero shares, zero options, zero variable component (DEF 14A, compensation table, 2024 and 2025). Someone owning 82% doesn’t need shares as an incentive — but $12M in cash, fixed, independent of performance, is exactly the structure Thorndike identifies as a sign of extraction. The 2024 equity plan has 15,000,000 units available and zero issued as of 12.31.2025 — so the rest of the team isn’t long-term co-invested either.
And the component that tips the balance: the sale of 3,100,000 shares in April 2026, with the company paying ~$1.3–1.4M of costs. An Outsider doesn’t sell at 4× profit; and if he sells, he doesn’t put the bill on the public company.
6. Proven contrarian behavior — NO. The reverse: a growth-through-acquisitions strategy paid cash across three jurisdictions simultaneously (Japan, Singapore, US), an AI narrative aligned with the fad of the moment (“SBC AI Lab,” a CTO appointed in March 2026 to turn data into a “strategic AI asset”), an announced entry into the “Longevity” market. The $20M investment in OrangeTwist (18.2% of a US med-spa operator, no control, already at a $0.57M loss in the first full quarter) is a non-liquid minority position in a market the company doesn’t know, paid with 5.3% of its own market cap — money that would have bought back 5.5 million shares at $3.66.
“How Outsider” score: 1 of 5
The point granted is for the absence of quarterly guidance and for the operating cost discipline shown by EBIT staying stable at $65–71M over four periods. The rest of the grid scores negative. Aikawa is a remarkably successful founder-operator in building the network; he isn’t a minority-shareholder-oriented capital allocator, and the evidence is from documents, not impression.
The key person-risk. It isn’t succession, although that’s real too (no disclosed succession plan, and the board just shrank to 4 members). The key risk is concentration and the related party: the same person sets the price at which the listed company sells services to his family’s clinics, controls 81.7% of votes, is paid $12M/year regardless of results, sold shares at the company’s cost, and the approval procedure for these transactions is declared deficient by management itself, two years running. In March 2026, the audit and compensation committees retroactively ratified the problematic arrangements — which means the control mechanism functions as recording a fait accompli, not as prior authorization.
Added: a shareholder class action filed 02.11.2026 in Delaware Chancery Court, alleging the company’s charter illegally restricts director removal to “for cause only” (10-K FY2025, Note 22). The proposed response in the May 2026 proxy is to eliminate the “for cause only” standard — i.e. resolving an illegal protection by removing the protection, plus an opt-out from DGCL §203 and officer exculpation. The direction is consistent: consolidating control, not strengthening minority protections.
Accounting red flags (accruals, dilution, one-offs, accounting policy changes)
1. Material internal control weaknesses, unremediated for two years (severity: MAXIMUM). Financial reporting controls were declared ineffective as of 12.31.2025, same as at 12.31.2024 (Item 9A). The two specific deficiencies are: (i) the absence of an operating control ensuring identification and prior approval of related-party transactions “including those involving close family members of management”; (ii) the absence of a procedure to bring exceptional executive compensation to the compensation committee before execution. The concrete manifestations in 2025: compensation of ¥240,000,000 ($1,604,028) paid to Yoshiko Aikawa, the CEO’s mother, as CEO of subsidiaries, not timely identified; and a bonus to the CFO not formally approved on time. Management states the payments had valid business purposes and were ratified in March 2026.
This is the most serious finding in the report, not for the size of the amounts ($1.6M doesn’t move the valuation), but because exactly the control mechanism that should protect the 18.3% minorities from the 81.7% is the one declared non-functional, by management, two years running.
2. Gains from related-party transactions bypassing the income statement (severity: HIGH). Three “deemed contributions” were credited directly to additional paid-in capital:
- FY2024: $1,473,571 on the disposal of subsidiaries to CEO entities (SBC Kijimadaira Resort, Skynet Academy, December 2024).
- FY2025: $10,353,587 — the markup on the sale of a plane to General Incorporated Association SBC, a CEO-controlled entity, via a memorandum in June 2025 to an 08.18.2023 contract (Note 20). The proceeds appear as a financing activity inflow, alongside $7,478,783 of “proceeds from disposal of property and equipment” in investing.
- H1 2026: $3,648,425 on deconsolidating the VIE (SBC Irvine MC).
Total $15.48 million over 2.5 years, i.e. 20.4% of total additional paid-in capital of $75.7M as of 06.30.2026. None went through the income statement, so none appears in EPS, adjusted EBITDA, or any multiples comparison. From a valuation perspective these are real cash inflows (good); from an information-quality perspective they’re transactions with the controlling party, at prices with no disclosed independent valuation, that increase equity without being tested through the income statement. O’Glove would highlight that a plane’s price marked up $10.35M two years after the contract has no market mechanism behind it. The tax cost was real: additional tax in Q2 2025 was ~$5.7M above a normal 40% rate.
3. Distorted basis of comparison in Q2 2026 communication (severity: MEDIUM). See the variations chapter: “+335% net income” is the product of an 82% tax rate in the base quarter, caused by red flag #2. “Growth reaccelerates” covers a core segment that grew 4.74% at the operating-income level. Not an accounting error — the 10-Q explains it all correctly — but a presentation choice that exploits a distortion created by a transaction with the company’s own related party.
4. Large positive accruals in 2024–2025 (severity: MEDIUM, improving). The Sloan ratio was +9.95% (2024) and +8.16% (2025) — thresholds at which the literature documents subsequent underperformance. LTM turned to −3.66%. The diagnosis (the affiliated-party float drain) is in the previous chapter and is closed. I keep it as a historical, not prospective, red flag.
5. Non-recurring gains in FY2025 profit (severity: MEDIUM). Of the $82,065,630 of pretax profit for fiscal 2025: $8,746,138 gain from the redemption of life insurance policies (10.7% of gross profit), plus a gain on the sale of the Irvine, California clinic’s land, included within $5,113,637 of “other income,” plus $2,002,789 FX gain, plus $815,328 gain on the previously held stake in Waqoo. Combined, over $14 million, 17% of gross profit, is non-recurring. “Clean” net profit for fiscal 2025 is ~$42M, not $51M — and normalized EPS ~$0.41, not $0.50. I used this normalization in the owner-earnings calculation.
6. Dilution: NOT a red flag. Shares fell from 103,020,816 (12.31.2024) to 102,576,943 (06.30.2026). Stock compensation was $13,022,692 in FY2024 (non-recurring warrants to the listing provider), $0 in FY2025 and $7,854 in H1 2026 — under 0.01% of revenue. The 2024 plan has 15 million shares available and none issued. This is one of the few criteria the company passes perfectly. The latent risk remains: 15 million authorized units = 14.6% potential dilution, at the discretion of a controlled board.
7. Three-month consolidation lag and the Waqoo segment (severity: MEDIUM — to watch). Waqoo and Aesthetic Healthcare Holdings are consolidated with a three-month reporting lag, meaning SBC’s “Q2 2026” contains Waqoo’s “Q1 2026.” The practice is permitted, but reduces comparability and delays the emergence of problems. More important: the Waqoo segment reports $3,756,172 operating income on $4,514,508 of revenue = 83% operating margin. I have no satisfactory explanation for an 83% margin at a direct-to-consumer cosmetics sales business; it could be an artifact of segment allocation or acquisition accounting (intangibles of $25.9M amortized over 11–16 years should weigh ~$0.47M/quarter). I flag this as an unexplained item, to clarify at Q3 2026. Waqoo was also excluded from the 2025 internal control assessment.
8. Zero provision on receivables from declining customers (severity: LOW-MEDIUM). See the O’Glove chapter, section 1.
9. A guarantee on the CEO’s personal debt (severity: LOW in amount, HIGH as a signal). A subsidiary guarantees $228,210 of the CEO’s personal debt as of 12.31.2025 (from $262,095 at 12.31.2024), with no provision (Note 22). The amount is immaterial; the fact that a listed company guarantees the personal debt of a man with a ~$307M stake at the current price isn’t.
10. Ongoing governance litigation (severity: MEDIUM). A shareholder class action in Delaware Chancery, 02.11.2026, over a DGCL §141(k) violation. The company can’t estimate a loss range. The research brief found no litigation — an omission; I document it from the 10-K, Note 22.
Triangulated valuation (conservative DCF with explicit assumptions + earnings power value + 5-year historical multiples + Monte Carlo from step 5; a range, not a point)
The FCF bridge — construction and choice
The issuer is US-GAAP (10-K, not 20-F), so interest paid appears in operating activities: $160,583 for FY2025, $243,435 in H1 2026. It’s de minimis (0.1% of CFO) and isn’t subtracted again. There’s no interest in investing or financing. The gap between definitions is under 1% — the AFYA case doesn’t apply here. What must be corrected, instead, is something else: the automated bridge doesn’t see advances for fixed assets (presented separately from capex in investing), which are real economic capex.
| Fiscal year | CFO | − PP&E capex | − PP&E advances | − Lease principal | = FCFE |
|---|---|---|---|---|---|
| FY2023 | $50,670,322 | $8,500,000* | n/a | — | ~$40.5M (including $1.7M intangibles) |
| FY2024 | $20,582,933 | $2,564,643 | $843,740 | $0 | $17,174,550 |
| FY2025 | $24,668,496 | $1,401,012 | $968,848 | $331,365 | $21,967,271 |
| H1 2025 | −$6,411,168 | $560,431 | $705,351 | $278,097 | −$7,955,047 |
| H1 2026 | $31,741,248 | $209,391 | $804,423 | $71,942 | $30,655,492 |
| LTM Jun. 2026 | $62,820,912 | $1,049,972 | $1,067,920 | $125,210 | $60,577,810 |
* FY2023 capex from the data pack’s XBRL tags; the FY2023 10-K isn’t in the downloaded set (the company was still a SPAC during that period from a consolidated-filing perspective).
Gross LTM FCFE = $60.6M. I don’t use it, for three verifiable reasons: (1) it contains $11,194,897 of tax deferral (the increase in the “income tax payable” balance in H1 2026), which will be paid in H2; (2) it contains the tail end of the related-party float drain, a non-recurring effect; (3) the gross three-year average, $26.6M, is depressed by the exact same effect in the opposite direction, so it isn’t representative either.
Normalization, step by step:
- LTM pretax profit (Jun. 2025 – Jun. 2026): $82,065,630 − $44,991,768 + $38,411,297 = $75,485,159
- minus the gain on selling the Irvine land (non-recurring, Q4 2025): −$3.1M (estimated from FY2025’s $5.11M of “other income,” of which $3.07M explicitly from related parties in 2024 doesn’t repeat; the exact figure isn’t isolated in the filing — flagged as an estimate)
- minus net FX gain LTM ($2,002,789 − (−$1,828,246) + $1,893,937): −$5.72M (non-recurring, non-operational)
- = normalized pretax profit ≈ $66.7M
- minus tax at the effective rate of 40.0% (the Q2 2026 rate, consistent with 37.8% for FY2025): −$26.7M
- minus minority interests (Waqoo, ~$1.0M annualized): −$1.0M
- = $39.0M
Cross-check via another route: LTM EBIT $65.4M + amortization $3.2M − total capex $2.1M − actual LTM cash tax paid ($37,190,188 − $19,637,454 + $8,660,822 = $26.2M) = $40.3M. The two routes converge to $39–40M.
External verification (mandatory). The 5-analyst consensus for 2026 is EPS $0.45 (Yahoo Finance, post-Q2), i.e. ~$46.2M net profit; the revenue consensus is $180.3M. My figure, $39.0M, is 15.6% below consensus — more conservative, not detached from reality. Implied FCFE yield: $39.0M / $375.4M = 10.4%. Brokers publishing targets of $7.00–9.00 are implicitly working with a yield near 12–13% on unadjusted figures. I’m not outside the published band.
Stated definition used in all DCF models below: normalized FCFE = CFO − PP&E capex − advances for fixed assets − lease principal, cleaned of non-recurring gains and FX. Value: $39.0 million. (The corresponding marker line is at the end of the report.)
Net debt used in the models
As of 06.30.2026: cash $184,311,213 + short-term investments $307,922 = $184.62M; interest-bearing debt $11,816,235 + $25,938,581 + finance lease $176,707 = $37.93M. Book net cash = $146.69M.
In the models I use $130.0M, i.e. a $16.7M discount, for: operating lease liabilities ($10,301,555 as of 06.30.2026, a real economic liability), the $5.0M commitment to OT Midco due December 2026, and the portion of consolidated cash belonging to Waqoo minority interests. The Monte Carlo script reports a balance-sheet net debt of −$112.5M (calculated from yfinance, on a different basis); my figure of −$130.0M sits between the two and is more conservative than the actual balance sheet. I don’t subtract debt separately in the DCF, because the flow used is FCFE — interest is already in CFO. Net cash is added to the present value of the flows.
Model 1 — Bear DCF: the fee compression repeats
Assumptions: OE $28.0M (a reversion to a reduced fee structure, as in April 2025, plus the yen at ¥170+), g1 1.0%/yr, r 14.0% (a governance and liquidity premium), gt 1.0%, net cash $130.0M.
Present value of years 1–5 flows: $98.8M · discounted terminal value: $118.7M · enterprise value $217.5M + $130.0M = $347.5M → $3.39/share → MOS −7.4%
Model 2 — Base DCF, Monte Carlo (step 5)
Central assumptions: OE $39.0M · g1 6.00% · r 12.00% · gt 2.00% · net debt −$130.0M · 102.577M shares · FX 1.0. Justification for deviations from the mechanical Excel base: (a) OE isn’t derived from margin × revenue, but from the bridge above, cleaned of $8.7M of policy-redemption gains, the gain on the Irvine land, and $5.7M of FX gains — exactly the AFYA error the mechanical base produces; (b) g1 of 6% is below the network’s structural growth (locations +13.4%, visits +10%, price/visit +9%) and below the core segment’s constant-currency growth of 14% in Q2, precisely because the fee is discretionary and has already been cut once; (c) r of 12% includes a premium above the Japanese cost of equity for revenue concentration and governance; (d) gt of 2% is capped by Japanese demography.
| Percentile | P5 | P10 | P25 | P50 | P75 | P90 | P95 |
|---|---|---|---|---|---|---|---|
| MOS (%) | +11.7 | +20.9 | +38.7 | +63.2 | +93.0 | +127.0 | +150.8 |
Median intrinsic value: $5.97. Probability the stock is undervalued: 98.5%. Probability of a margin of safety above 30%: 83.1%. Dispersions: OE ±25% lognormal, g1 σ=3.0%, r σ=1.0%, gt σ=0.50%. 20,000 scenarios.
The interpretation, which matters more than the figures. The range is extremely wide — from +11.7% to +150.8% between P5 and P95, a ratio of 13:1. What widens it isn’t OE (±25% alone would produce a band of only ~±35% on MOS), but the r × gt combination in terminal value: at r=12% and gt=2%, terminal value represents ~62% of total value, and a 1-point deviation in r or 0.5 in gt moves that component by 15–20%. Effectively, more than half of the median intrinsic value depends on what happens after 2031 at a company whose revenue swung ±16% over two consecutive years at a single person’s unilateral decision.
P(undervalued) = 98.5% must be read with an essential caveat: the simulation measures uncertainty about the company’s cash flows, not the probability those flows reach the minority shareholder. The model has no parameter for “the controlling owner diverts value.” The four models below do have one. For position sizing, the 98.5% signal doesn’t justify a large position: it justifies a small position in an asset with favorable asymmetry, where the dominant risk isn’t the one modeled.
Model 3 — Earnings Power Value (Greenwald)
The method ignores growth and capitalizes current, sustainable earnings power. Normalized EBIT: the average of the four recent periods ($70,660,066 · $70,303,710 · $67,486,398 · $65.4M LTM) is $68.5M; I use $66.0M, closer to LTM, because FY2023–24 were realized on a then-different fee structure since revised. Tax 40% → NOPAT $39.6M. Maintenance capex ($2.1M LTM) is below amortization ($3.2M), so no further adjustment is needed — the business model is genuinely capital-light.
Operating EPV = $39.6M / 0.12 = $330.0M · plus net cash $146.7M · minus the OT Midco commitment $5.0M = $471.7M → $4.60/share → MOS +25.7%
Model 4 — Governance-adjusted DCF (minority perspective)
Models 1–3 implicitly assume the flow belongs to all shareholders proportionally. The verified history says otherwise: over the last 2.5 years, $15.5M has passed through equity from related-party transactions, $26.5M has gone out as the CEO’s salary, $1.3–1.4M was paid for his share sale, $0 was distributed to shareholders. The model applies two adjustments: flow available to the minority = $35.0M (a permanent ~10% loss through escalating compensation and transfer prices, consistent with the observed trend) and cash valued at 50% ($73.3M instead of $146.7M) — because a balance that has produced neither dividend, nor buyback, nor significant interest in two years doesn’t deserve to be counted at par.
Assumptions: OE $35.0M, g1 4.0%, r 13.0%, gt 1.5%, valued cash $73.3M. Discounted flows years 1–5: $137.4M · discounted terminal value: $204.0M · total $341.4M + $73.3M = $414.7M → $4.04/share → MOS +10.5%
Model 5 — Historical multiples (limitation: only 23 months of relevant history)
Here, what’s missing must be stated explicitly. SBC doesn’t have five years of multiples. The quote before 09.17.2024 is SPAC Pono Capital Two’s, which traded between $9.96 and $10.91 from September 2022 through December 2023 — a trust price, not a business price. Usable history starts at the merger and includes a post-SPAC bubble (a high of $16.00 on 08.12.2024) that isn’t a valuation reference.
Post-normalization trading band (January 2025 – August 2026): low $2.87 (05.18.2026), high $5.65 (Q3 2025), quarterly-range median ~$4.10. On stable annual EPS of ~$0.48–0.50, that corresponds to an average trading P/E of ~8.5×. Applied to the LTM EPS of $0.478: $4.06/share → MOS +10.9%.
As an enterprise-multiple cross-check: current EV $228.7M / LTM EBIT $65.4M = 3.5×. Japanese medical-services and MSO comparables generally trade at 6–9× EV/EBIT; at 6× the result would be ~$5.25/share (+43%). I don’t use this figure in the triangulation, since I couldn’t verify a comparables set at a primary source in this session — I mention it as an order of magnitude for the relative discount, not as a model.
Summary: the range, not the point
| # | Model | Value/share | MOS vs $3.66 |
|---|---|---|---|
| 1 | Bear FCFE DCF (repeated fee compression) | $3.39 | −7% |
| 2 | Base FCFE DCF — Monte Carlo, median | $5.97 | +63% |
| 3 | Greenwald EPV (no growth) | $4.60 | +26% |
| 4 | Governance-adjusted DCF (minority) | $4.04 | +11% |
| 5 | Own historical multiples (23 months) | $4.06 | +11% |
Value range: $3.39 – $5.97. Defensible central point: $4.05 – $4.40. Median MOS of the five models: +11%.
The difference between model 2 (+63%) and the other four (−7% … +26%) isn’t a methodological inconsistency; it’s the thesis itself. Model 2 values the business. Models 1, 4 and 5 value the minority stake in that business. The ~$1.9-per-share difference between the Monte Carlo median and the central point of the others is, numerically, the price of governance at SBC — approximately a 32% permanent discount. Whoever believes the discount closes (a buyback actually executed, a dividend, normalized affiliate transactions) has a double-bagger. Whoever believes it doesn’t close has a fairly valued stock with a small bonus.
Pre-mortem (why the thesis could be wrong — 3 concrete scenarios)
Scenario 1 — The $15M fee increase doesn’t materialize, because the MCs can’t afford it (estimated probability: 30–35%).
The mechanism: on 08.13.2026, the company announced fee revisions for call-center services to five affiliated medical corporations and for expanded support at Rize Clinic and Gorilla Clinic, with an estimated annualized impact of ~$15 million at ¥158.1/$. This is an 8.6% increase on the revenue base and, being a service fee with near-zero marginal cost, would flow almost entirely to operating income — pushing normalized EBIT from $65M toward $78M and owner earnings from $39M toward $48M. The consensus bull case ($8.25–8.67) essentially assumes this happens and repeats.
How it fails: exactly as happened in reverse in April 2025. Then, management reduced fees “based on each clinic’s size, scale and performance,” and franchise revenue fell 24.7% and management revenue 44.2%. The undisclosed but most plausible reason is that the MCs could no longer afford the flat fee at small, new clinics. If the 2026 increase hits the same constraint, it will show up first in receivables — the affiliate balance will stay above $45M at 09.30.2026 instead of falling toward $30M — and only then in revenue, with a 2–3 quarter lag. The alarm signal is in the balance sheet, not the income statement.
Impact on the thesis: OE falls toward $30–33M, the bear model becomes the central scenario, value drops to $3.40–3.70 and the stock is fairly valued at today’s price. It’s not a major capital loss, it’s a dead thesis — the opportunity cost of money locked in an asset that doesn’t move.
Scenario 2 — The $184M of cash never reaches the minority shareholder; it turns into mediocre acquisitions (estimated probability: 40–45%, the most likely of the three).
The mechanism: 39% of market cap is cash. Almost all the apparent cheapness comes from there — without cash, EV/EBIT would be 3.5× instead of looking absurd; with cash valued at zero, the target price falls by ~$1.43/share. Models 2 and 3’s implicit assumption is that cash is worth 100 cents on the dollar.
How it fails: not through fraud, but through being used exactly as it has been so far. The 2025–H1 2026 path is already this: $20M into 18.2% of a US med-spa operator that immediately produced a $568,652 loss and requires another $5M in December 2026; $19.4M for Waqoo, partly bought from the CEO through an off-market transaction whose price isn’t disclosed separately; $14.2M for MB career lounge; $0 dividend; $0 buybacks in 2026 out of a $20M program. The expansion narrative — the “Longevity market,” ASEAN via Thailand, “SBC AI Lab” — describes exactly the kind of program that consumes cash into illiquid, uncontrolled positions in markets the company doesn’t know.
Impact: intrinsic value theoretically stays the same, but is never realized. The stock stays at 7–9× profit indefinitely, like many Japanese companies with large cash and a dominant owner. Investor return = profit growth, ~5–8%/yr, with no multiple re-rating. You don’t lose money; you don’t earn the discount. This is the most likely and least dramatic risk.
Scenario 3 — The related-party structure breaks, through regulation or family conflict (estimated probability: 8–12%, but with severe loss).
The mechanism: the entire company exists because Japanese medical-care law prohibits corporate ownership of clinics, and SBC routes around the constraint via franchise and service contracts with medical corporations whose “members” are the CEO’s relatives. SBC holds “mochibun” in six MCs, but explicitly with no voting rights. The CEO himself ceased being a member in July 2023 — likely precisely to clean up the structure before the listing.
How it fails, on two paths. (a) Regulatory: Japanese authorities have periodically examined MSO models where the management company captures most of the clinic’s economics. A reinterpretation limiting management-fee levels would hit 91.5% of revenue directly. I found no concrete legislative initiative in this session — flagged as an unverified structural risk, not a predictable event. (b) Family: the MCs’ “members” are the CEO’s relatives, not the CEO. A succession, a divorce, an inheritance dispute, or a simple change of interest at a single MC could detach 20–24% of revenue with SBC having no voting right to oppose it. SBC can’t consolidate the MCs precisely because it has no control — and what you can’t consolidate, you can’t retain either.
Impact: losing the largest MC (Shobikai, 23.6% of revenue) would cut ~$41M of revenue and, at their high margin, perhaps $25–30M of EBIT — nearly half of earnings power. Value would fall below $2.50/share, i.e. a loss of over 30% from the current price, even with cash intact. The probability is small over a one-year horizon, but not negligible over five, and the investor has no monitoring tool: family relationships aren’t reported quarterly.
What would invalidate the thesis fastest (the single indicator to watch): related-party receivables as of 09.30.2026. Under $35M = the pattern works, the fee increase is collected, the thesis holds. Above $45M = the MCs aren’t paying, and scenario 1 is underway.
Verdict compared to the tracker’s GBL score (convergence/divergence and why)
Mandatory preliminary observation: the tracker has no GBL score for SBC. Row 125 of the “Sumar” sheet in Pregatire_investitii_21.xlsx contains only auto-populated yfinance data as of 08.17.2026 — price $3.66, NASDAQ exchange, RSI 14 = 72.2, position ↑EMA50 ↑EMA200, PEG (ttm) 2.62, “Disc. vs Consensus %” = +132.24%, status ACTIVE, next report 11.12.2026. The Score %, Verdict, Graham, Buffett, Lynch, Disc.GN%, MoS DCF %, F-Score, Penalty, Altman Z, governance flag and triangulated DCF columns are empty. SBC doesn’t appear on the “Analiza” sheet either. Therefore I can’t compare against an existing score; I calculate one myself, per the GBL_criterii_30.md grid (v21, Graham ×1, Buffett ×2, Lynch ×3 weighting, denominator 60), on this report’s normalized figures.
GRAHAM (×1) — 7.0 raw points
| No. | Criterion | Verified value | Score |
|---|---|---|---|
| G1 | Current P/E | 7.66× on LTM EPS $0.478 (9.0× on normalized EPS $0.41) — well below the 30th percentile of medical services | 1 |
| G2 | P/B | $3.66 / BVPS $2.572 = 1.42× < 1.5× | 1 |
| G3 | P/E × P/B | 7.66 × 1.42 = 10.9 < 22.5 | 1 |
| G4 | Dividend paid | Never | 0 |
| G5 | Dividend yield | 0% | 0 |
| G6 | Debt (D/E) | Interest-bearing debt $37.9M / equity $263.8M = 0.14×; net cash $146.7M | 1 |
| G7 | ROIC vs. cost of capital | LTM EBIT $65.4M × 0.60 / (equity $279.9M − net cash $146.7M) = 29.5% > 20% | 1 |
| G8 | 5-year EPS growth | Only 3 years of consolidated history: $0.42 (2023) → $0.48 (2024) → $0.50 (2025), LTM $0.478. Growing, but no 5-year series | 0.5 |
| G9 | Earnings consistency | Positive profit in all 3 documented years; can’t verify 5 | 0.5 |
| G10 | Price vs. Graham Number | GN = √(22.5 × 0.478 × 2.572) = $5.26; price $3.66 = 30.4% below GN | 1 |
BUFFETT (×2) — 4.5 raw points → 9.0 weighted
| No. | Criterion | Rationale | Score |
|---|---|---|---|
| B11 | Durable moat | Real procurement scale + a 26-year-old brand, but the “contract” isn’t a moat, it’s dependence on a single family | 0.5 |
| B12 | 10+ year durability | Regulatory risk of the MSO structure + Japanese demography + Korean competition | 0.5 |
| B13 | Consistent ROE | 27.6% (2024) · 23.0% (2025) · 19.1% (LTM), no leverage | 1 |
| B14 | Superior margins | Operating margin 37–39% vs. medical-services sector 5–10% | 1 |
| B15 | Management quality | Material control weaknesses two years running, exactly around related parties; a mother’s salary unidentified; the company pays for the CEO’s share sale | 0 |
| B16 | Skin in the game | Owns 81.7% but sold 3.1M shares in April 2026; the grid penalizes the sale | 0.5 |
| B17 | Predictability | Revenue −15.5% then +1.7%, at the unilateral decision of the same person | 0 |
| B18 | Efficient capital reinvestment | Bought back at $4.88, stopped at $3.00; $20M in OrangeTwist already at a loss; $184M cash unused | 0 |
| B19 | Pricing power | Price/visit at clinics +9% (real), but SBC cut its own fee 25% in 2025 | 0.5 |
| B20 | DCF margin of safety | Triangulation median +11% < the 20% threshold; only the mechanical model gives +63% | 0.5 |
LYNCH (×3) — 5.0 raw points → 15.0 weighted
| No. | Criterion | Rationale | Score |
|---|---|---|---|
| L21 | PEG | 2.62 (tracker, auto) > 1 | 0 |
| L22 | 3-year profit growth | EPS $0.42 → $0.50 = +9.1%/yr, and LTM ($0.478) is below FY2025; normalized, nearly flat | 0 |
| L23 | Lynch category | A declared turnaround (“restructuring complete”) after a 15.5% revenue drop | 0 |
| L24 | Ignored sector | Japanese aesthetic clinics listed in the US — a niche sector, but with 5–7 active analysts | 0.5 |
| L25 | Simple story | The model explains itself in 2 minutes; the network of 24 disclosed related parties doesn’t | 0.5 |
| L26 | Insider ownership >10% and buying | Owns 81.7% but is selling, and the company isn’t buying | 0 |
| L27 | Solid balance sheet | Net cash $146.7M, current ratio 3.31× | 1 |
| L28 | Growing market share | 287 locations (+34 y/y), visits +10%, ~31% of mid/large-group locations | 1 |
| L29 | Low coverage | 5–7 analysts, all small boutiques; Maxim Group, which publishes estimates, was the bookrunner of the CEO’s share sale | 1 |
| L30 | Dilution through stock compensation | $0 in FY2025, $7,854 in H1 2026 — under 0.01% of revenue | 1 |
Final score and verdict
Weighted sum = Graham 7.0 ×1 + Buffett 4.5 ×2 + Lynch 5.0 ×3
= 7.0 + 9.0 + 15.0 = 31.0
Volatility penalty: HV30 = 57.5% (band 45–70%) −0.5
Other penalties (negative FCF 3yr / dilution >5% / Altman Z distress): none
Final weighted sum = 30.5
Score% = 30.5 / 60 = 50.8% Equivalent /30 = 15.25GBL verdict: MONITOR (threshold ≥48%; BUY would require ≥72%).
Convergence and divergence
My deep verdict and the calculated GBL score converge, but by different roads — and that’s worth explaining, because it’s informative.
The GBL score reaches 50.8% because Graham scores nearly maximum (7/10) — on every mechanical safety and cheapness test the company passes: P/E, P/B, P/E×P/B, debt, ROIC, Graham Number. And Lynch scores only 5/10, with the heaviest weight (×3) — because Lynch’s grid is the one asking whether insiders are buying (L26 = 0), whether profit is growing (L22 = 0), and whether the story is growth or turnaround (L23 = 0). The v21 weighting (Lynch ×3) makes the growth and insider-behavior failures cost 9 weighted points out of 60 — and that alone pulls the score out of the “BUY” zone into the “MONITOR” zone.
My deep analysis reaches the same conclusion through a more explicit mechanism: the mechanical model (Monte Carlo, +63% MOS, 98.5% undervaluation probability) and the governance-adjusted models (+11% median) differ by ~32 percentage points, and that difference is exactly what the GBL grid captures via B15 = 0, B18 = 0, B16 = 0.5 and L26 = 0. In other words: the four management-behavior criteria, worth 3.5 weighted points lost, are the score’s translation of the governance discount I quantified in the valuation.
The significant divergence is against the market consensus, not the tracker. The “Disc. vs Consensus %” column in the tracker shows +132.24% — i.e. an average analyst target around $8.50 (consistent with the $8.25–8.67 from the brief, from 7 analysts, with 83% buy recommendations). Not even my most optimistic model (Monte Carlo P90 = +127%, i.e. $8.31) reaches there except in the top decile of scenarios. The consensus target implicitly assumes both the full materialization of the $15M fee increase and cash valued at par and the absence of any governance discount — three assumptions that can’t all be true simultaneously at a company whose controller just sold shares at $3.25 at shareholders’ expense. Also noting a relevant conflict of interest: Maxim Group, a source of coverage on the stock, was the sole bookrunner of the CEO’s April 2026 secondary offering, and Roth Capital was co-manager and collected a consulting fee from the company. The consensus on SBC isn’t an independent consensus.
Recommended positioning. A small speculative buy (≤1–1.5% of portfolio), justified by real asymmetry: at $3.66 you pay ~5.9× normalized owner earnings ex-cash for a business with 29.5% ROIC and net cash equal to 39% of market cap, and even the most pessimistic model gives only −7%. The dominant risk isn’t capital loss, it’s capital immobilization. Liquidity, in turn, imposes discipline: at ~$452,000 average daily volume, any position over $75,000 is hard to liquidate quickly.
Re-evaluation triggers, in order of importance: (1) the first buyback actually executed from the $20M program — would validate model 2 and justify increasing the position; (2) related-party receivables at 09.30.2026 under $35M — confirms the fee increase is being collected; (3) the declared remediation of material control weaknesses in the FY2026 10-K (March 2027); (4) any inaugural dividend. Exit triggers: a new downward “fee structure revision”; affiliate receivables above $45M at 09.30.2026; an acquisition above $30M paid cash to a CEO-related entity.
All financial figures come from the cited SEC filings (10-K FY2025 of 03.27.2026, 10-Q Q1 and Q2 2026, 10-K FY2024, DEF 14A of 05.28.2026, 424B7 of 04.17.2026, 8-K/EX-99.1 of 08.13.2026), the XBRL data pack generated 08.17.2026, and the simulation file mc-SBC-20260817.json. Figures flagged as estimated (the Irvine land gain, FY2023 capex, the 31% market share, the size of the Japanese aesthetics market) couldn’t be isolated at a primary source and are flagged as such in the text. The reference price is the $3.66 close from 08.14.2026.
DEEPFCF;SBC;FCFE normalizat (CFO minus capex PP&E si avansuri pentru imobilizari minus principal de leasing, curatat de castiguri nerecurente si de curs valutar);39000000;dobanda e in operating la un emitent US-GAAP si sub 0,1% din CFO deci nu se mai scade, iar FCFE brut LTM de 60,6M contine 11,2M amanare de impozit si media pe 3 ani de 26,6M e deprimata de drenarea float-ului de la parti afiliate, acum incheiata
DEEPDCF;SBC;-7;11;63;5 DEEPDCFM;SBC;1;DCF FCFE bear;-7 DEEPDCFM;SBC;2;DCF FCFE baza MC;63 DEEPDCFM;SBC;3;EPV Greenwald;26 DEEPDCFM;SBC;4;DCF ajustat guvernanta;11 DEEPDCFM;SBC;5;Multipli istorici proprii;11
DEEPMC;SBC;20.9;38.7;63.2;93.0;127.0;98.5;20000