Natural Food International Holding Limited 1837

Deep-value analysis — 1837.HK · Natural Food International Holding Limited (五谷磨房食品國際控股有限公司)

Analysis date: September 4, 2026 · Reference price: … (last full close: … on 09/03/2026) · Exchange: HKEX Main Board, code 1837, listed 12/12/2018 Reporting currency: RMB (CNY) · Trading currency: HKD · Rate used: … = RMB0.8558 (market rate 09/04/2026, yfinance HKDCNY=X; verified via USDCNY 6.7001 / USDHKD 7.84079 → 0.85451) Shares issued: 2,188,564,000 at 08/31/2026 (Monthly Return, 09/04/2026) · Market cap: … mn = RMB2,687.7 mn · Net cash: RMB1,256.8 mn · Enterprise value: RMB1,430.9 mn

New primary source for this analysis: INTERIM RESULTS FOR THE SIX MONTHS ENDED 30 JUNE 2026, published on HKEXnews on 08/27/2026 at 22:43, reviewed by Ernst & Young per HKSRE 2410 (/listedco/listconews/sehk/2026/0827/2026082702341.pdf). Rest of the file: 2025 Annual Report (04/24/2026), FY2025 annual results announcement (03/31/2026), Business Updates (03/18/2026), Positive Profit Alert (08/17/2026), three announcements of the controlling shareholder increasing its stake (06/25, 07/02, 07/23/2026), Next Day Disclosure Return (07/21/2026), Monthly Return (09/04/2026), plus the half-year series from stockanalysis.com — which reconciles exactly with filings (CFO H1+H2 2024 = 271.45 vs 271.452 reported; 2025 = 236.00 vs 235.990; capex 24.63 vs 24.634 and 34.19 vs …).

This analysis continues the report from 08/23/2026 (2026-08-23-deep-1837.md), written four days before the interim results were published. Material that hasn't changed (the five-year history, ownership structure, audited FY2025 notes) is not repeated here; only what the 06/30/2026 test confirmed or refuted is revisited, and the valuation is fully redone from a price 14.9% lower.

Corrections to the received research brief (research-1837-20260904.md), in order of severity. (1) The brief says "H1 2026: revenue RMB1.45 bn, … YoY." False. H1 2026 revenue grew …; … is net income growth. The brief pulled yfinance's earningsGrowth field and labeled it revenue growth — exactly the error that would have anchored the entire valuation to a growth rate double reality. (2) "H1 2026 net income: unavailable" — it had been published for eight days: RMB155.898 mn. (3) "PepsiCo cancelled the acquisition of the 25.8% stake" — false; PepsiCo, Inc. holds 566,506,000 shares = 25.88%, and Ms. TSE Cheung On Anne, CEO APAC Foods at PepsiCo, appears as a non-executive director in the board list of the 08/27/2026 announcement. (4) The target prices "…" (article from 03/31/2022) and "…–0.87" (unattributed) are wrong or expired: current consensus, across two analysts, is … (stockanalysis.com / yfinance targetMeanPrice 2.3719). (5) "DCF fair value … vs price …" is in Canadian dollars for a Hong Kong stock — a localized page artifact, unusable. (6) The flagged leadership change (NG Benjamin Jin-Ping, undated) isn't the relevant one: the current board is GUI Changqing and ZHANG Zejun (executives), TSE Cheung On Anne and WANG Duo (non-executives), ZHANG Senquan, OUYANG Liangyi and CHOU Hung Chi (INEDs); the only change disclosed since the annual report is that Mr. WANG Duo became vice-president at Persagy Technology Co., Ltd. as of May 2026. (7) The brief's observation that rttnews labels FY2025 as FY2024 is correct — the figures are FY2025.

---

Executive summary

What happened in the 12 days since the prior analysis. On August 27, after close, the company published its half-year results. The figures were good: revenue RMB1,449.7 mn (…), net income RMB155.9 mn (…), net margin 10.75% versus 9.41%, operating cash flow RMB190.8 mn (…). Profit landed at the midpoint of the range guided on August 17 (RMB150-160 mn). The next day the stock opened at … hit … and closed at … (…) on 10.4 million shares — roughly seven times the month's average volume. Selling continued for four sessions: 1.550 → 1.470 → 1.450 → 1.430. From the pre-results close (…) to today: …, on roughly 26 million shares traded, i.e. over 4% of the actual 584-million-share free float.

Why it fell, if the figures were good — the figure nobody published. On March 18 the company had announced January-February 2026 sales of RMB655-670 mn, … to …. Subtracting that period from the full half-year:

Period	2026 (RMB mn)	2025 (RMB mn)	Growth
January-February (announced 03/18/2026)	655 – 670	442.0 – 442.8 (derived)	… … …
March-June (residual)	779.7 – 794.7	691.9 – 692.7	… … …
Full half-year (reported 08/27/2026)	1,449.7	1,134.7	…

March-June growth was roughly 13…, not 48%. The market had extrapolated the seasonal peak and, in month eight, received proof it had been exactly that — a seasonal peak. This is the only coherent explanation for the sell-off — profit was in line with guidance, the balance sheet cleaned up, there was no negative surprise in the income statement.

The thesis, rebuilt. At … you're buying a business that produces EBIT of RMB376.7 mn over the trailing twelve months for an enterprise value of RMB1,430.9 mn — 3.80x EV/EBIT, 3.45x EV/EBITDA, 0.50x EV/sales. Net cash of RMB1,256.8 mn represents 46.8% of market cap, i.e. … of the … price. Capital actually deployed in operations (average equity minus net cash) is RMB446.9 mn and produces NOPAT of RMB305.3 mn — a 68% ROIC. Trailing-twelve-month P/E is 8.48x versus the five-year year-end average P/E of 7.59x: an 11.7% premium, versus 53% at the date of the prior report. Dividend yield is 4.81%.

Estimated value (triangulation across five models, net cash treated explicitly and separately from flows):

Model	Key assumption	IV/share (HKD)	MOS vs …
DCF FCFE — bear	OE 160, g1 2%, r 13.5%, gt 1.5%, cash haircut 25%	1.232	…
5-year historical multiples	Average P/E 7.59x on 2026E profit of RMB343.0 mn	1.390	…
EPV Greenwald (no growth)	EBIT normalized to 3-year average margin (10.64%), r 11%	1.793	…
DCF FCFE — base	OE 232, g1 9%, r 11%, gt 2.5%, haircut 10%	2.673	…
DCF FCFE — bull	OE 265, g1 14%, r 10.5%, gt 2.5%, full cash	4.004	…

Median of the five: …. Range: … to …. Monte Carlo (20,000 scenarios on top of the base model): median MOS ** …**, P10 …, P90 …, probability of undervaluation …%.

Verdict: 🟢 BUY — a position of roughly 70% of the normal allocation, with a second tranche below … This is a change from "🟡 INTERESTING, 40… of allocation" on 08/23, and the change has three verifiable causes, not a change of mood: (1) the price fell 10.3% versus … while trailing-twelve-month profit rose 18.5% versus FY2025 — a 27% de-rating of the multiple in twelve days; (2) the explicit test written in the prior report was passed: DIO dropped from 53 to 43 days on the company's methodology (from … to 29.5 on annualized period-end balances), the finished goods built for New Year sold through, and receivables fell in absolute value by 10.9% while revenue grew 27.8%; (3) the model that assumes no growth at all — EPV — went from … to …, i.e. the market is no longer paying even for current earning power.

What keeps me under full allocation. Not the valuation, but three things the valuation doesn't resolve: trailing-twelve-month operating cash flow is … versus the prior period while net income grew 54.0% (positive accrual for the second consecutive half-year, … of average assets); underlying growth, stripped of seasonality, is 13…, not 28%; and the RMB1.26 bn of cash — half the thesis — is held in mainland China subsidiaries by a management that has never bought back a single share for cancellation in eight years of listing.

---