3393 — Wasion Holdings Limited

Deep-value analysis — Wasion Holdings Limited (威胜控股), HKEX 3393

Analysis date: 08/25/2026 · reference price … (close 08/25/2026) · market cap … ≈ RMB 14,676 mil · 1,038.99 mil shares (excl. treasury) · DIFFERENT currencies: reporting RMB, quoted HKD (rate used: … = 0.8571 CNY) · analysis, not investment advice.

Issuer identification warning. “3393” is an ambiguous code. On 08/24/2026, the triangulated-DCF agent mistakenly analyzed 3393.T = Startia Holdings (Tokyo, IT services) and wrote into the tracker a range of … … … belonging to a completely different company. The issuer requested here is 3393.HK — Wasion Holdings Limited, Cayman Islands, listed on the HKEX Main Board, a Chinese manufacturer of smart meters and network equipment, headquartered in Changsha, Hunan. There are no SEC filings (not a US issuer, no registered ADR), so this report’s primary dossier is the HKEX filings: the FY2025 annual results announcement (03/26/2026), the 2025 annual report (04/22/2026, 275 pages, auditor EY, unqualified opinion), the H1 2025 interim results (08/26/2025) and nine voluntary announcements from 2026. The final markers replace the wrong entries in the tracker.

The group structure matters more than any figure in this report. Wasion Holdings is, de facto, a three-legged holding company, of which one leg is already separately listed and a second is in the process of listing:

Consequence: 28.6% of the group’s 2025 profit (422.9 mil RMB of 1,481.1 mil) does NOT belong to Wasion’s shareholder, but to Willfar’s minorities (330.6 mil) and Wayon’s (92.2 mil). Any valuation comparing the consolidated flow to Wasion’s market cap overstates value by ~40%. This is the report’s central thread.

Executive summary

The thesis. Wasion is a real industrial business, genuinely improving operationally — revenue … in 2025 to RMB 10,073.5 mil, gross margin 35.7% (from 32.9% in 2022), profit attributable to the parent … to RMB 1,058.2 mil, ROE 18.1%, ROIC ~21%, almost no net bank debt. The founder holds 51.5% after the placement, and export orders grow 17%/year. But at … the market is paying for three overlapping stories — data centers, the Wayon spin-off, and “AI + energy” — not for the existing cash flow. The real flow to Wasion’s shareholder, after interest, leasing and the minority share, was RMB 462 mil in 2025 and RMB 396 mil on a three-year average, i.e. a yield of 2.7… on today’s market cap, against an apparently “cheap” P/E of 13.9x. The gap between these two figures — a 67% divergence between “CFO − capex” and the flow attributable to the parent — is the whole thesis.

What broke in the last three months. The stock hit … on 04/13/2026 and closed at … on 08/25/2026: … in four and a half months, of which … in the week of 08/17 alone. The trigger is verifiable in the filings: on 08/17/2026 Wasion published the half-year results of its subsidiary Willfar — H1 2026 revenue of RMB 1,066.7 mil, …, profit RMB 202.4 mil, …. Broken down by quarter (Willfar had already reported Q1 on 04/29/2026: revenue …, profit …), this implies a catastrophic Q2 2026 at Willfar: revenue RMB 534.9 mil, … YoY, profit RMB 54.1 mil, … YoY. Willfar is the group’s highest-margin segment (23.1% segment margin) and 37% of aggregate segment results. Wasion group’s half-year results are approved by the board on 08/27/2026 (two days away) and published by 08/31/2026 — meaning this report is being written exactly in the blind window before the one document that would resolve the uncertainty.

Estimated value. Five triangulated models, all calibrated to the flow attributable to the parent (FCFE), give a range from … (pessimistic DCF) to … (optimistic DCF), with a median of … (EPV Greenwald). The Monte Carlo simulation across 20,000 scenarios around the base case (OE RMB 500 mil, g1 8%, r 11%, gt 2.5%) gives a median intrinsic value of … and a probability of undervaluation of 4.5%. The only model close to the market price is sum-of-the-parts (Willfar at its own STAR Market quote + the rest at 10× profit): … — but that model imports the Shanghai exchange’s 22.5× trailing-profit valuation for a subsidiary whose profit just fell two-thirds in a quarter.

The verdict. OVERVALUED at the current price; a business to watch, not to buy now. Not because it’s a bad company — it isn’t — but because the price already embeds the re-rating from an average P/E of 7.3× (the 2021–2025 average) to 13.9× today, at a moment the group’s best-margin leg is contracting 34% quarter-on-quarter and profit-to-cash conversion for the holding’s shareholder is around 40%. The level at which the file becomes deep-value interesting again: below … (the simulation’s median intrinsic value), last reached in August 2025. A reasonable intermediate reassessment level: …–12 (EPV Greenwald), where the simulation’s probability of undervaluation rises toward 25….