DN Agrar Group S.A. DN

DN Agrar Group S.A. (BVB: DN, AeRO Premium) — Deep-Value Report (REFRESH)

Analysis date: 2026-09-04 · Current price: RON … (09/03/2026, 17:54, primet.ro / tracker) · Market cap: RON 555,238,842 (~EUR 105.9 mn) · Shares: 159,094,224 · Regime: REFRESH over the baseline thesis from 2026-07-27 (2026-07-27-deep-DN.md)

New primary source: the H1 2026 interim report (published 08/24/2026, unaudited, RAS + parallel IFRS) — RAPORT-H1-2026-ENG-SITE.pdf, dn-agrar.eu — plus the 08/25/2026 conference call presentation and the audited FY2025/FY2024 consolidated statements (BS-si-IS-consolidat-31.12.2025.xlsx, BS-si-IS-consolidat-31.12.2024.xlsx). A BVB issuer, with no SEC/EDGAR filing — the automated data pack (data-pack-DN-20260904.md) is empty for balance sheet/P&L/CF because yfinance has no financial statements for DN.RO; all figures below are read directly from filings.

What is carried over from the 07/27/2026 thesis (not redone)

Carried over, without re-verification, because nothing in the new filing touches them: the business description and operating structure (9 wholly owned subsidiaries, >7,000 ha, an integrated livestock farm model with no consumer brand), the moat verdict (weak, based on scale, not differentiation — raw milk is a commodity, concentration on Lactalis/Albalact >70% of 2024 sales), the 2021-2025 capital allocation history (the Lacto Agrar / Apold / Calnic acquisitions, zero dividends, 100% credit financing), the structural sector context (a fragmented Romanian market, a ~EUR 700 mn dairy trade deficit, the APIA reference frozen at 2018) and pre-mortem scenarios 2 and 3 (CUT2 execution risk and geographically concentrated sanitary/climate risk) — remain valid as stated, updated only with the S1 2026 project status.

Re-derived (because the H1 2026 filing moved them): the entire quarterly-variance chapter, the entire earnings-quality analysis (a new chapter, absent from the old report), the CEO profile on the Thorndike grid (new chapter), the red flags, the FCF bridge and the entire valuation — including the DCF base, which shifts from an EBITDA projection aligned with company guidance to normalized owner earnings. This last change is a shift in opinion, not in fact, and is stated explicitly as such in the valuation chapter.

The price move (… → …, … in 38 days) is context, not the reason for the resumption. The reason is that the H1 2026 filing became available and the thesis's five falsifiers were unverifiable.

---

Executive summary

DN Agrar Group remains the EU's largest integrated cow-milk producer by scale (18,300 head, 41 mn liters delivered in H1 2026, … YoY), with operational execution that continues to be good: unit cost has clearly fallen (operating expenses … on milk volume … ⇒ cost per liter …), the herd is growing, and the four major projects — the processing plant at Straja, CUT2, doubling compost capacity, the vertical wheatgrass farm — are all on schedule or close to it.

What broke in H1 2026 and wasn't visible in July. The average milk price fell 30% YoY, and the effect was concentrated in Q2. Reconstructing the quarter (H1 minus Q1, consolidated RAS figures):

	Q2 2025	Q2 2026	Δ
Revenue	50,148,905	43,465,380	…
Operating profit	19,899,155	12,962,656	…
Operating subsidies	7,708,430	14,301,562	…
Operating profit EXCLUDING subsidies	+12,190,725	−1,338,906	reversal
Net income	13,112,877	4,164,044	…
Net margin (on revenue)	26.1%	9.6%	−16.5 pp

The second quarter of 2026 was, on Romanian accounting standards, operationally loss-making before subsidies. For the half-year, operating profit excluding subsidies fell from RON 24.37 mn to 4.57 mn (…), and subsidies (RON 21.69 mn, …) represent 82.6% of operating profit — versus 37.5% in H1 2025. The company itself correctly flags that part of the subsidy increase is a 2025-related supplement collected in Q2 2026 and should not be treated as recurring.

The thesis, restated: DN Agrar is a real, operationally well-run business, in a phase of aggressive investment, which at the bottom of the milk-price cycle does not cover its core-activity costs without subsidies and has not self-funded in any year since listing. The stock price already prices in full execution of the 2025-2030 plan (doubling EBITDA by 2030), not current earning power.

Estimated value. Five triangulated models give a range of … … …, with the median at … (Monte Carlo, 20,000 scenarios, median IV RON …/share). Four of the five models are negative and cluster between … and …; the sole positive model is the DCF built directly on management's plan. Asset anchors confirm the low zone: RAS equity RON 1.455/share, IFRS RON 1.686/share. Simulation probability of undervaluation: …%.

Verdict: AVOID at the current price (revised from MONITOR). Not because the business has broken — operating indicators are good — but because (a) TTM profit fell 18.7% since the prior analysis, so the stock has gotten more expensive without growing (RAS P/E from 11.8x to 14.27x), (b) current earning power, cleaned of the accounting recognition of capitalized own-raised heifers and of subsidies, is well below what the market is paying, and (c) the margin of safety has turned negative in four of five models. A reasonable re-entry point: below RON 2.20 (base MOS ≈ …) or confirmation, in Q3-Q4 2026, of a milk-price recovery accompanied by the first non-negative FCF year.

---