Look at the chart of SOJA3 — shares of Boa Safra Sementes, listed on the B3 (the Brazilian stock exchange) — and one question jumps out: if Boa Safra is the market leader and moves more seed every season, why has the stock fallen roughly 67% from its 2023 high, from above R$18 to under R$6? The answer is blunt: selling more stopped translating into earning more. In 2025 the company posted record revenue of R$2.62 billion (up 42%) and a record volume of 215,000 big bags (up 34%), yet the profit attributable to controlling shareholders — the slice of the bottom line that actually belongs to a SOJA3 holder, once fund partners and affiliates are stripped out — collapsed from R$301.5 million in 2023 to R$20.0 million in 2025.
What broke was not demand; it was profitability. Adjusted EBITDA margin — what the business keeps before interest, taxes and depreciation, a clean read on operating efficiency — fell from 10% to 5.9%. And below the operating line, interest on debt exploded 254% (R$132 million in 2025) after the company issued roughly R$1 billion in CRAs (agribusiness receivable certificates, debt instruments a company sells to raise cash) at CDI plus 3% a year — CDI being Brazil's interbank benchmark rate — with the Selic (Brazil's benchmark interest rate) running high. The stock fell because earnings fell, and earnings fell far faster than revenue rose.
What Boa Safra actually does, from scratch
Before planting soybeans, a farmer has to buy the seed. Boa Safra is the largest independent seller of that seed in Brazil: of every ten bags of certified soybean seed sold in the country, one is theirs — a record 10% market share. The company does not farm for grain. It contracts around 320,000 hectares of partner growers to multiply the seed, processes that material in its own plants (capacity of 280,000 big bags a year) and distributes through 996 dealers across Brazil. Founded by the Colpo family in Formosa (Goiás state), it went public on the B3 in April 2021, on the Novo Mercado, the exchange's highest corporate-governance tier.
One feature of this business changes how you must read the numbers: the seasonality is brutal. Roughly 89% of revenue lands in the second half of the year, because seed is delivered close to planting time (September to December). The first half is inventory-building season — the company burns cash, net debt peaks, and quarterly EBITDA can even turn negative. That is normal.
Never annualize a Boa Safra first quarter. Net debt jumped from R$151 million in December 2025 to R$848 million in March 2026 — not a sign of deterioration, but the seasonal cycle of buying grain to build inventory. What matters is the full-year close and the comparison against the same quarter a year earlier.
Anatomy of the fall: three blows at once
The run of results tells the whole story. Revenue grows while profitability caves — and every line in the table below hides a cause.
| Period | Net revenue (R$ mn) | Gross margin | Adj. EBITDA margin | Controlling profit (R$ mn) | Big bags (000) |
|---|---|---|---|---|---|
| 2023 | 2,079 | 14.8% | — | 301.5 | 164 |
| 2024 | 1,841 | 13.1% | 10.0% | 93.5 | 161 |
| 2025 | 2,622 | 10.3% | 5.9% | 20.0 | 215 |
| Q1/26 (qtr.) | 132 | 21% | n/m (seasonal) | 3.7 (ex-SNAG11) | — |
| LTM to Mar/26 | 2,591 | — | 5.1% | 15.2 (ex-SNAG11) | — |
The first blow was price. The whole sector produced too much seed — a structural surplus of about 30% — and with supply overflowing, the selling price dropped. Add pricier grain on the buying side and an above-normal quality discard (15% versus a historical 10%, due to erratic weather), and gross margin sank from 14.8% to 10.3%. EBITDA per big bag, which measures how much the company earns on each unit sold, fell from roughly R$1,900 (2023) to R$1,100 (2025).
The second blow landed below the operating line: the R$132 million of interest (up 254%) on the roughly R$1 billion of CRAs at CDI+3%. That figure is nearly seven times the year's controlling profit — the expensive debt taken on to fund growth now eats almost everything the operation generates.
The third blow was the customer. Brazilian agriculture is living through the worst credit crisis on record: 1,990 judicial reorganizations in agribusiness in 2025 (up 56%), with rural-credit delinquency climbing from 1.2% to 3.3%. A cash-strapped grower does two things that hurt Boa Safra at the same time: postpones the purchase and falls back on saved seed — seed kept from the farmer's own previous harvest and replanted without paying for certified seed. Today 27% of the country's soybean area already uses uncertified seed. That is the permanent competitive ceiling of the business.
Beware the consolidated-profit trap. The 2025 consolidated profit shows up as R$101 million in some sources — but that is an accounting illusion. Roughly R$81 million of that total belonged to minority partners in the SNAG11 fund, which the company already sold in early 2026. The number that matters for a SOJA3 shareholder is the profit attributable to controllers: R$20.0 million. Always use that one.
So why is there still a buy thesis?
Because the market pays less for the company than its equity is worth on paper. The P/BV (price-to-book value — what the share costs relative to book equity per share) sits at 0.63x, against book value of R$9.59 per share. In plain terms: the market values all of Boa Safra at R$853 million — a market-leading operation that billed R$2.6 billion and has already cleared SNAG11 off its balance sheet.
The turnaround levers are visible: the order book hit a record R$1.5 billion in Q1/26 — unusual for a first quarter and a rare window of visibility into second-half deliveries. Soybeans in Chicago have climbed again (US$12.26 per bushel), the Selic has started to fall (14.25%, projected at 12.5% by the end of 2026 — each 1-point cut is worth R$10 to R$13 million a year of interest relief), and the company is buying back up to 4.95% of its free float, signaling that management itself considers the stock cheap.
There is also a deeper structural catalyst: Bill 1702/19, which would create a royalty on saved seed. If it advances in Congress, it changes the economics of the sector — every real charged on pirated seed pushes acreage back into the formal market, exactly where Boa Safra leads. But it depends on Congress, with no set timeline.
What is it actually worth? The valuation, assumptions on the table
SOJA3 is the classic case where each multiple tells a different story. On today's depressed earnings the stock looks expensive (a P/E around 35x). On book value it looks very cheap (0.63x). The way out is to estimate a mid-cycle profit — neither the 2023 peak nor the 2025 trough — and cross-check three independent methods.
| Method | Fair value | Core assumption |
|---|---|---|
| P/BV × normalized returns | R$7.40 | Book value of R$9.59/share, ROE normalizing at 7-9%, fair P/BV of 0.75-0.80x |
| Normalized profit × P/E | R$6.70 | 4% net margin on R$2.6 bn = ~R$105 mn of profit, at a P/E of 9x |
| Normalized EV/EBITDA | R$6.20 | EBITDA of ~R$220 mn (8.5% margin) × 6x, less average net debt |
| Reference: sell-side consensus | R$6.50 to R$12 | Citi R$6.50 · Bradesco R$9 · Itaú R$10 · XP R$11.80 · BTG R$12 (sanity check only) |
The three proprietary methods converge on a R$6.20 to R$7.40 band, centering on R$7.00 — a +17% upside over today's R$5.97, not enough to offset the risk without signs the margin is confirming. The full fair-value range runs from R$5.00 (floor) to R$10.00 (ceiling). At R$5.97 the market is pricing in almost the bear case; buying below R$5.50 embeds a real margin of safety even in the base case, while paying above R$7.00 means paying for a recovery before it exists.
Three scenarios, three observable triggers
| Scenario | Target price | Probability | What has to happen |
|---|---|---|---|
| Bull — the cycle turns | R$10.00 | ~25% | EBITDA margin back to 8-10% in H2/26, soybeans above US$12.50, Selic at 12%, Bill 1702/19 advancing |
| Base — slow recovery | R$7.00 | ~50% | Order book converted with price still under pressure, partial 7-8% margin, normal seasonal deleveraging |
| Bear — the squeeze becomes a crisis | R$4.80 | ~25% | Rural credit worsening, seed glut persisting, covenants breached, Selic ceasing to fall |
In the bear case, one warning rarely gets mentioned: the book-value discount stops being a floor. A seed company's equity is made of inventory and receivables from stressed farmers — in a liquidation, that is worth less than the balance sheet claims. Cyclical small caps tend to break the rational floor during moments of panic.
Dividends: this is not an income stock
Do not count on meaningful dividends from SOJA3. The company has no formal payout policy and channels profit into its "Expansion Reserve" — the April 2026 shareholders' meeting retained R$19 million out of R$20 million in profit. The R$40 million of JCP paid in December 2025 (JCP, "interest on equity," is a Brazilian way of paying shareholders with a tax advantage for the company) actually exceeded the year's profit — a gesture that consumed reserves and cannot repeat without an earnings recovery. There is even a contractual clause that blocks distributions if the group defaults. Any return from this stock, if it comes, will come from price, not income.
The test of the thesis: November 2026
Everything above resolves into a single observable variable. The acid test is the Q3/26 result, released in November 2026, when roughly 44% of the year's revenue runs through the numbers. The trigger is objective:
An adjusted EBITDA margin of ≥8% in Q3/26 confirms the recovery thesis. Below 6% denies it — signaling the problem is structural rather than cyclical, and the stock would lose its valuation argument.
Beyond the margin, watch: net debt in September and December (the seasonal pattern demands it plunges in the second half — if it doesn't, that's a red flag), agribusiness delinquency and the company's loan-loss provisions (already up from R$0.6 million to R$13 million, healthy while below 3% of receivables), the Chicago soybean price, and the progress of Bill 1702/19.
In one sentence: SOJA3 is a bet that the agribusiness cycle has turned before the numbers prove it — cheap on book, expensive on current earnings, for a small position, a patient wallet and an eye on the November print. Whoever already holds it finds reasons to stay in the record order book and the improving macro. Whoever doesn't can wait for margin confirmation in Q3/26, perhaps paying 15-20% more in exchange for far less risk.
Verdict: UNDER REVIEW — score 5.5/10. Boa Safra is a legitimate, market-leading business, but it is going through a leveraged turnaround in the middle of a rural-credit crisis. Estimated fair value of R$7.00 (range R$5.00 to R$10.00). This is not an investment recommendation; the decision is yours.
This piece summarizes the full analysis. The detailed dossier — fundamentals, risks, governance, scenarios and valuation — lives on the /acoes/soja3/ page (Analysis tab). There is also a PDF report, produced by the site with the help of artificial intelligence — it is not an official document of Boa Safra Sementes S.A.
