Cosan Puts Control on the Table — Then Picks It Back Up
A 0.3x coverage ratio, a departing CEO, and the word 'control' entering the divestment vocabulary for the first time.
CSAN · Earnings Call · 2026-11-16
The Coverage Ratio Is the Whole Quarter
Strip out the pleasantries and Cosan's Q3 2026 call is a single number wearing a suit: the debt service coverage ratio came in at 0.3x. Fernando Tinel delivered it without flinching — “the debt service coverage ratio, DSCR reached 0.3x in the third quarter of 2026, an improvement of 0.1x compared to Q2 '26” — Fernando Tinel, Executive · 2026-11-16 — and immediately explained why it is not as bad as it looks. The dividend calendar is back-loaded: Radar's Mato Grosso land sales (up to BRL 586 million) and Compass's recently announced ordinary dividends (BRL 377 million) have not yet landed in the calculation. Guidance for December was reaffirmed at 0.8x to 1.2x, and management nudged analysts toward the upper half.
The balance sheet itself is moving in the right direction, just slowly quarter-on-quarter. Expanded net debt was BRL 9.5 billion versus BRL 9.2 billion in Q2, and the year-on-year comparison is the flattering one — down 48% against Q3 2025 as the deleveraging initiatives of the past twelve months bite. Average maturity sits at six years, average cost stable at CDI plus 1.14%. Nothing here is broken. But a 0.3x ratio is a holding company living off a lumpy dividend schedule, and everyone on the call knew it.
Boundary Conditions
Matheus Enfeldt of UBS asked the question that set the tone for the entire Q&A: what are the guardrails on asset sales, and will Cosan give up control? Cesario's first answer was expansive. “We are focusing on divesting and that can include any of our assets as long as there's a fair price and boundary conditions are adequate. And that includes control.” — José Cesario Menezes de Barros Sobrinho, Executive · 2026-11-16 He then spent the next question walking it back: Cosan prefers not to sell controlling stakes, the deals under discussion do not involve giving up control, and the real goal is bringing net debt close to zero.
That is a meaningful wobble, and it is new. The keyword boundary conditions and terms of control both landed in the company's top ten this quarter — vocabulary that simply was not in the deck three quarters ago. Compare the register. In August, Rafael Bergman described Radar's monetisation with “we don't have any intentions to liquidate assets at any cost” — Rafael Bergman, Executive · 2026-08-17. A year earlier, Marcelo told the same audience “there is no need for any fire sale of assets… We don't want to give anybody the impression that we're rushing around trying to sell assets” — Marcelo Martins, Executive, likely CEO or CFO · 2025-11-17. And back in August 2025: “We're not going to sell any company in the portfolio in its entirety.” — Unidentified Company Representative, Company Representative (likely IR or Executive) · 2025-08-15
Those were all statements about price discipline. This quarter's first answer was a statement about scope — and scope is the more consequential variable. When a controlling shareholder's lieutenant says out loud that control is negotiable, buyers hear it, and the bid they bring next changes.
We prefer not to sell the controlling share. And the deals we are looking at do not include giving up control of any of our businesses. But our main focus is on divesting part of the assets in our portfolio so that we can bring our net debt close to 0.
What Went Quiet
The loudest thing on this call was a silence. For several quarters the Shell relationship and Cosan Nove were the centre of gravity — Cosan Nove was the single highest-momentum topic in the 2025 second-quarter call and Shell was a recurring top-ranked theme through 2025. On this call, Gabriel Barra of Citi asked directly for an update on Cosan Nove post-Raizen restructuring, and got: “As for Cosan Nove, no news. Basically, it's what we have been telling the market… there is no ongoing negotiation with anyone to address that.” — Unknown Executive, Executive · 2026-11-16
Raizen is now framed purely as a completed event — the court restructuring is described as something the company "went through or is going through the conclusion of," with Cosan on the sideline. Given that Marcelo spent the November 2025 call saying he had a sense of urgency measured in six months, not two years, on Raizen capital structure, the quiet is itself informative. Either the problem is being solved outside Cosan's balance sheet, or the urgency has been reframed as patience.
The other defection is Moove. A quarter ago, Marcelo was categorical: no IPO, no stake sale, “We don't expect to sell our stake at Moove at the moment.” — Marcelo Martins, Executive · 2026-08-17 Now the question is dividends, and the answer has softened into “dividends may be paid out. That's an ongoing conversation. Nothing has been approved yet, but the dividends may be paid out pending some approvals still.” — Unknown Executive, Executive · 2026-11-16 Moove earned it: EBITDA up 21%, net revenue up 26%, leverage at 1.3x, with Cesario crediting a decentralised post-fire production platform and market conditions in oil prices feeding through to lubricant and byproduct prices. That is a genuine tailwind in a market where high fuel prices have been punishing most of the tape — Moove is on the benefiting side of the same price move that has hurt airlines and truckers all year.
The Land Machine and the New Mandate
Radar is now the deleveraging engine, and management was unusually specific about the bid. A BRL 808 million November land sale was disclosed, and on next year's pipeline: “we expect to sell, if not the same amount of land, very similar. So we're very optimistic because we have been seeing considerable interest by some players to buy land.” — Unknown Executive, Executive · 2026-11-16 That considerable interest is doing real work — Radar's net revenue fell 2% on weaker sugarcane lease revenue tied to TRS prices, but net income rose 14% on a larger cash balance. The asset is worth more as a source of liquidity than as a recurring earner right now.
Which is exactly the mandate Guilherme Penin inherits. Marcelo's sign-off was part farewell, part instruction manual: “no assets are excluded from a potential sale, full or partial… we need to be more efficient to reduce Cosan's size, to focus on expanding the businesses, which are becoming increasingly more independent” — Marcelo Martins, CEO · 2026-11-16. Penin answered with value creation, transparency and governance — and a telling phrase about arriving with a different mandate than his predecessor's. The CEO change is the real news under the numbers: Cosan is moving from a builder of a portfolio to a seller of one, and the departing CEO just said so more clearly than the incoming one will be able to for a few quarters.
For investors, the honest read is that the operating assets are fine — Rumo's EBITDA was stable on a comparable basis once BRL 23 million of simplification costs and last year's BRL 60 million of insurance and tax credits are stripped out, Compass's free-market volumes grew 20% — while the holding company remains the risk. A 0.3x coverage ratio with a 0.8x–1.2x year-end target is a narrow corridor, and the only lever that widens it is the one management just publicly opened to negotiation.