Petrobras Round 1 position

The Braskem Room - Run 4

Cutoff: 18 August 2026, approximately 20:50 CT. Public information only.

Classification of statements: [V] verified public record; [R] reported (press / court-docket reporting, not a company 6-K); [I] Petrobras negotiation inference from that record.

This memorandum is the opening Petrobras chair. It is not an affiliation, mandate, or offer. It does not commit Petrobras S.A. to any instrument that has not been independently approved under Lei 13.303.

1. Objective and realistic alternative

Objective. Preserve Braskem as a going-concern Brazilian petrochemical offtaker and jointly controlled affiliate, without taking the US$10.3 billion corporate stack onto Petrobras’s balance sheet, without nationalizing the company, and without writing a cheque that pays legacy financial creditors at the expense of Petrobras shareholders. [V] Petrobras holds 47.03% of voting capital and 36.15% of total capital and jointly controls Braskem with Shine I under a consensus shareholders’ agreement, equity-method accounting. [V] Management has said it does not intend to “statize” Braskem or consolidate its debt and prefers to remain near the current voting percentage.

The only live Petrobras instrument in this run is unsigned, capped naphtha working-capital terms: a related-party payable extension on Brazilian naphtha, at market price, with a hard dollar cap, no Mexico use, and no loss-absorbing feature. [R] Valor/Bloomberg 17–18 August: Petrobras is weighing capped payment-term relief (purchases now often cash / short-dated, conceptually toward six months). Not a loan. Not signed. Cap, price and duration remain open. [R] Pipeline Valor, 18 August 12:16 BRT: Tuesday bondholder meeting demanded a Petrobras commitment to invest in Braskem if financial metrics are missed, as the condition for supporting an extrajudicial filing and 90-day talks; no willingness from Petrobras or IG4 to inject capital or to nationalize.

The following are rejected, not parked for Round 3:

Realistic alternative if these talks fail. Braskem protocols a thin plan-to-a-plan EJ next week, or, if Monday 24 August is empty or the EJ cannot be filed or homologated, Braskem S.A. files RJ. [I] Petrobras does not recapitalize in that scenario. Petrobras continues to sell naphtha on court-supervised commercial terms (cash / short-dated or ordinary-course terms the court authorizes). Brazilian RJ is debtor-in-possession, with weaker absolute priority than Chapter 11, headcount plus Lei 11.101 art. 58, and equity generally left in place. [I] A process discount of 15–25 percent, LC and trade disruption, and utilization risk are real; they are cheaper for Petrobras shareholders than consolidating US$10.3 billion or signing a conversion DIP that drops the joint controllers below 10 percent.

Ranking of bad outcomes for Petrobras (best to worst among failures):

1. Thin EJ or RJ with joint control and the equity stub intact; naphtha sold on court terms. [I]
2. Consensual extension with capped naphtha and creditor warrants, control retained. [I]
3. Cure-and-continue that simply pays the disclosed defaults into a December RCF wall. [I]
4. Conversion DIP or all-asset lien package that drops Shine/Petrobras below 10 percent, or a Petrobras put. [I]

[I] We prefer (1) to (4). RJ that leaves 47.03% / 36.15% in place is superior to a “rescue” that is a disguised change of control. A foreign-led conversion that wipes Shine also creates a non-trivial IFRS 10 question: whether Petrobras, as the remaining 47 percent holder with a strategic offtake, is deemed to have de facto control and must consolidate. That question is a hard stop, not a drafting point.

Political overlay (not a guarantee). [V] First round 4 October 2026; possible second round 25 October. [V] Lei 13.303 requires governance, transparency and defensible related-party economics; a political preference does not authorize an uneconomic rescue. [V] Lula has said Petrobras should consider Brazil’s priorities and that the government does not command the company. Treat that as orientation, not an instruction to buy Braskem bonds. The election-year double bind is bailout-of-IG4 versus jobs/supply-chain. Capped, documented naphtha terms sit in the only politically usable gap. An overt put does not.

2. Do unsigned naphtha terms satisfy burden-sharing after Idesa - and at which cap?

No, not as loss-absorbing capital, at any of the three caps. Supplier credit is liquidity. It is not equity, not a junior loan, and not a substitute for the put Elliott/Contrarian asked for on Tuesday. [I]

[V] Braskem Idesa 18 August 6-K: Braskem will contribute US$476 million, of which about US$126 million was already made available; about US$350 million remains. Treat as live parent / Netherlands cash use unless a later filing proves it is only a restatement of the existing term loan (US$180 million committed / US$129 million disbursed) or the US$82 million Idesa-secured WC loans. [V] Separate estate; no disclosed parent guarantee of the remaining about US$1.6 billion Idesa senior debt; Braskem keeps a majority; Idesa haircut does not recapitalize the US$10.3 billion corporate stack.

That remaining ~US$350 million reduces Petrobras’s willingness to stretch naphtha, including to the US$950 million mathematical ceiling, for four reasons. [I]

1. Parent cash that could have rebuilt Brazilian liquidity or paid the disclosed ~US$98 million defaults is instead supporting a Mexican unrestricted subsidiary. Financing that leak with longer Petrobras payables is a related-party transfer to Mexico by another name.
2. Lei 13.303 independent-committee economics get worse as Braskem simultaneously funds Idesa and asks Petrobras for 180-day credit while already in uncured default and Fitch RD.
3. If a parent EJ or RJ is filed before Idesa’s ~55-day effective-date milestone, the remaining contribution / DIP may freeze anyway. [R] Disclosure statement (first-day reporting) flags that certain Braskem actions could be reviewed in Brazil if the parent enters EJ or RJ. Stretching naphtha now to “make Idesa close” is not a Petrobras objective.
4. Naphtha terms that can be drawn into Mexico, or whose cash release is used to fund the remaining Idesa cheque, are prohibited in any package we will sign.

Cap tests (Brazilian naphtha only; market price; no Mexico)

[R] Mathematical ceiling of moving about nine-day payables to 180 days is roughly US$950 million. Implied average daily naphtha offtake ≈ US$950m / 180 ≈ US$5.3 million/day. Baseline ~9-day payable stock ≈ US$47 million. Incremental stock at each cap:

Burden-sharing score after Idesa. At US$250 million, naphtha is a partial restoration of supplier finance that Q2 already showed was lost; it does not meet the June AHG demand for shareholder burden-sharing. [V] June AHG required positive-NPV treatment, shareholder burden-sharing, cash controls, information rights and Petrobras at the table. Sitting at the table with unsigned US$250 million terms is presence, not burden-sharing. At US$500 million it is still liquidity, not capital, but it is a real WC stock (about 16 percent of June unrestricted cash plus the Idesa remainder, or about 5 percent of adjusted net debt). It can be described to an independent committee as financing profitable PE-naphtha conversion in Brazil, if cash conversion is evidenced and creditors extend. It still does not replace a put, and we will not let counterparties pretend otherwise. [I]

Pricing. Market naphtha. No embedded discount, no PRESIQ/REIQ capitalization, no election-window subsidy. Duration: 12–24 months, cancellable on conversion, on a Petrobras voting-control event, or on use of proceeds for Mexico, dividends, or unsecured principal. [I]

Sequence. No incremental outstanding above ordinary-course (~9-day) until near-dated financial debt is extended. Otherwise Petrobras payables simply fund the December US$1.0 billion standby and the remaining Idesa cheque. [I]

3. Is IG4 / Shine cash available at all?

No material IG4 cash is available for a parent recapitalization. [R] Pipeline Valor 18 August: no IG4 aporte. [V] Shine acquired 50.1108% of voting capital and 34.3234% of total capital from NSP/Novonor through a distressed exchange, not a cash recap of Braskem S.A. [I] IG4 disclosed AUM around US$1 billion; that is not uncalled Braskem dry powder and will not be treated as such.

[V] The Idesa US$476 million is company / Braskem Netherlands cash, not an IG4 fund cheque.

Petrobras will not put a positive number on a Shine junior facility in this run. Doing so would require explaining why Pipeline is wrong and why Petrobras would allow IG4 to be the only cash sponsor - which we would not. A Petrobras-only cash rescue of IG4’s option is the bailout narrative Lei 13.303 and the 4 October campaign make indefensible. [I]

What IG4 did buy, and what it may keep if it dilutes, is joint control. Conversion that drops the controllers below 10 percent remains rejected by this chair whether or not IG4 later finds a cheque. [I] If Shine wants to retain 50.1% voting, it must accept warrant dilution, cash controls, and the absence of a Petrobras backstop - not a keep-well from the 47.03% partner.

4. Will Elliott / Contrarian support a one-third filing without a Petrobras put? If not, how does a filing still happen?

No. [R] Pipeline 18 August: the Tuesday ask was a Petrobras contingent injection as the price of supporting an EJ filing and 90-day negotiation. [R] Bloomberg 10 August: Contrarian is in a hard group with Elliott on the Petrobras-cash ask. [V] July: bondholders including Elliott offered (i) a DIP that converts and drops controllers below 10 percent or (ii) a maturity extension at existing coupon with all assets as collateral; Braskem rejected both. [V] June AHG support for a temporary plan-to-a-plan EJ was conditional on protections, not adhesion to Braskem’s five-year / three-year grace / no-haircut economics.

Do not assume a public blocking percentage for Elliott. Do not assume the steering committee is 51 percent of the notes. [I] Treat Elliott/Contrarian as hostile to any filing that is not a put or a conversion, and as the dominant voice on the AHG steering chair - not as the whole book.

How a filing still happens without them

[V] Lei 11.101 art. 163 §7: more than one-third of each affected espécie to protocolar an EJ; more than 50 percent of each class to homologate. Pipeline’s “two-thirds to file” is wrong. If NY notes are one class, more than about US$2.35 billion of the listed ~US$7.02 billion parent securities is needed to file that class. Homologation of that class needs more than ~US$3.51 billion. Estadão has reported that without some bondholders the company cannot reach one-third.

Listed parent securities at 30 June (US$ million): ’28 1,250; ’30 1,500; ’31 850; ’33 1,000; ’34 850; ’41 587; ’50 750; hybrid ’81 231. Total ~US$7.02 billion. Remainder of the US$10.3 billion stack is the drawn US$1.0 billion RCF plus debentures, CRAs and bank / ECA lines. [V]

Path to a notes-class one-third without Elliott/Contrarian adhesion [I]:

• Peel long-only accounts named in public reporting - Capital Group, AllianceBernstein, PGIM - plus SVP (reported as holding notes and some RCF, not a formal Elliott pairing). A coalition of those holders plus non-AHG accounts can, in principle, clear US$2.35 billion even if Elliott is the largest single name. It cannot be assumed; it has to be counted per série / espécie.
• [R] Local banks (Itaú, Safra, KfW, Bladex, DZ, SMBC and others) are described as supportive of an EJ. Debenture / CRA / bank classes are the easier espécies.
• The US$1.0 billion standby / RCF, drawn October 2025, matures 31 December 2026; Elliott is reported as its largest lender. [R]/[V] If the RCF is an affected class and Elliott holds a majority of it, Elliott can block filing of an EJ that includes the RCF, not merely homologation. The mechanical answer is a notes-and-local-bank EJ that leaves the RCF unaffected (or a later RJ that does not need Elliott’s one-third). A notes-out EJ does not stay Elliott after 24 August; it also does not stay the RCF wall.
• [R] O Globo, 17:36 BRT: company-side sources say Braskem is preparing to file an EJ next week before 24 August; first filing is a 90-day process plan, not a recap; no asset sales in the first filing. Do not treat this as locked adhesion. The URL slug overstates creditor approval; the body still recites the one-third bar.

Petrobras condition on being used as the peel. We will not supply the put that Elliott demanded in order to manufacture a one-third. We will not sign naphtha terms as a side letter that long-onlys can wave as “sponsor support” unless the cap, Brazil-only use, and no-guarantee language are in the four corners. A thin EJ filed with local banks plus a notes peel, Elliott/Contrarian outside or hostile, is acceptable to this chair. [I] Homologation at 50 percent is a later problem; we do not pre-commit economics to buy Elliott’s vote.

Tutela clock. [V] 60-day Brazilian tutela granted 26 June; reported expiry Monday 24 August 2026. Chapter 15 provisional relief from 30 June. Stay covers invited Wind Chamber financial creditors only; trade is not stayed. If next week is empty, acceleration risk on the disclosed defaults rises. [V] After July cure periods Braskem disclosed R$507 million / about US$98 million of defaults under certain financial instruments; no public acceleration as of this cutoff. [V] Public bond terms generally: 30-day interest cure; 25 percent of a series can accelerate after an uncured event of default. [V] Fitch revised the parent to RD on 17 August; a subsequent EJ or RJ would take ratings to D. Ratings are not a reason to inject common.

5. Opening, first concession, narrowest acceptable package

Opening (Round 1)

• Petrobras: no cash equity, no put, no keep-well, no guarantee. Willingness to negotiate capped Brazilian naphtha payable terms at US$250 million, market price, ~60-day terms, 12-month tenor, borrowing-base logic on financed Brazilian inventory, receivables and proceeds only. Unsigned until creditor duration is locked. [I]
• Creditors: five-year extension on near and intermediate maturities (including the 31 December 2026 RCF and the ’28 / ’30 notes); no coupon cut as a Petrobras-sponsored gift; 50 percent PIK for 12 months then existing coupon. Legacy financial debt remains unsecured. [I]
• Burden-sharing: IG4/Shine dilute via 5.0–7.5 percent warrants to extending creditors. Petrobras dilutes pro rata on existing Braskem shares; Petrobras does not issue new Braskem common for cash. [I]
• Filing: a 90-day plan-to-a-plan EJ is tolerable. Braskem’s current five-year / three-year grace / no-haircut term sheet is not an economic deal Petrobras will dress up as sponsor-supported. [R]/[I]
• Sequence: extension documentation before incremental naphtha outstanding. Cash controls and information rights from day one of any EJ. [I]
• Mexico and Alagoas: see §§6–7. No Petrobras dollars, no Petrobras guarantee. [I]

First concession (Round 2, only if duration is real)

• Naphtha cap to US$500 million; terms toward ~100–120 days; still Brazil-only, still market price, still no Mexico. [I]
• Warrants to 10 percent. [I]
• After the PIK window, existing coupon +100 bp. That is creditor NPV from the debtors’ economics, not from a Petrobras fill. [I]
• Accept a thin EJ that peels long-onlys and local banks with Elliott/Contrarian outside, provided the EJ does not contain a springing Petrobras put, a keep-well, or an obligation to raise the naphtha cap to US$950 million. [I]

Narrowest acceptable (final)

• Naphtha: US$500 million cap. US$950 million remains rejected after Idesa. [I]
• Maturity: five years on near/intermediate; RCF extended at least 24 months past 31 December 2026 before naphtha outstanding exceeds US$250 million. [I]
• Coupon: 12 months 50 percent PIK, then existing coupon +150–200 bp. No mandatory principal haircut from this chair; we will not block a modest par-for-par NPV exchange if it is what peels long-onlys and keeps equity above 10 percent. [I]
• Warrants: 15 percent, stepping to 22.5 percent if agreed cash-conversion, liquidity or PIK milestones are missed. Fully diluted, Petrobras voting ≈ 40.0 percent at 15 percent and ≈ 36.4 percent at 22.5 percent if pro rata; both remain far above 10 percent; joint control by SHA can be preserved. Immediate conversion that takes controllers below 10 percent remains rejected. [I]
• Collateral: Petrobras naphtha super-priority only on the financed Brazilian inventory/receivables/proceeds. Reject blanket liens over core Brazilian plants (Camaçari and other cracker/PE assets) for legacy unsecured debt. [I]
• No converting DIP. No all-asset package that is economically the July (ii) offer dressed as an EJ. [I]
• Maximum Petrobras exposure in this run: US$500 million of naphtha payable stock, not US$750 million of mixed WC-plus-hybrid from prior internal framing, and not common. A deeply subordinated hybrid or rights issue is not on the table while Pipeline reports no capital injection and while IFRS 10 / Lei 13.303 / TCU / 4 October remain live. [I]

If the choice is this narrow package versus a conversion DIP, we refuse the DIP and accept RJ. [I]

6. Package grid

7. Mexico treatment (required in every package)

TQPM. [R] Pipeline cites a TQPM equity-support agreement and Slim as that lender as the reason they say no holding cross-default. Treat as company-friendly spin, not as a parent-guarantee analysis, and not as a Petrobras obligation.

Utilization in Mexico was 43 percent in Q2; Mexico Q2 EBITDA was US$57 million. Do not annualize Mexico. [V]

8. Operating cases

No Q3 result is assumed. [V] June plan Q3 EBITDA about US$586 million is the base operating case; US$750 million is a sensitivity, not a reported result; US$1.0 billion is an upside sensitivity.

Starting stock. [V] 30 June: corporate gross debt US$10.3 billion; adjusted net debt US$9.5 billion; adjusted corporate net leverage 6.74x (ex-Idesa project debt). Cash and cash equivalents R$3.931 billion, of which R$353 million inside Idesa. June plan unrestricted cash about US$795 million. Q2 recurring EBITDA US$1.043 billion (Brazil/SA US$869 million, US/Europe US$147 million, Mexico US$57 million); H1 about US$1.235 billion. Q2 working capital consumed US$547 million; price/volume components are level resets if prices and inventory stabilize; lost supplier finance is persistent. H2 contractual debt service in the June plan US$2.349 billion including LCs and the standby. Q3 contractual debt service US$878 million including US$572 million LC runoff; June plan unrestricted cash about US$337 million at 30 September under the status quo.

Implied Q3 status-quo cash burn at plan EBITDA: 795 − 337 = US$458 million. [V]

Remaining Idesa use. Unless later 6-Ks prove restatement, assume US$350 million leaves parent/Netherlands cash inside the Idesa ~55-day window (late August–early October). [I] That is not in the sentence of the June plan we can verify. Treat September unrestricted cash as US$337 million minus whatever portion of the US$350 million is actually funded in Q3.

Naphtha as a stock, not a flow. A signed cap releases WC once: +US$250 / +US$500 / not +US$950. It does not raise EBITDA. [I]

Case A - Q3 EBITDA US$586 million (June plan) [V]

[I] Case A with Idesa funded and no naphtha is a September liquidity event. That is an argument for a thin EJ or RJ, not for a Petrobras put. US$250 million naphtha in Case A is a bandage on Brazilian payables; it does not pay LCs or the RCF. US$950 million would look like filling the entire Q3 hole plus Idesa - which is exactly why it fails Lei 13.303.

Case B - Q3 EBITDA US$750 million (sensitivity, not reported) [V]

Delta vs plan: +US$164 million EBITDA. [I] If 60–70 percent converts after persistent SCF loss, incremental cash ≈ US$100–115 million. September unrestricted cash ≈ US$440–450 million before Idesa; ≈ US$90–100 million after a full US$350 million Idesa funding.

Case C - Q3 EBITDA US$1.0 billion (upside sensitivity)

Delta vs plan: +US$414 million. [I] At 60–70 percent conversion, incremental cash ≈ US$250–290 million. September unrestricted cash ≈ US$590–630 million before Idesa; ≈ US$240–280 million after Idesa.

All three cases. The December US$1.0 billion standby is not a Petrobras take-out. Pay in full: not in A, not in B, not in C on verified cash. Extend: requires an EJ/amendment Elliott may not give on the RCF class. Restructure: thin EJ carving out the RCF, or RJ. [I] Petrobras naphtha does not change that binary.

9. Alagoas

[V] Alagoas remains legally live. In June 2026 Braskem and former executives became defendants in a federal proceeding related to the Maceió mining disaster. Braskem continues to disclose administrative proceedings and socio-environmental obligations.

Petrobras terms for every consensual case. [I]

• Ring-fence budgeted safety, stabilization and remediation cash. Those payments are not a coupon-negotiation chip.
• No dividends, no related-party leakage, and no warrant cash-out while Alagoas budgeted cash is unpaid and leverage is above a going-concern threshold.
• No Petrobras guarantee, indemnity, or keep-well for Alagoas. An election-year “national champion” story that socializes Maceió onto Petrobras is politically and legally worse than RJ.
• Information rights: creditors get the same Alagoas cash reporting we would insist on as a jointly controlling shareholder.
• A deal that appears to subordinate victims to NY noteholders will not be supported by this chair and would not survive public-company scrutiny.

Alagoas uncertainty also limits the collateral value of “unencumbered” Brazilian cash. It is a further reason not to grant blanket plant liens to legacy unsecured debt. [I]

10. Outcome probabilities

Must sum to 100 percent. These are Petrobras-chair probabilities over the next 90 days, not a market-implied distribution. [I]

Conditional on a thin EJ being filed next week: homologation of a process plan inside 90 days is a separate, lower-probability event than protocol. A filed thin EJ that later fails homologation is counted here as pathing into RJ, not as a completed economic EJ. [I]

Fitch. RD already. EJ or RJ → D. Not a decision variable for naphtha caps or equity. [V]/[R]

11. The single fact that would most change this position

A legally binding, cash, parent-level junior commitment from IG4/Shine, board-approved and large enough to be loss-absorbing (order of several hundred million dollars), contradicting Pipeline Valor of 18 August 12:16 BRT.

That is the only fact that would let Petrobras describe capped naphtha as shared commercial support rather than a one-sided related-party rescue, and the only fact that would move this chair from US$250 million toward US$500 million with speed. It would not clear a contingent put, common equity, keep-well, parent guarantee, nationalization, Mexico use, or the US$950 million ceiling. It would not make a conversion DIP acceptable.

Secondary facts that would move caps or probabilities but not red lines: (i) a later 6-K proving the remaining US$350 million Idesa item is only a restatement of the existing term loan / US$82 million WC, which would raise willingness to treat US$500 million as the working cap rather than the stretch; (ii) a Q3 cash-conversion print showing September unrestricted cash already below US$200 million after Idesa, which would cut even the US$250 million discussion until the RCF is extended, because the release would otherwise pay Elliott; (iii) acceleration by 25 percent of a notes series after 24 August, which would raise the RJ probability and end any residual cure path.

Until (and unless) the IG4 cash fact appears, this chair’s instrument is unsigned capped Brazilian naphtha working-capital terms, tested at US$250 million / US$500 million / US$950 million and accepted only in the first two boxes, and the realistic alternative is court-supervised naphtha sales in EJ or RJ - not a recapitalization of Braskem by Petrobras.

Appendix A - Hard stops (non-negotiable in this run)

1. No Petrobras common equity. [R]/[I]
2. No contingent metric put. [R]/[I]
3. No keep-well. [I]
4. No parent guarantee (financial debt, Idesa, Alagoas, LCs). [I]
5. No nationalization; do not take voting through 50 percent; do not break joint control with Shine. [V]/[I]
6. No conversion or other mechanic that drops the joint controllers below 10 percent. Prefer RJ. [I]
7. No Mexico use of naphtha terms or naphtha-released cash. [I]
8. Remaining Idesa ~US$350 million is a reason to tighten, not loosen, the naphtha cap. [I]
9. US$950 million ceiling is a math identity, not a bid. [R]/[I]
10. Lei 13.303, TCU scrutiny, the 4 October 2026 election, and IFRS 10 at 47.03% voting are the legal-political basis for 1–6, not talking points. [V]/[I]

Appendix B - Core sources used

• Braskem ownership: https://www.braskem-ri.com.br/a-companhia/estrutura-societaria/

• Petrobras Q2 6-K: https://www.sec.gov/Archives/edgar/data/1119639/000129281426004133/pbrfs2q26usd_6k.htm

• Idesa 6-K, 18 August: https://www.sec.gov/Archives/edgar/data/1071438/000129281426004279/bak20260818_6k.htm

• Idesa PR: https://www.prnewswire.com/news-releases/braskem-idesa-sapi-takes-strategic-action-through-consensual-restructuring-to-reduce-its-debt-by-more-than-us920-mm-302853634.html

• Idesa claims agent: https://restructuring.ra.kroll.com/BraskemIdesa/

• Pipeline Valor, 18 August: https://pipelinevalor.globo.com/negocios/noticia/o-impacto-do-chapter-11-da-braskem-idesa-para-a-holding.ghtml

• O Globo, 18 August: https://oglobo.globo.com/economia/negocios/noticia/2026/08/18/braskem-consegue-aval-de-credores-e-prepara-pedido-de-recuperacao-extrajudicial.ghtml

• Lei 13.303: https://planalto.gov.br/ccivil_03/_ato2015-2018/2016/lei/l13303.htm

• TSE 2026 calendar; Lula/Petrobras remarks; PRESIQ Law 15.294/2025; Alagoas federal proceeding (Reuters via UOL) as cited in the shared political-context brief.

• End of Petrobras Round 1 -