Sylvamo Corporation (SLVM)

1. Executive summary

Sylvamo is the largest pure-play producer of uncoated freesheet paper in the western hemisphere: the cut-size reams that feed office printers, the offset rolls that become books and direct mail, the envelope and forms grades that still move through commerce. The company was spun out of International Paper in October 2021 with eight mills across three regions, a portfolio of leading brands (Hammermill and Accent in North America, Chamex and Chambril in Latin America, REY and Multicopy in Europe), and an investment case built entirely on cash: run structurally advantaged low-cost mills hard in a shrinking market, spend little, and return the difference. For its first three years the case worked. Between the spin and the end of 2024 Sylvamo generated roughly $2.5 billion of cumulative adjusted EBITDA and more than $960 million of free cash flow (Q4 FY25 call), cut gross debt by about half, initiated and quadrupled a dividend, and retired roughly a tenth of its shares. The stock peaked above $90 in November 2024.

2025 broke the pattern and 2026 is breaking it harder. Full-year 2025 adjusted EBITDA fell 29% to $448 million on an 11% revenue decline to $3,351 million, free cash flow collapsed from $248 million to $44 million, and Europe, the segment acquired and expanded through the 2023 Nymölla purchase, swung to a $78 million adjusted EBITDA loss (FY25 release). The first quarter of 2026 then printed $29 million of adjusted EBITDA at a 4% margin, a $3 million net loss, and negative $59 million of free cash flow (Q1 FY26 release). Three separate forces are compounding: a cyclical trough in European paper pricing against elevated wood costs; a self-inflicted but deliberate transition year in North America, where International Paper's conversion of its Riverdale mill ended a 260,000-ton supply agreement in April 2026 while Sylvamo's own $145 million Eastover investment programme requires a 45-day fourth-quarter outage; and roughly $85 million of one-time costs, sourcing frictions and war-driven input inflation layered on top. Management has suspended buybacks for two consecutive quarters and is funding the $72 million annual dividend from the balance sheet. Net debt has risen to $791 million, 2.0x trailing adjusted EBITDA, and will approach $900 million by year-end on our numbers.

The variant question is not whether 2026 is a trough. It almost certainly is: the Riverdale transition costs (a $65 million adjusted EBITDA drag this year at current tariff rates) do not repeat, the Eastover outage does not repeat, and the Eastover projects themselves should add more than $50 million of annual EBITDA from 2027 at a stated internal rate of return above 30% (Q1 FY26 release; Q2 FY25 release). Price increases of 4–8% are being realised across all three regions, and the Riverdale conversion removed roughly 7% of annual North American industry supply just as imports rolled over. The variant question is what the recovery is worth. Our model rebuilds adjusted EBITDA to $470 million in 2027 and about $485 million in 2028, with free cash flow of $160–180 million a year. Discounted at an 8.5% weighted cost of capital with a modest perpetual fade, that stream is worth about $35 per share. The reverse-DCF is the uncomfortable part: at $37.45, the market is already paying for sustainable EBITDA of $490–540 million depending on the fade assumed, which is management's aspiration rather than the filings' demonstrated run-rate. The optically low forward multiple, 4.9x our FY27 EBITDA, is not a mispricing. It is the correct price of a terminal-decline business carrying $0.8–0.9 billion of net debt.

Two things keep us from a negative rating. First, the dividend and the asset value provide a floor of sorts: eight mills, Brazilian forestlands, and a demonstrated capacity to produce $250–300 million of free cash flow in ordinary years support the low-$30s even on unheroic assumptions. Second, and more interesting, Atlas Holdings, a Greenwich private equity firm with a thirty-company industrial and paper portfolio, holds 16% of the shares outright and 21.5% economic exposure including derivatives. In November 2025 Atlas pulled its two board designees, terminating a 2023 cooperation agreement and freeing itself of standstill restrictions; the board responded with a one-year poison pill expiring 9 November 2026 (10 Nov 2025 release). Atlas took Office Depot's parent private for about $1 billion the following month, giving it ownership of one of the largest cut-size paper retail channels in North America. A bid, a proxy contest, or a negotiated transaction before or after the pill lapses is a live possibility. We assign it no explicit weight in the base case, and we note that our bull scenario ($53) approximates what a control transaction at six times mid-cycle EBITDA would deliver.

The probability-weighted value across our bear, base and bull cases is roughly $34 against a $37.45 close, a spread the 4.8% dividend yield approximately fills. That arithmetic describes a Hold: the market has already de-rated the shares 34% over twelve months and now discounts most of what can be responsibly underwritten, while the tail outcomes in both directions are wide. What would change our mind is set out in Section 11; the shortest version is evidence that European margins are normalising faster than priced, or a price below roughly $30 that restores a genuine margin of safety against the base case.

Sources: FY22–FY25 earnings releases and 10-K filings; Q1 FY26 release. 2021 income figures are continuing operations excluding Russia; 2021 free cash flow included only one quarter of standalone cash interest and taxes and is not meaningful. 2022 net income of $336m is continuing operations; including the $218m after-tax loss in discontinued Russian operations, total 2022 net income was $118m. Net debt for 2021–2023 is approximate, reconstructed from disclosed gross debt and cash; 2023 gross debt was $950m against $60m of restricted escrow cash (Q4 FY23 presentation). n.d.: not compiled here; n.m.: not meaningful.

2. How the company got here

Sylvamo exists because International Paper wanted out of printing papers. On 1 October 2021 IP distributed its global uncoated freesheet business to shareholders as a standalone company, loading it with roughly $1.4 billion of debt and two umbilical cords: multi-year offtake agreements under which IP's Georgetown and Riverdale mills would keep producing paper for Sylvamo to sell, and a transition-services scaffold typical of spins. The market treated the newco as an orphan. It began trading at a market value of about $1.1 billion, less than two times the EBITDA it would earn in its first full year. Management, led by IP veteran Jean-Michel Ribiéras, set out a deliberately simple thesis: Sylvamo is a cash-flow story, the phrase used in the first full-year release (FY22 release), and every subsequent capital-allocation decision was framed against it.

Three portfolio moves defined the early years. First, the exit from Russia: the Svetogorsk mill, a large and profitable operation, was classified as discontinued in mid-2022 and sold on 2 October 2022, crystallising a $218 million after-tax loss in the discontinued line but removing sanction and expropriation risk and funding accelerated debt paydown (FY22 10-K). Second, the January 2023 acquisition of the 500,000-ton Nymölla mill in Sweden from Stora Enso for $167 million, a counter-cyclical purchase of an asset that had just completed a $40 million pulp-mill modernisation and generated about $50 million of cash before allocated overhead in its first year (Q4 FY23 call). Third, the steady redirection of capital into the lowest-cost assets: Brazilian forestlands, the Saillat and Nymölla mills in Europe, and above all Eastover, South Carolina, which management describes as its lowest-cost mill and where the current $145 million programme (a $100 million paper-machine optimisation adding 60,000 tons, a $45 million replacement cut-size sheeter, and a woodyard modernisation executed with an external operating partner that avoids about $75 million of capital over five years) concentrates.

The shareholder register changed as fast as the portfolio. Atlas Holdings accumulated a large position rapidly in 2022; a 2023 cooperation agreement, the negotiation of which cost Sylvamo $4 million in professional fees (FY23 release), placed two Atlas-affiliated directors on the board and bound Atlas to a standstill. That settlement held until 10 November 2025, when Atlas directed its designees to resign, terminating the agreement; the board adopted a one-year shareholder rights plan the same day. Leadership turned over almost simultaneously: Ribiéras retired on 31 December 2025, chief operating officer John Sims became chief executive on 1 January 2026 with David Petratis, lead independent director since the spin, taking the chair. Sims inherited the Riverdale wind-down, the Eastover build, a European segment losing money, and a 21.5% holder with no remaining contractual constraints. That inheritance, more than any macro variable, is what 2026 is about.

One judgement on the record deserves revisiting with the benefit of three years' hindsight, because it bears directly on how much trust to extend to the current management's capital decisions. The Nymölla purchase looked shrewd at signing: $167 million for half a million tons of modernised capacity, roughly three times the cash the asset generated in its first year under Sylvamo ownership. It looks considerably less shrewd today, with the European segment losing money in aggregate, an impairment already taken in France, and Nymölla itself contributing boiler and debarking failures to the 2026 reliability bill. The fair verdict is probably that the asset was bought well and the region was bought badly: the mill's economics were and are sound relative to its European peers, but the purchase doubled the company's exposure to the one geography where it holds no structural advantage, at the exact moment that geography's cost base was resetting against it. The lesson we carry into the valuation is not that management overpays, the price was demonstrably low, but that counter-cyclical courage in a structurally declining region is a different and more dangerous virtue than the same courage applied at Eastover, where every dollar deepens an existing advantage. The current programme spends where the moat is. That is the right lesson learned, whether or not it is ever stated as one.

3. The business and how it earns money

Sylvamo makes one thing at scale: uncoated freesheet, the paper family that runs through office copiers, laser printers, commercial presses, book manufacturing, envelopes and business forms. It also sells market pulp opportunistically when integration economics favour it. The company operates seven paper mills plus a converting site: Eastover, South Carolina and Ticonderoga, New York in the United States, with sheeting at Sumter, South Carolina; Saillat in France and Nymölla in Sweden; and Luiz Antônio, Mogi Guaçu and Três Lagoas in Brazil, supported by company-managed eucalyptus forestlands that feed the Brazilian system (FY25 10-K). Around 6,500 people work across the network. Revenue splits into three reported segments that behave like three different businesses sharing a product.

North America is the profit engine: $1,754 million of 2025 net sales, 52% of the total, with $334 million of segment adjusted EBITDA at a 19% margin (FY25 release). The model here is brand plus cost position. Hammermill is the dominant branded cut-size ream in the channel; the Eastover mill sits low on the cost curve; and distribution runs through office-products retailers, e-commerce, merchants and direct commercial accounts. Pricing is semi-administered, published in trade indices and negotiated around announced increases, which is why the company can describe a 5–8% increase communicated to customers and then report realisation quarter by quarter (Q1 FY26 call). Critically, until this year a meaningful slice of North American volume, about 260,000 tons in 2025, was not made by Sylvamo at all but purchased from International Paper's Riverdale mill under the spin-era offtake. That agreement ended in April 2026 when IP converted Riverdale's No. 16 machine to containerboard, an event that simultaneously cost Sylvamo purchased supply and removed roughly 7% of North American industry capacity, tightening the market it sells into (Aug 2025 joint statement; Q1 FY26 release). The Georgetown offtake, the smaller of the two, had already wound down, with termination charges recognised in early 2025 (Q1 FY25 release).

Latin America is the growth-adjacent piece: $904 million of 2025 sales, 27% of the total, and $192 million of adjusted EBITDA at a 21% margin, down from 23% in 2024. Brazil is one of the few large markets where uncoated freesheet demand is broadly stable, supported by education, government and a printing culture that digitised later; Chamex is the category-defining brand. The three Brazilian mills are integrated into low-cost eucalyptus fibre from owned and managed plantations, giving structural margins the other regions cannot match, and the segment exports into other Latin American countries, the Middle East and Africa when domestic seasonality softens. The mix within the segment matters enormously: domestic Brazilian cut-size is the most profitable ton, exports the least, so first quarters (seasonally weakest at home, heaviest in exports) always screen poorly and fourth quarters well. In 2026 the export book acquired a second function: after United States tariff changes in late February, when a court ruling invalidated the prior tariff framework and a flat 10% rate replaced Brazil's previous 50%, Sylvamo began redirecting Brazilian tons into the US to bridge the Riverdale gap, a swap management expects to save about $20 million of transition cost this year versus the original plan of importing from Europe (Q1 FY26 release).

Europe is the problem: $741 million of 2025 sales, 22% of the total, and negative $78 million of adjusted EBITDA, an 11% negative margin, after a $47 million positive contribution in 2024 (FY25 release). Saillat and Nymölla sell REY, Multicopy and related grades into a market where demand has fallen faster than anywhere else, where cut-size prices exited 2025 a full €100 per tonne below the prior year-end, and where Nordic wood costs stayed stubbornly high. The fourth quarter of 2025 included an $11 million goodwill impairment in the France reporting unit, a small number but an honest signal about the carrying economics (FY25 release). Two price increases have been pushed through in 2026, 4% in the first quarter with roughly half realised by April, and 8% effective May, and pulp prices have started recovering, but the segment lost another $44 million at the operating line in the first quarter alone, worsened by a third-party turbine failure at Saillat and boiler and debarking-drum problems at Nymölla (Q1 FY26 release and call). Europe is where the bull and bear cases diverge most violently, and we treat it accordingly throughout this note.

The revenue model end to end is straightforward: fibre (owned eucalyptus in Brazil, purchased wood in Europe and the US) is pulped, formed into paper on large machines whose economics are almost entirely utilisation-driven, converted into reams, rolls and folio sheets, and sold through a mix of direct commercial relationships, paper merchants, office-products retail and e-commerce. Roughly speaking the company sells on annual and multi-year commercial arrangements repriced around announced increases, so revenue in any quarter is volume shipped times a lagged realisation of announced price, and EBITDA is that revenue less wood, chemicals, energy and freight (all volatile), fixed mill costs (heavily sensitive to downtime, planned or otherwise), and maintenance outage expense that the company schedules and discloses by quarter. There is no subscription, no backlog beyond order books measured in weeks, and no meaningful services attach. It is a manufacturing spread business. What distinguishes a good year from a bad one is the gap between announced-price realisation and input costs, and the number of days the machines run.

Source: FY25 fourth-quarter release, segment schedules. The entire $184m year-on-year decline in total adjusted EBITDA is explained by Europe ($125m swing) and Latin America ($36m), with North America down only $23m despite a $275m revenue decline, evidence of the price discipline the Riverdale supply removal is now reinforcing.

The channel structure completes the picture of how the money is made. In North America, cut-size reams reach the end user through office-products retail and e-commerce, through paper merchants and distributors serving commercial printers and enterprises, and through direct contracts with government and education buyers. The branded ream is the closest thing in this industry to a consumer product: the Hammermill name on the wrapper carries a price premium of several percent over private label for functionally similar sheet, sustained by decades of advertising sediment and by the low absolute stakes for the purchaser, who will not risk a paper jam in the office printer to save forty cents a ream. In Brazil the dynamic is stronger still; Chamex is close to a generic term for copy paper in parts of the market, a brand position built over forty years that a new entrant cannot buy at any tolerable cost. Offset rolls and folio sheets sold to commercial printers, by contrast, are specified on price and delivery alone. The margin difference between the branded cut-size ton and the commodity offset ton is substantial, unquantified in the disclosures, and central to why the two American segments earn what they earn. It is also why the Office Depot channel's change of ownership in December 2025, into the hands of the company's own largest shareholder, is a strategic fact and not a footnote: a meaningful share of Sylvamo's most profitable product line now flows through a retailer controlled by Atlas.

4. Unit economics and key operating metrics

Sylvamo does not disclose shipment volumes, realised prices per ton, or capacity by mill in its quarterly releases, a genuine gap for an outsider trying to model a tonnage business. What it does disclose, with unusual granularity, is the sequential earnings bridge (price/mix, volume, operations, outages, inputs, in dollars) every quarter, plus specific tonnage figures around discrete events. From those disclosures and the segment accounts, the unit economics can be triangulated with reasonable confidence, and we flag every estimate as such.

Start with what is disclosed. Riverdale supplied about 260,000 tons in 2025 and will supply roughly 100,000 in 2026 before ending entirely; the extended Eastover outage costs about 30,000 tons of production this year; European operations are shipping roughly 80,000 tons across the Atlantic to bridge the gap, at a stated $20 million cost to the European segment (about $250 per ton of margin sacrifice); and the net effect is around 55,000 fewer North American tons sold in 2026 (Q4 FY25 call; Q4 FY25 presentation). Nymölla is a 500,000-ton mill (FY23 release). The Eastover optimisation adds 60,000 tons from 2027 (Q1 FY26 call). North American revenue of $1,754 million in 2025, set against total regional volumes that management's bridge implies were on the order of 1.3–1.4 million tons including purchased Riverdale product, suggests average realisation somewhere near $1,250–1,350 per ton across the cut-size, offset and specialty mix; this is our estimate, not a disclosure, but it is consistent with published cut-size transaction prices and the roughly $36 million revenue effect the company attributed to the first-quarter volume decline. On the same arithmetic, the $65 million transition drag equates to roughly $6–7 per share of pre-tax earnings parked in 2026 that returns in 2027 at a rate of about $1.10 per share after tax.

The pricing metrics are the closest thing this business has to a net-revenue-retention figure, and management's realisation disclosure deserves attention because it is testable against trade publications. In the first quarter of 2026: North America announced 5–8% and began realising within that range from March, with the bulk landing in the second quarter; Brazil announced 5% on cut-size for January and realised about two-thirds in the quarter; other Latin American markets announced 7% and realised about a third, with a second 7% effective in the second quarter; Middle East and Africa took 4% fully realised, with a second increase following; Europe announced 4%, saw prices fall in early January before capturing about half of the increase through April, then announced a further 8% effective May (Q1 FY26 call). Realisation of announced increases in a declining-demand industry is only possible when supply leaves faster than demand, which is precisely the North American setup after Riverdale, and much less reliably the European one.

Cost-side operating metrics are equally observable. Planned maintenance outage expense is disclosed and scheduled: 2026 carries a $20 million sequential increase into the second quarter and, unusually, more than half of the full-year outage cost in the fourth quarter because of the 45-day Eastover shutdown, inverting the normal first-half weighting of more than 80% (Q1 FY26 call). Reliability, the unplanned kind, cost about $9 million in the first quarter across power-plant and digester failures at Mogi Guaçu and Luiz Antônio, a third-party turbine trip at Saillat, and boiler and debarking problems at Nymölla, with the Nymölla debarking drum running $1–2 million per quarter of extra chip-purchase cost until a September repair (Q1 FY26 call). War-driven input inflation, energy, chemicals, diesel and ocean freight following the Middle East conflict, adds about $15 million in the second quarter, split roughly $5 million per region, versus $1–2 million in the first (Q1 FY26 call). None of these numbers is individually large against a $3.2 billion revenue base; collectively they explain why a 13–17% margin business is printing 4% quarters.

The flywheel, to the extent a declining commodity business has one, is capital-project driven and quantified by management with unusual precision: the $145 million Eastover programme is guided to more than $50 million of incremental annual adjusted EBITDA, an internal rate of return above 30% (Q2 FY25 release); the woodyard partnership avoids about $75 million of capital over five years; Nymölla carried a $20 million run-rate synergy target achieved through 2024; and the lean transformation launched in Latin America in the first quarter of 2026, extending to North America and headquarters in the second, carries a three-year rollout with unquantified but recurring cost ambitions (Q1 FY26 call). We treat the Eastover $50 million as bankable (it is mechanical: more tons through the lowest-cost mill plus internalised sheeting), the lean savings as optionality, and the European normalisation as the swing factor no company disclosure can settle.

Sources: quarterly earnings releases, FY24 Q4 through FY26 Q1. Q2 2025 net sales derived as the residual of full-year and other reported quarters; Q2 2025 adjusted operating EPS of $0.37 is adjusted operating earnings of $15m over diluted shares. The Europe line is the story of the table: five consecutive loss-making quarters, deepening rather than stabilising, against a North American franchise that even in the transition quarter remained profitable.

A word on the physics behind these numbers, because paper-mill economics reward and punish in ways the income statement only reports after the fact. A modern uncoated freesheet machine is a fixed-cost instrument of extraordinary intensity: it wants to run flat out, every hour, forever, and every unplanned stop destroys margin twice, once through the lost tonnage and once through the restart waste and energy. The first quarter of 2026 was a live demonstration. Four separate incidents, a power-plant and recovery-digester problem at the Mogi Guaçu and Luiz Antônio complex in Brazil, a third-party turbine trip that took down Saillat in France, and boiler and debarking-drum failures at Nymölla, together cost roughly $9m in a single quarter, with the Nymölla items continuing at $1m to $2m a quarter until repairs complete around September. None of these was individually large, and their simultaneity was bad luck; but a system of eight mills, several of them decades old, generates a base rate of such events that the guidance bridge does not carry a line for. We handle this in the model by holding roughly $15m a year of unspecified operational friction against the bridge arithmetic, which is one of the reasons our estimates sit below what the company's own disclosed components would sum to. The other side of the same physics is the prize at Eastover: a machine already among the lowest-cost in the world being tuned to produce 60,000 more tons a year through the same fixed-cost base, which is why a $100m optimisation can credibly carry an internal rate of return above 30% when nothing about the demand environment justifies building anything new.

Sources: Q4 FY25 call and presentation; Q1 FY26 release and call. The one-time cost and transition-drag lines overlap in part; we model the combined 2026 effect, and its 2027 reversal, explicitly in Section 7 rather than double-counting.

5. Industry structure and competitive position

Uncoated freesheet is a product in managed retreat. The grade's largest end uses, office printing, direct mail, forms, envelopes and commercial print, have been losing share to screens for two decades, and no serious participant in the industry pretends otherwise. North American demand has declined at a rate that has averaged somewhere between 3% and 5% a year through the cycle, with a violent air pocket in 2020, a partial restocking recovery in 2021 and 2022, and a return to trend thereafter. Latin American demand is the exception within Sylvamo's footprint: Brazilian and wider regional consumption of cut-size paper has been broadly stable to modestly growing, supported by education spending, government procurement and a consumer base still earlier in the digital transition. Europe sits between the two, with structural decline in demand compounded since 2022 by an energy cost base that reset materially higher after the loss of cheap Russian gas.

Declining demand does not, by itself, condemn an industry to poor economics. What matters is whether capacity leaves faster than demand, and who owns the capacity that remains. Here the 2026 picture is more interesting than the secular narrative suggests. International Paper's conversion of the Riverdale mill in Selma, Alabama to containerboard, the same event that deprives Sylvamo of roughly 160k tons of purchased supply this year, removed about 7% of North American uncoated freesheet industry capacity in one step. Georgetown's exit took further tonnage out before that. Imports into North America, which had filled part of the gap and peaked in June 2025, have declined since, helped by freight costs and by the tariff regime described below. The result is a North American supply picture in 2026 that is tighter, relative to demand, than at any point since the spin, and it is the proximate justification for the 5% to 8% price increases Sylvamo announced from March. Whether those increases hold through the second half is one of the two or three facts on which the whole 2027 earnings recovery rests.

The import mechanism deserves unpacking, because it is the pressure valve that determines how long domestic pricing discipline can last. Uncoated freesheet travels well: it is dense, standardised and sold to distributors and large retailers who will switch supplier for a small landed-cost advantage. When North American prices rise, tonnage from Asia, Brazil and Europe flows in with a lag of roughly two to three quarters, which is the time it takes to book vessels, clear channels and win shelf placements. The 2024 to mid-2025 import wave followed exactly this script, peaking in June 2025 after the prior round of domestic increases, and its subsequent decline reflects three brakes applied at once: the 10% universal tariff, elevated trans-ocean freight, and the fact that the largest single potential importer into the US market is now Sylvamo itself, routing Brazilian and European tonnage through its own brands and its own channel rather than through opportunistic traders. That last point is subtle and matters: the Riverdale gap is being filled substantially by the company's internal system, which means the marginal ton entering North America in 2026 carries Sylvamo's margin discipline rather than a trader's clearing price. It is the closest thing the industry has to vertical control of its own import threat, and it is temporary, lasting only as long as the company has surplus southern and European capacity to route.

It is also worth being precise about what kind of demand is being lost, because the mix within the decline is not neutral for Sylvamo. The fastest-eroding end uses, forms, envelopes, transactional print, are commodity offset grades where brand is irrelevant and price is everything; the slowest-eroding are branded cut-size reams sold through office-products retail, education and government procurement, where Hammermill and Chamex hold their positions. Each year of decline therefore shifts the surviving demand pool slightly toward the segments where Sylvamo is strongest, a mix effect that has quietly supported realised price per ton even in years when list prices went nowhere. The effect is worth perhaps half a point of revenue a year, it is already embedded in the history rather than a hope about the future, and it is one reason the company's revenue decline since the spin has consistently run shallower than the industry's volume decline.

The competitive set is concentrated and rational, which is the industry's redeeming feature. In North America, Sylvamo competes principally with Domtar (private, under Paper Excellence ownership), Packaging Corporation of America's remaining white-paper operation, and imports. Sylvamo's Hammermill, Accent and Springhill brands hold leading positions in branded cut-size, a segment where private-label pressure exists but brand still commands a measurable premium at the point of purchase. In Latin America the company is the clear leader in Brazilian cut-size through Chamex and Chambril, produced from some of the lowest-cost fibre in the world: plantation eucalyptus grown on company and partner land in São Paulo state, on growth cycles roughly a third the length of northern-hemisphere pulpwood. That fibre position is the single most defensible asset in the portfolio and the reason the Latin American segment has produced 20%-plus EBITDA margins through conditions that pushed Europe to losses.

Europe is where the competitive position is weakest. The Saillat mill in France and Nymölla in Sweden are well-invested assets, but they sit in a regional market with excess capacity, demand falling faster than in North America, and an energy and wood cost base that Nordic and Iberian competitors, above all Navigator with its integrated Portuguese pulp position, can better absorb. Navigator earned roughly €145m of net income on €2.0 billion of revenue in 2025 while Sylvamo's European segment lost $112m at the operating line. The gap is not managerial. It is structural: fibre cost, energy cost and freight to the customer. Sylvamo's response, sourcing roughly 80k tons of European production into North America this year and pushing two rounds of European price increases (4% from the first quarter, a further 8% from May), treats the symptom. The May 2026 restructuring conversations around lean deployment address the cost base at the margin. Neither changes the fact that Europe is the segment where a strategic decision, fix, shrink or exit, remains genuinely open.

Tariffs cut both ways for this company and deserve a paragraph of their own. The Supreme Court's invalidation of the IEEPA-based tariff schedule and the substitution of a flat 10% rate on all trading partners in early 2026 was a net positive for Sylvamo: the prior regime had put 50% on Brazilian imports, which would have made the company's most natural response to the Riverdale gap, shipping Brazilian tonnage north, punitively expensive. At 10%, Brazil-to-US sourcing works, and management quantified the saving at roughly $20m against its original transition plan. The same 10% floor on European imports gives modest protection to North American pricing. Tariff policy remains a live political variable, and we treat the current regime as a fact of 2026 rather than a permanent feature.

Where does this leave the competitive position overall? Sylvamo is, by scale, the largest pure-play uncoated freesheet producer in the world, with roughly 2.4 million tons of North American capacity, a Brazilian system that is structurally advantaged on cost, and a European system that is structurally disadvantaged. The company's own framing, that it holds low-cost positions in the Americas and premium brands globally, is accurate as far as it goes. The honest completion of the sentence is that around a fifth of revenue sits in a region where the company earns nothing through the cycle, and that the global grade it serves shrinks every year. A fair reading of the industry is that Sylvamo can generate cash for a long time, that the cash generation requires continuous capacity discipline from competitors it does not control, and that the terminal value of the franchise is genuinely uncertain in a way that the terminal value of, say, a containerboard producer is not.

Margins are segment adjusted EBITDA on segment sales, FY2025, from company reporting. Demand trends are our characterisation of industry data referenced on company calls; precise regional decline rates vary by year and source.

6. Financial analysis: earnings quality, balance sheet and cash flow

Sylvamo's reported numbers are, by the standards of the sector, clean. The gap between GAAP net income and the company's adjusted operating earnings has been modest in most years, and the add-backs are largely items we would ourselves exclude when judging run-rate earning power: discrete tax items, the Georgetown offtake termination charge in early 2025, the $11m France goodwill impairment in the fourth quarter of 2025, and severance associated with the lean programme. Two items deserve scrutiny rather than automatic acceptance. First, stock-based compensation runs at roughly $14m to $18m a year and is not added back to the company's adjusted operating earnings, which is to management's credit; we flag it only because adjusted EBITDA, the metric the company leads with, necessarily excludes it, and a reader moving between the two measures should remember that roughly half a turn of the P/E gap between GAAP and adjusted in a normal year is timing and tax rather than substance. Second, the first quarter of 2026 included a $19m foreign-exchange gain on an intercompany note with the Brazilian business, which the company excluded from adjusted results; the exclusion is correct, and it cuts against the company (adjusted operating EPS of a negative $0.53 was worse than the reported negative $0.08), which we take as evidence that the adjustment framework is applied symmetrically rather than opportunistically.

The reported effective tax rate in the first quarter, 50%, looks alarming and is not. It is the arithmetic of a near-zero pre-tax base interacting with the geographic mix of profit (Brazil taxed, Europe generating losses with limited current benefit) and the excluded FX gain. Management points to an operational rate near 13% for the quarter and a normalised rate in the mid-to-high 20s across the cycle; our model uses 27% from 2027. Nothing in the tax accounts suggests aggressive positioning.

FY2025 figures from the fourth-quarter release; TTM figures are our arithmetic across the four quarters to 31 March 2026 and are approximate where quarterly special-item detail is incomplete.

The cash flow statement tells a harder story than the income statement. Free cash flow of $44m in 2025 against $155m returned to shareholders means the company paid out roughly 350% of the cash it generated, funding the difference from the balance sheet; net debt rose from $599m at the end of 2024 to $707m at the end of 2025 and to $791m by 31 March 2026 after a negative $59m first quarter. Some of that first-quarter outflow is seasonal and some is deliberate: inventories built to $483m ahead of the Riverdale supply cliff and the fourth-quarter Eastover outage, and management guided to a $25m working-capital investment for the year. But the direction is unambiguous. On our 2026 estimates the company will finish the year at roughly $900m of net debt, about two times the depressed 2026 EBITDA and 1.9 times our mid-cycle estimate, from 1.2 times at the end of 2023. The dividend, $72m a year at $0.45 a quarter, is not covered by 2026 free cash flow and is being paid, this year, out of borrowings. Management suspended buybacks after the third quarter of 2025 (none in the fourth quarter of 2025 or the first quarter of 2026, with $150m of authorisation intact) precisely because it can see the same arithmetic.

Working capital merits its own paragraph because 2026 is a year in which the balance sheet is deliberately absorbing operational risk. Inventories of $483m at 31 March represent both finished-goods safety stock built ahead of the loss of Riverdale supply and the pre-positioning required for the fourth-quarter Eastover outage, when the company's largest and lowest-cost North American mill will be down for 45 days while customers still expect delivery. This is inventory as insurance, and the premium is real: roughly $25m of guided working-capital investment for the year, carrying cost on the incremental debt, and obsolescence exposure if demand disappoints while the warehouses are full. The reversal should arrive in 2027 as the system normalises, and our free cash flow estimates credit most of it, but the timing of working-capital releases in this industry has a habit of slipping a quarter or two behind plan. Receivables and payables are unremarkable; the securitisation facility, now extended to 2029, funds a portion of receivables efficiently and its extension removed a quiet refinancing dependency.

Free cash flow quality through the cycle is better than the 2025 and 2026 prints suggest, and worse than the 2022 and 2023 prints suggested. Across the four full years since the spin the company converted roughly a third of adjusted EBITDA into free cash flow on average, a ratio suppressed in the strong years by catch-up capex on assets that International Paper had run for cash before the separation, and flattered in no year by asset sales or working-capital raids. Capex splits roughly $110m of maintenance and regulatory spend, $40m of Brazilian reforestation (which is genuinely maintenance in economic substance, since the plantations are the raw-material inventory of the southern system), and a discretionary high-return layer that management flexes between $40m and $95m depending on the opportunity set. The honest way to read that structure is that roughly $150m of annual capital expenditure is non-negotiable, so every dollar of EBITDA decline below about $400m comes almost entirely out of free cash flow. That operating leverage in reverse is what 2026 is demonstrating, and it is why the difference between our $470m mid-cycle and the bear case's $395m is worth twenty-two dollars a share rather than five.

The balance sheet itself is in reasonable order, and the May 2026 refinancing removed the one near-term pressure point. The term loan F due 2027 was replaced with a new facility maturing 2032, the receivables securitisation was extended to 2029, and the roughly $400m revolver remains substantially undrawn. Gross debt stood at $921m at 31 March 2026 ($155m short-term, $766m long-term) against $130m of cash. There are no covenant issues at current leverage, and no maturity of consequence before the end of the decade. The risk in the balance sheet is not refinancing risk; it is the slower risk that a structurally declining business carries two turns of leverage into whatever demand environment the 2030s bring, with the dividend absorbing $72m a year of flexibility on the way.

Return on invested capital is the metric management is paid on and the one that best captures the swing in the business: 23% in 2024, 12% in 2025, and on our estimates roughly 4% in 2026 before recovering toward the mid-teens in 2027 if the transition-year costs prove as temporary as guided. The three-to-five-year ambition management set out alongside the first-quarter results, more than $300m of annual free cash flow at better than 15% ROIC, implies EBITDA in the $560m to $590m range. We treat that as the bull case, not the base case, for reasons Section 8 makes explicit.

Company reporting; 2023 net debt is stated after the $60m escrow treatment used by the company at the time, and 2023–24 cash figures are approximate. The May 2026 refinancing extended the principal term facility to 2032 and the securitisation to 2029; the multiple shown for 31 March 2026 uses trailing-twelve-month adjusted EBITDA of $387m.

7. Forecasts, 2026e to 2028e

Our forecast treats 2026 as the trough it is guided to be and tests, rather than assumes, the recovery. The revenue build starts from the first quarter's $755m, adds the March North American price increases (5% to 8% announced, of which we credit roughly two thirds realised by the fourth quarter), the Brazilian and other Latin American increases already partly in the run rate, and the two European rounds, and nets off the lost Riverdale tonnage, the 45-day Eastover outage and the net 55k ton North American volume decline the company has quantified. That produces full-year 2026 revenue of $3,205m, down 4.4%, with a quarterly EBITDA shape of roughly $29m, $60m, $128m and $75m: a weak second quarter carrying $20m of sequential outage cost and roughly $15m of war-related input inflation across the three regions, a third quarter that captures the price increases with the mills running, and a fourth quarter suppressed again by Eastover. Full-year adjusted EBITDA of $292m at a 9.1% margin sits below the level implied by simply annualising guidance components, deliberately; the company's transition-year estimates have already been revised once, and the first quarter's $9m of unplanned reliability cost at four separate mills is a reminder that ageing asset bases do not respect bridge charts.

2027 is the year the thesis is decided. We rebuild EBITDA to $470m on revenue of $3,310m: the $65m transition drag reverses, the roughly $85m of one-time cost falls away, the Eastover programme contributes a first tranche of its guided $50m-plus annual benefit with the softwood woodyard line commissioning in the first quarter, and Europe improves from a $78m EBITDA loss toward roughly breakeven on the two 2026 price rounds and the lean programme, without ever reaching the black at the operating line. Against that we assume underlying North American volume resumes its 3% to 4% decline and that perhaps a third of the 2026 price gains erode as the post-Riverdale tightness normalises. The result, a 14.2% margin, is below every year the company reported from 2022 through 2024 and below the since-spin average; we regard it as a conservative mid-cycle rather than a recovery high. 2028 adds modest Eastover completion benefit and normal decline, for $485m of EBITDA on $3,335m of revenue. Free cash flow follows with the usual lag: negative $48m in 2026 on $245m of capex and the working-capital build, then $162m in 2027 and $179m in 2028 as capex steps back toward the $180m to $200m maintenance-plus-reforestation level.

Our estimates. Year-end net debt assumes the dividend is maintained at $1.80 per share throughout and no buybacks resume before 2027. The 2026 leverage spike is an artefact of trough EBITDA against transition-inflated debt; on our mid-cycle EBITDA of ~$470–485m the same debt is roughly 1.9x falling to 1.5x.

The assumptions we would defend hardest, because the valuation depends on them, are three. First, that the 2026 one-time costs are genuinely one-time: the company's own bridge is detailed enough to audit, and the largest components (outage timing, sourcing friction, severance) have identifiable end dates. Second, that North American pricing holds most of its 2026 gains into 2027; the Riverdale capacity removal is permanent and the import response has so far been muted, but this is a commodity grade and discipline is rented, not owned. Third, that Europe stops bleeding without necessarily healing: our 2027 number requires roughly a $75m EBITDA swing from a segment that lost money in nine of the last twelve quarters, achieved mostly through price already announced and cost already being removed. Anyone who finds that third assumption generous should read the bear case in Section 8 and 9, which prices its failure.

By segment, the 2027 rebuild distributes as follows. North America moves from an estimated $270m of adjusted EBITDA in 2026 (depressed by the outage and the transition friction) to roughly $360m, recovering to but not beyond its 2024 level, on the Eastover contribution, full-year price realisation and the restoration of internally supplied volume, net of underlying demand decline. Latin America holds roughly flat at $190m to $200m: the Brazilian cost position does not change, the January and mid-year price increases roughly offset local cost inflation, and the segment's variability is mostly currency, which we do not forecast. Europe improves from an estimated $65m loss in 2026 to a modest loss of $10m to breakeven, the swing that carries the most uncertainty per dollar, as discussed. Corporate and eliminations absorb the remainder. Readers who prefer to stress-test should note the modularity: each $10m of European miss subtracts roughly $0.60 from our DCF value, and each point of North American price that fails to hold into 2027 is worth roughly $17m of EBITDA and $1.00 of value. The forecast is not a black box; it is three regional bets of very different quality stapled together, which is also a fair description of the company.

On the shape within 2026, the second quarter now in progress is the year's operational low point by design: it carries the sequential $20m of planned outage cost, the residual war-related input inflation, the first full quarter without Riverdale tonnage and only partial realisation of the March price increases. Our $60m estimate for the quarter would represent the weakest print since the spin, and we flag it in advance because a company reporting its worst-ever quarter in early August, in line with a bridge it published in April, is a different event from a company missing. The information in the August release will be almost entirely in the price-realisation and Europe lines, not the headline. The third quarter is where the year is won or lost: mills running, prices in, no major outages scheduled, and our $128m estimate implicitly annualising above $500m, which is the first evidence the market will get that the 2027 recovery case is more than a bridge chart.

8. Valuation

We value Sylvamo four ways: against peers on current-year and recovery-year multiples, through an explicit discounted cash flow with a declining terminal assumption, in reverse (asking what the market price already assumes), and through probability-weighted scenarios. The methods disagree in instructive ways, and the disagreement is the analysis.

8.1 The market's own arithmetic

At $37.45 (close of 2 July 2026), the market capitalisation is $1,488m on 39.7m shares, and enterprise value is $2,279m against the 31 March net debt of $791m. On trailing numbers the stock looks inexpensive: 5.9 times trailing adjusted EBITDA of $387m and 14.8 times trailing GAAP earnings. On our 2026 estimates it looks expensive, 7.8 times EBITDA, because the denominator has collapsed. On our 2027 recovery estimates it looks cheap again: 4.85 times EBITDA, 9.0 times earnings, and a free cash flow yield on the market capitalisation of 10.9%. All three statements are true simultaneously. The question a multiple cannot answer is which denominator deserves to be capitalised, and at what rate a shrinking perpetuity should be discounted.

SLVM computed from company reporting and the 2 July 2026 close; peer figures from public market data compiled in June 2026 and, for the sector aggregate, a December 2025 industry compilation showing a trailing EV/EBITDA of about 8.0x against a five-year median near 8.4x. International Paper's earnings multiple is distorted by conversion and restructuring charges; its EBITDA multiple is the informative one. Navigator figures derived from its FY2025 results (revenue €1.97bn, net income €144.7m) and late-June market values.

The peer comparison makes one point cleanly: Sylvamo trades at a persistent, roughly two-to-three turn EBITDA discount to both its diversified former parent and the paper sector aggregate. Some of that discount is deserved. Sylvamo is the purest secular-decline exposure in the group; International Paper's multiple reflects a portfolio now overwhelmingly weighted to packaging, and Navigator's reflects pulp integration and a lower-cost European position. We are not prepared to argue the discount should close. We do note that even holding the discount constant, a 5.5 to 6.0 multiple on 2027 EBITDA of $470m less year-end 2027 net debt implies equity value of roughly $44 to $50 a share, which is the arithmetic that keeps the bull case honest rather than fanciful.

The stock's own multiple history since the spin frames what the market has been willing to pay at different points in the cycle. At the November 2024 high near $91, the shares traded at roughly seven times that year's $632m of EBITDA on an enterprise basis, the market briefly extending mid-cycle credit at the top of the earnings cycle, which is how the most money is lost in commodity equities. At the trough of the 2022 spin-related indiscriminate selling, the multiple touched the low fours on then-current EBITDA. The present 5.9 times trailing sits almost exactly at the midpoint of the since-spin range, on an EBITDA figure that is itself near the midpoint of the since-spin earnings range once 2026's distortions are set aside. In other words, neither the multiple nor the earnings base is at an extreme, which is a further, cruder way of arriving at the same place as the reverse DCF: the market is pricing this company roughly fairly on the information available, and the investor's edge, if any, must come from a differentiated view on Europe or on Atlas, not from the arithmetic everyone can do.

We considered and set aside a sum-of-the-parts approach, and the reasons are themselves informative. A SOTP would value Latin America generously: a structurally advantaged, stable-demand, 21%-margin business with captive fibre could plausibly command seven to eight times its roughly $195m of EBITDA from a strategic or local acquirer, call it $1.4 to $1.6 billion, which is most of the current enterprise value on its own. North America at five times mid-cycle adds perhaps $1.6 to $1.8 billion, and Europe, valued honestly, is worth something between modestly positive (asset value to a consolidator such as the parties that have acquired distressed European graphic-paper capacity this decade) and negative (if exit requires funding closures and social plans, particularly in France, where mill closures carry statutory obligations that have surprised acquirers before). Summed, the parts exceed the whole by 30% to 50%, which is the standard conglomerate arithmetic, and it is exactly the calculation we assume Atlas has done with better information than ours. We decline to underwrite it in the fair value because the discount only closes through a corporate event, the board has armed itself against precisely that event for the next four months, and value that requires someone else's decision to surface belongs in a scenario weight, not a base case. The bull case's $53, at roughly six times mid-cycle EBITDA, is functionally our SOTP expressed with humility.

8.2 Discounted cash flow

Our DCF discounts unlevered free cash flow at an 8.5% weighted average cost of capital: cost of equity of 10.4% (a 4.4% risk-free rate, 0.9 beta against a 5.0% equity risk premium, and a 1.5% size and liquidity premium for a $1.5 billion company with a 31m share float), after-tax cost of debt of 4.4%, weighted 65/35. The explicit period runs 2026 through 2030 with free cash flow to the firm of negative $14m, then $185m, $202m, $206m and $206m. The load-bearing assumption is the terminal growth rate, which we set at negative 0.5%: a permanent, gentle decline that assumes pricing and mix offset most, but not all, of the volume erosion forever. On those inputs enterprise value is $2,187m, equity value $1,396m, and fair value $35.14 per share, with the terminal value contributing 72% of the total. A DCF of a declining business is an exercise in disciplined pessimism, and we would rather present the machinery openly than pretend the answer is more precise than its inputs.

Our model. Using the 40m diluted count rather than basic shares moves the figure by roughly $0.25; we quote the basic-count figure and treat the difference as noise at this level of precision.

Our model; the highlighted row is our base WACC. The grid's message is that no defensible combination of inputs produces a fair value materially above the mid-$40s, and combinations consistent with a faster secular decline sit in the twenties.

8.3 The reverse question

Rather than argue about our assumptions, invert them. At the current enterprise value of $2,279m, an 8.5% WACC and our capex and tax structure, the market is capitalising sustainable adjusted EBITDA of roughly $506m if one accepts our negative 0.5% terminal decline, roughly $489m at zero terminal growth, and roughly $523m to $540m if the decline is steeper at negative 1.0% to negative 1.5%. Every one of those figures sits at or above our mid-cycle estimate of $470m to $485m. The stock, in other words, is not pricing the 2026 trough; it is already pricing a full recovery to mid-cycle and then some. The apparent cheapness of the 4.85 times 2027 multiple is the compensation a rational market demands for holding a melting asset, not a mispricing. This single calculation is the spine of our Hold recommendation.

Our model, holding the enterprise value of $2,279m, an 8.5% WACC and our capital-intensity assumptions fixed while solving for the steady-state EBITDA that justifies the price at each decline rate.

8.4 Scenarios and the weighted answer

The bear case assumes the two fragile legs give way: North American pricing fades through 2027 as import volumes and private-label pressure return, and Europe never reaches breakeven, leaving group EBITDA in the $390m to $400m range and fading at 2% a year. On a matching DCF that is worth $13.28 a share, 65% below the current price, and we emphasise that nothing in that scenario requires a recession, only the failure of two recoveries that have not yet happened. The base case is the $35.14 described above. The bull case assumes Europe normalises to a modest profit, the Eastover benefits land in full, and mid-cycle EBITDA settles near $560m with a flat terminal profile; that is worth $53.04, 42% above the current price, and it is no coincidence that it approximates what a strategic or financial acquirer, Atlas being the obvious candidate, might pay at roughly six times mid-cycle EBITDA. Weighting these 25/50/25 produces $34.15, 8.8% below the 2 July close.

Our model. The asymmetry is worth staring at: the bear case loses 65% while the bull case gains 42%, and the weighted value sits below the market price. A stock on 5.9 times trailing EBITDA can still offer negative expected value when the tail risk is a melting terminal.

Our fair value is $35 on the base case, with a $13 to $53 range, and the shares at $37.45 sit modestly above both the base and the weighted figure. That is the definition of a Hold: not cheap enough to buy against the reverse-DCF evidence, not expensive enough to press against a 4.8% yield, credible 2027 estimates and a 21.5% economic holder with form.

9. The bear case, adjudicated

A serious note on a declining-grade paper company must give the prosecution its full day in court. There is no published short thesis on Sylvamo of the activist-report variety, and short interest is unremarkable at 5.2% of shares outstanding (6.8% of the smaller float, 5.4 days of volume, as of the late-June settlement data), so the bear case must be assembled from first principles. We assemble it in five counts and rule on each.

Count one: the product is dying, and no amount of self-help changes the slope. Largely true, and the most important sentence in this report. North American uncoated freesheet demand declines 3% to 5% a year through the cycle; the company's growth since the spin has come entirely from price and mix, never volume. The defence is not that demand will stabilise but that supply has left faster: the Riverdale conversion alone removed roughly 7% of industry capacity in 2026, and every conversion or closure resets the price floor higher for the survivors. This is the managed-decline playbook that tobacco and directory businesses ran, some profitably and some ruinously. Our ruling: the count stands, and it is why our terminal growth is negative, our multiples discount peers, and our recommendation is not Buy at a 10.9% forward free cash flow yield. The stock's cheapness on 2027 numbers is the market pricing exactly this.

Count two: Europe is not cyclical, it is broken. The strongest count. The segment has lost money at the operating line in most recent periods, culminating in a $112m operating loss in 2025 and a further $44m in the first quarter of 2026 alone; a goodwill impairment has already been taken in France; and the structural cost gap to integrated Iberian and Nordic producers is fibre and energy, which no lean programme closes. The defence rests on price (two rounds in 2026, with the industry exiting 2025 roughly €100 a ton below the prior year and now recovering), on the Nymölla boiler and debarking repairs completing around September, and on the sourcing of 80k tons into North America improving mill loading. Our ruling: genuinely open. Our base case takes Europe only to approximate breakeven, never to profit, and our bear case, weighted at a full 25%, assumes the segment stays loss-making and drags the group to $13. Investors who believe Europe is unfixable should not own the stock at any price near the current one; the reverse DCF shows the market is not paying them for that risk.

Count three: the 2026 bridge is too convenient; one-time costs at this company recur annually. Partially true. The company has now had transition charges, offtake terminations, reliability incidents and severance in successive years, and the first quarter's $9m of unplanned outages across four mills suggests an asset base whose maintenance capital may be structurally understated. The defence is that the largest 2026 items have contractual end dates (the Riverdale agreement is over; the Eastover outage has a calendar) and that the company revised its transition drag estimate in the favourable direction, from $85m to $65m, which convenient bridges rarely do. Our ruling: we haircut rather than dismiss. Our 2027 EBITDA of $470m sits below the arithmetic the bridge implies (2025's $448m plus $65m drag reversal plus $50m Eastover plus European pricing would exceed $560m), precisely because we assume some of the one-time layer is in fact recurring friction.

Count four: the dividend is uncovered and the buyback is dead; capital returns were a fair-weather promise. True as to 2026 fact, unproven as to policy. Free cash flow this year will not cover the $72m dividend, buybacks stopped after the third quarter of 2025 with $150m of authorisation unused, and net debt has risen $192m in five quarters. The defence: paying a maintained dividend through a guided trough year out of a just-refinanced balance sheet at two times trailing EBITDA is not distress, it is a choice, and the 2027 free cash flow estimate covers the dividend 2.3 times. Our ruling: the count fails on solvency but succeeds as a discipline signal. A management that returned 350% of free cash flow in 2025 and then stopped entirely has told you its own visibility is limited; we take the halt as honest and the history as a caution.

Count five: the rights plan protects the board, not the shareholders, and the Atlas situation will end in dilution of governance rather than a premium. Unresolved by construction. The board adopted a one-year rights plan within hours of Atlas terminating the cooperation agreement in November 2025, with a 15% trigger set conspicuously just below Atlas's 16% beneficial position, and the May 2026 annual meeting re-elected the incumbent slate unopposed. A pill that expires 9 November 2026 is not entrenchment in perpetuity, but it did remove the near-term possibility of the one event, a control contest or bid, that the bull case partly relies on. The defence is that the board is buying time for the Eastover and Europe evidence to arrive before any negotiation, which is what a board maximising value against a 21.5% holder with complete information about the assets should do. Our ruling: we assign the Atlas situation option value in the bull case and zero in the base, and we note the pill expiry as the single most interesting date on the calendar.

Count six: the shareholder register is deteriorating around the story. A minor count, but real. The stock was removed from the Russell growth indices at the June 2026 reconstitution, mechanically forcing passive selling into a 31m share float in the final week of June, which likely contributed to the pressure into the $35.53 fifty-two-week low. Sell-side coverage is thin, five firms of varying reach, with published targets scattered from $44 to $75 over the past year, a dispersion that itself measures how little consensus exists about mid-cycle earnings. Average daily volume near 380,000 shares means an institutional position of any size takes weeks to build or exit. The defence is that all of this is liquidity, not value: index flows wash through, and a thin float cuts both ways when a 21.5% holder or the company's own $150m authorisation eventually returns to the bid. Our ruling: no effect on fair value, a modest widening of the range around it, and a practical note that limit orders are not optional in this name.

10. Catalysts and the calendar

The catalyst path is unusually legible for the next twelve months, which is part of why we are comfortable holding rather than avoiding a name whose base case sits below the market price. The dated events are set out below; the undated ones, a European price announcement holding or failing, an Atlas filing amendment, a strategic decision on the European system, matter more.

Company disclosures and our assessment. The August results date is indicated by the company's reporting pattern; it had not been formally confirmed at the time of writing.

Of the undated catalysts, three carry more weight than anything on the calendar. The first is the European strategic question. Management has so far framed Europe as a turnaround, but the France impairment, the lean programme's sequencing and the routing of European tonnage to North America are all consistent with a portfolio being groomed for a decision, and the investor day is the natural venue to make one. An announced exit or structural downsizing of the European system, even at a cash cost, would in our estimation add several dollars to the stock by removing the bear case's largest limb; the market pays quickly for the amputation of a loss-maker. The second is any amendment to Atlas's ownership filings: a move in the derivative position, a conversion of economic to voting exposure, or the first public letter would each reprice the situation within hours, and the direction would depend entirely on whether the board's posture reads as engagement or entrenchment. The third is the behaviour of North American list prices in the trade press through the autumn, which will tell holders whether the March increases held before the company confirms it in February; in this industry the pricing data leads the earnings by a full quarter, and the diligent get paid for reading it.

11. Risks, ranked

In order of expected impact on our fair value rather than probability alone. First, European normalisation failing: the difference between our base and bear cases is mostly Europe, and the evidence to date favours the bear. Second, North American price erosion in 2027: the post-Riverdale tightness is real but rentable, and a resumption of import growth or a demand step-down would unwind the single largest positive in our bridge. Third, execution at Eastover: a $145m programme with a 45-day outage in a mill that supplies the company's most profitable region concentrates operational risk into one quarter; the first quarter's reliability record across four other mills does not reassure. Fourth, capital allocation drift: a dividend maintained too long against the 2026 cash burn, or buybacks resumed prematurely to defend the stock against Atlas, would convert balance-sheet strength into a governance instrument. Fifth, the demand slope itself steepening: our negative 0.5% terminal is a judgement, not a datum, and the sensitivity grid shows a move to negative 2% is worth roughly $9 a share. Sixth, currency and country: roughly 27% of revenue is Latin American, the Brazilian real drives both costs and the intercompany note that produced the first quarter's excluded FX gain, and tariff policy has already changed once this year. Seventh, and cutting the other way, the principal upside risk to a Hold: Atlas, unconstrained after 9 November, moving on a company whose bull-case takeout arithmetic starts $15 above the current price.

Two of these risks interact in a way that deserves explicit statement. The Eastover outage and the balance-sheet trajectory converge in the fourth quarter of 2026: the company will run its largest cash-consuming quarter of the year, outage costs plus the working-capital peak plus the dividend, at the precise moment its EBITDA is most depressed and, coincidentally, in the weeks around the rights-plan expiry. Nothing about that convergence threatens solvency; the refinanced facilities carry it comfortably. But it does mean the company will be negotiating whatever it negotiates with Atlas from its position of minimum optical strength, with trailing numbers at their worst and leverage at its screen-shot peak of roughly three times trough EBITDA. A holder with Atlas's sophistication understands the difference between trough optics and mid-cycle reality perfectly well, which is arguably the point: if an approach comes, the fourth quarter is when the price asked would be lowest and the board's public defence hardest to make. We flag the window rather than predict the event. The currency risk also deserves one further sentence: because Brazilian costs are in reais while much of the Latin American segment's pricing follows dollar-linked pulp and paper benchmarks, a sharp real appreciation compresses the segment's margin from both directions at once, and the segment's $36m EBITDA decline in 2025 owed as much to currency as to volume. We model currency flat and disclose the sensitivity rather than pretending to a view.

12. Management, incentives, ownership and capital allocation

The leadership transition completed cleanly on paper. Jean-Michel Ribiéras, who led the company from the spin, retired on 31 December 2025; John Sims, the long-serving chief operating officer, became chief executive on 1 January 2026, with David Petratis taking the chairmanship and Don Devlin continuing as chief financial officer. Sims is an operator by background, which fits a year whose entire content is operational execution, and his first two earnings calls have been notable for specificity: the transition bridge is quantified line by line, the reliability failures were disclosed mill by mill with repair dates, and the three-to-five-year framework (more than $300m of free cash flow, better than 15% ROIC) was stated as a commitment to be tested at the coming investor day rather than left as atmosphere. We hold the view, unfashionable but repeatedly vindicated in this sector, that in a declining industry the quality question is not whether management can grow but whether it will tell the truth about shrinking. The evidence so far is that this one does: the Q1 FX gain excluded against the company's own optics, the transition estimate revised, the buyback halted rather than defended.

Incentives are adequately aligned with the shape of the business. The proxy structure weights annual bonuses to free cash flow and adjusted EBITDA margin (the company calculated 13.7% for compensation purposes in the relevant year) and longer-term awards to return on invested capital (18.7% on the compensation calculation) and relative shareholder return. Paying operators of a declining asset on cash and returns rather than on revenue growth is the correct design; the risk in such schemes, that maintenance is deferred to flatter near-term cash metrics, is exactly the risk the reliability incidents keep raising, and it is why we watch maintenance capex as closely as any income-statement line. Insider ownership is thin at 0.92%; institutions hold 79%; and one holder dominates the register.

That holder is Atlas Holdings, and no assessment of this company's governance is complete without it. Atlas is a Greenwich-based private equity firm with an operating roster concentrated in paper, packaging and industrial businesses, and its December 2025 completion of the roughly $1 billion Office Depot take-private gave it control of a major retail channel for the very cut-size product Sylvamo makes. Atlas accumulated its position rapidly in 2022, took two board seats under a 2023 cooperation agreement, and then in November 2025 withdrew its directors and terminated the agreement, freeing itself of the standstill. The board's same-day response was the one-year rights plan with its 15% trigger. Atlas holds 16% beneficially with a further 5.5% in derivative form, 21.5% economically. It has made no offer, filed no proxy contest, and said little. A firm with Atlas's paper-sector operating depth does not hold a fifth of a paper company's economics passively as a portfolio yield position. We do not underwrite an outcome; we observe that every strategic decision the company makes in the next eighteen months, on Europe, on the dividend, on the pill, will be made in Atlas's presence.

The board that will make those decisions was re-set at the May 2026 annual meeting, where shareholders elected a seven-member slate unopposed: Breves, Bruce, Desmond, Johnson, Petratis, Sims and Zallie. The composition is competent and conventional, industrial operating experience, finance, and consumer-adjacent backgrounds, and its most notable feature is what it lacks: any Atlas presence, for the first time since 2023, and any director whose seat depends on the largest economic holder's goodwill. A seven-person board with a chief executive five months into the job, defending a pill against a 21.5% holder while deciding the fate of a loss-making continental segment, is carrying an unusually heavy agenda for a $1.5 billion company. We make no criticism of the individuals; we note that governance risk at Sylvamo is concentrated not in any structural defect but in the sheer number of consequential, irreversible decisions scheduled for a single eighteen-month window, and that boards under siege have historically erred toward decisions that preserve independence rather than decisions that maximise the sale value of the evidence.

It is also fair to credit the record of the management system this team inherits. In four and a half years as a public company, Sylvamo has generated $2.5 billion of cumulative adjusted EBITDA and more than $960m of free cash flow, exited Russia at a real but contained cost within a year of the invasion, bought Nymölla for $167m at what was then a low multiple of its normalised earnings, returned over half a billion dollars, and never missed a debt covenant, cut the dividend or restated a number. The 2025 to 2026 deterioration is real and the reliability record is a legitimate worry, but the base rates from this organisation's own history favour execution over failure. Our base case, which credits the transition bridge at roughly 80 cents on the dollar, is calibrated to exactly that record: better than a coin flip, worse than the guidance.

Company reporting; annual splits partly our reconstruction and marked approximate. The dividend was initiated at $0.1125 a quarter in late 2022 and has been raised in steps to the current $0.45; 2022 buybacks of roughly $80m retired about 1.56m shares in the fourth quarter. No buybacks have been executed since the third quarter of 2025 despite $150m of standing authorisation.

13. Bull and bear, side by side

14. What we would watch, and position notes

For holders, we would stay put and mark three tests. The August second-quarter print must show North American price realisation and a European operating loss visibly narrower than the first quarter's $44m; a second consecutive quarter of that magnitude would move our Europe assumption toward the bear case and our fair value toward the twenties. The fourth-quarter Eastover outage must complete on its 45-day calendar. And the 9 November pill expiry must resolve into something other than a routine renewal; a renewal without a stated engagement process would tell us the board prefers the siege to the negotiation, and we would reassess the governance discount accordingly.

For prospective buyers, the entry arithmetic is mechanical. At our $35 base fair value the stock offers nothing at $37.45; below roughly $30 the weighted scenario value turns positive against the price even on our cautious Europe assumption, and the dividend yield approaches 6% with 2027 coverage above two times. We would rather pay $31 after bad second-quarter news, with the bear partially priced, than $37 before it. For those inclined to the bull case on Europe or on Atlas, we would still size the position to the bear case's negative 65%, because the distinguishing feature of this name is not its upside, which is ordinary, but the width of its distribution, which is not.

For income-oriented holders, who plausibly own much of the non-Atlas register at a 4.8% yield, the honest framing is that this is a dividend paid on the balance sheet's credit for one more year, then on cash flow's if 2027 arrives as modelled. The board's decision to maintain the rate through the trough while halting buybacks tells you the dividend ranks first among discretionary uses, and the 2027 coverage of 2.3 times on our numbers restores comfortable headroom; but a dividend whose coverage depends on a recovery not yet delivered is a different instrument from one covered at the trough, and it should be sized in an income portfolio accordingly. A cut is not our expectation in any scenario short of the bear case persisting into 2027, at which point a board watching net debt pass $1 billion against fading EBITDA would have every justification. In the bear case the stock's fall to the teens would owe as much to the income register's forced exit as to the earnings themselves, which is how a 65% downside scenario becomes mechanically plausible for a company that would still, even then, be generating positive free cash flow after normal-year capex. Distribution risk, in both senses of the word, is the reason our scenario weights are not more generous to the middle.

We see no case for a short position despite the negative weighted return: the borrow earns a 4.8% dividend against you, the reverse DCF overhang is a slow-burn argument rather than a catalyst, and an unconstrained 21.5% strategic holder is precisely the situation in which shorts are carried out on stretchers. Reduce would require either a demanding valuation or a probable negative catalyst; we find neither. Hold, market-weight, review at the second-quarter results.

15. Sources and methodology

This report draws on Sylvamo's annual reports on Form 10-K for 2022 through 2025 and the quarterly report for the period ended 31 March 2026; the fourth-quarter 2025 and first-quarter 2026 earnings releases and presentations; management commentary from the fourth-quarter 2025 and first-quarter 2026 earnings calls; the 2026 proxy statement and annual-meeting results; the November 2025 rights-plan adoption and related beneficial-ownership filings; and company announcements regarding the May 2026 refinancing, the Eastover investment programme and the Nymölla acquisition. Peer figures derive from International Paper and Navigator Company public reporting and market data compiled in late June 2026, and the sector multiple aggregate from a December 2025 industry compilation. Market prices, share counts, short interest and float are from public market data as displayed on 2 and 3 July 2026; the share price of $37.45 is the closing price of 2 July 2026, the last full trading session before publication, and is dated separately from the report date of 5 July 2026 throughout. All forecasts, the DCF, reverse DCF and scenario values are our own. Figures are in US dollars in millions unless stated; adjusted measures are reconciled to GAAP in Exhibit 6; approximations are marked. Where quarterly disclosure was incomplete, notably second-quarter 2025 segment detail, derived figures are identified as such in the exhibits.

16. Disclosures

Thracian Research is an independent research publication. This report is provided for information only and does not constitute investment advice, an offer, or a solicitation to buy or sell any security. The author may hold positions in securities discussed. Estimates, fair values and scenario weights reflect judgement applied to public information believed reliable but not guaranteed; they will be wrong in ways that matter and should be treated as a framework, not a forecast. Recommendation vocabulary: Buy/overweight, Speculative Buy, Hold/market-weight, Reduce/underweight, Avoid. Fair values are twelve-month horizons unless stated. Data errors brought to our attention will be corrected in subsequent versions.