SXC · Suncoke Energy Inc · Stock research

Current Thesis

Three weeks of tape have settled the question. SXC closed near $8.39 on 2026-06-27 and near $8.37 on 2026-07-17 a 0.2% move over roughly fifteen sessions, with the 2026-07-17 session ranging $8.23–$8.55. The coal-policy sympathy leg that carried the stock to a 52-week high of $9.74 into the 2026-06-03 $700M Defense Production Act package is not consolidating for a second leg; it has stopped moving. That is the honest read on a name whose narrative was never tethered to its own P&L: the DPA money funds thermal power plants and an Oakland export terminal, while SunCoke sells metallurgical coke into steel blast furnaces and moves coal through Convent Marine. The headline beta mean-reverted, and there was nothing underneath it.

What has changed materially since the last review is the clock. On 2026-07-16 the company confirmed Q2 2026 results for Thursday 2026-07-30 before the open, with an 11:00am ET call five days earlier than the prior ~08-04 estimate. The single dated variable in this name is now roughly eight trading days out, which puts the whole position of the stock into a binary print rather than a trend. Market cap sits at ~$711M, trailing P/E is negative at -10.87 after the Q1 loss, and the dividend yields ~5.68%. Sell-side price targets cluster at $9.00–$9.50 against $8.37 spot single-digit percentage upside on thin coverage, which is not the shape of a momentum trade.

Bullish and bearish views on Suncoke Energy Inc

The model's bull view on Suncoke Energy Inc (SXC), in brief: Guidance intact: FY2026 Adjusted EBITDA reaffirmed at $230–250M and net income $18–36M on the 2026-04-30 Q1 release. The bear view: No narrative velocity. A stock that moves 0.2% in three weeks while its sector story fades is not a momentum candidate at any conviction level. The June policy spike produced no SXC-specific follow-through and management said as much on the Q1 call, attributing 2026 terminal… Both cases follow in full.

Bull Case

• Guidance intact: FY2026 Adjusted EBITDA reaffirmed at $230–250M and net income $18–36M on the 2026-04-30 Q1 release. Against a $711M market cap, that EBITDA run-rate is the reason the $8 shelf has held rather than broken.
• Top line beat in Q1: revenue $455.1M vs ~$422.3M consensus (+$32.8M) despite winter weather, the Middletown turbine failure, and the Haverhill I shutdown.
• Granite City volume is contracted and quantified: the 2026-01-22 extension covers ~590,000 tons of met coke to US Steel through 2026-12-31, with minimum steam supply obligations maintained. US Steel announced a Granite City Blast Furnace B restart in December 2025 on higher demand.
• Cleveland-Cliffs Haverhill: 500k tons/yr met coke on a three-year term from 2026-01-01 a second contracted volume block extending past the Granite City decision point.
• Balance sheet and cash conversion: Q1 operating cash flow $72.7M, quarter-end liquidity $262M, 27th consecutive $0.12 quarterly dividend paid 2026-06-02.
• Phoenix Global: the $325M acquisition closed 2025-08-01 adds EAF mill-services revenue, giving the company a revenue line that grows as blast furnaces retire rather than shrinking with them.

Bear Case

• No narrative velocity. A stock that moves 0.2% in three weeks while its sector story fades is not a momentum candidate at any conviction level. The June policy spike produced no SXC-specific follow-through and management said as much on the Q1 call, attributing 2026 terminal throughput to international coal demand rather than Section 303 or regulatory policy.
• Q1 earnings miss: EPS -$0.05 vs +$0.07 consensus, versus $0.20 EPS a year earlier; Adjusted EBITDA slipped to $56.5M from $59.8M. Trailing P/E of -10.87 reflects a company currently not earning.
• Secular met-coke decay: the migration of US steelmaking from blast furnace to scrap-fed EAF structurally shrinks the addressable market for SunCoke's core product. Phoenix Global exists because of this.
• 2H26 contract overhang: the Granite City agreement terminates 2026-12-31. Whether the ~590k-ton block renews or winds down is the largest single unhedged variable in the FY27 model, and it will not be resolved on this print.
• Binary risk into 2026-07-30. A name with no trend going into a pre-market print offers the downside of an earnings gap without the compensation of an existing move to ride.
• Sell-side ceiling: $9.00–$9.50 targets and a Hold consensus (as of 2026-06-29) put analyst-implied upside at roughly 8–13%, with coverage thin enough that a single revision swings the average.

Setup & Price Structure

The structure is a base, not a launch pad. Price has been pinned to the ~$8 shelf that the June spike launched from, roughly 14% below the $9.74 high and about 52% above the $5.52 52-week low. RSI has fully cooled from the ~80 reading that got the name flagged as overbought on 2026-05-26 and is now unremarkable. The rising 20-EMA has caught up to price rather than price pulling back to it, which compresses the risk band: $7.90 now sits close enough that a single bad session resolves the question. Volume has thinned through July.

On the beginner-trap matrix, this name is nowhere near peak retail sentiment and is not stretched above its moving averages those traps are cleared. The live trap is different: a 5.68% yield and a contracted volume book create the temptation to hold or accumulate a going-nowhere cyclical through a binary print because the story sounds defensible. Four years of that dividend is erased by one 20% drawdown. The second trap is the calendar the print is ~8 trading days out and the blackout window for fresh entries opens around 2026-07-27.

Catalyst Calendar (next 30 days)

• 2026-07-30 (confirmed, announced 2026-07-16): Q2 2026 results before the NYSE open; earnings call 11:00am ET. The only dated catalyst in the window. Watch the FY26 Adjusted EBITDA guide ($230–250M) for a cut, hedge, or reaffirm, and listen for Convent Marine throughput commentary tied to international coal demand.
• ~2026-07-30, same release: Q3 dividend declaration the 28th consecutive $0.12 quarterly payout is the base case; any deviation would be a genuine signal.
• ~2026-07-27: the practical entry blackout begins, three trading days ahead of the print.
• Undated, 2H26: US Steel decision on the Granite City cokemaking agreement past 2026-12-31. Any commentary on the 07-30 call is the first real read on ~590k tons of FY27 volume.

What Would Change Our Mind

The constructive case requires an SXC-specific driver rather than sector beta. Concretely: a 2026-07-30 print that raises the FY26 Adjusted EBITDA guide above $250M, returns EPS to positive, and pairs it with Convent Marine throughput data confirming international coal demand is translating into terminal volume. An early Granite City renewal announcement covering 2027 would remove the largest structural overhang and justify a re-rating toward the $9.50 upper target. On the other side, a daily close below $7.90 loses the shelf and the 20-EMA together and confirms the June move round-tripped in full. A guide cut on 07-30 with no Granite City visibility turns a cheap cyclical into a declining one, and the yield stops being a reason to own it.

Correlation Notes

SXC trades as a derivative of US blast-furnace steel utilization and seaborne coal, not as an independent story. Primary correlations run to US Steel (Granite City volume, blast furnace restarts) and Cleveland-Cliffs (Haverhill contract), with a second-order tie to seaborne met coal pricing and Atlantic-basin export rates through Convent Marine. It reacts to coal-policy headlines with high beta and zero durability the 2026-06-03 DPA episode is the clean example, a 14% round trip on a package that funds none of its assets. Correlation to broad energy is misleading: thermal-coal news moves the stock, thermal-coal economics do not touch its earnings. The Phoenix Global mill-services line introduces a small offsetting exposure to EAF operators, meaning the same secular shift that erodes coke demand modestly feeds the services segment. Against index beta the name is low-correlation and low-liquidity, which cuts both ways around a print.