Graphic Packaging filed its 10-K on March 2. It disclosed an ICFR adverse opinion from PwC, a covenant amendment jacked from 4.25x to 5.0x four days before filing, and confirmed that 2026 will be the worst earnings year in recent memory. The stock is down 59% in twelve months.

Two days later, the new CEO bought $501,000 of stock. A director added $199,000. Open market purchases at 52-week lows with perfect information on every skeleton in the closet.

That's either insane or the most informative signal in packaging right now.

What the Filing Actually Says

The 10-K is a confession document. Former senior management hid capital expenditure overruns from the Board on the Waco greenfield facility — total cost $1.67 billion against an original budget that management apparently preferred not to discuss. PwC issued an adverse opinion on internal controls. The culprits are gone. An interim CFO signed the filing. No permanent replacement named.

The covenant amendment is the tell: management obtained a waiver on February 26 — four days before filing — raising the maximum leverage ratio to 5.0x through December 2026. At year-end 2025, GPK was at 3.63x. In compliance. They sought the waiver anyway because they know 2026 EBITDA is going off a cliff.

How far off? Adjusted EBITDA guided $1.05-1.25 billion, down from $1.54 billion. The drivers: $260 million in production curtailments to work down bloated inventory, $100 million in incentive compensation restoration (from literally zero in 2025 — nobody earned their bonus), and continued pricing pressure in bleached paperboard. Peak leverage will hit 4.7-4.9x in H1 2026.

The Capex Cliff Is Real

Here's the number that matters: capital expenditure drops from $935 million to $485 million in 2026. That's $450 million of free cash flow appearing because GPK finally stopped building things.

The FCF trajectory:

Three years of negative free cash flow, and the company is about to inflect hard. Management is guiding $700-800 million in adjusted FCF. On a $3.3 billion market cap, that's a 21-24% adjusted FCF yield. At trough earnings.

Even if you haircut management's adjusted number aggressively — say operating cash flow drops $300 million from curtailments and comp restoration — you still get $540 million minus $485 million capex = $55 million positive GAAP FCF. The capex reduction dwarfs any plausible operating decline.

Bleached Is Sector, Not Alpha

New CEO Rietbroek said the bleached paperboard segment "doesn't earn good return on capital" and is "a little bit below cost of capital." Former CEO Doss said the same thing three months earlier on the Q3 call — recycled can "replace more expensive bleached paperboard." Two consecutive CEOs flagging the same asset as non-core over six months.

Sounds like alpha. It's not.

Cross-ticker corroboration shows industry-wide SBS overcapacity: Smurfit WestRock permanently closed 127,000 tons at La Tuque in February. Clearwater Paper — the buyer of GPK's Augusta mill for $711 million in May 2024 — wrote down 100% of Augusta goodwill ($48 million), is guiding Q1 to breakeven, and the CEO said margins are "not sustainable." Five hundred thousand tons of new SBS capacity flooded the market in 2025. Operating rates are running around 80% against a 90% norm.

Everyone in bleached paperboard is dying. GPK's remaining Texarkana mill is a divestiture candidate, but the buyer pool is thin — CLW can't bid (financially unable), SMFT is closing capacity not buying it. Augusta sold for $711 million during better market conditions. Texarkana might sell for less, or not at all.

The Covenant Is Idio

Here's what IS company-specific: the leverage stress. Peer comparison tells the story.

GPK is down 59% while PKG is up 10% and Sonoco is up 17%. This is not a packaging sector selloff. Factor regression confirms: 79-86% idiosyncratic variance across three model specifications. The decline is company-specific — driven by the Waco scandal, governance failure, and resulting leverage stress that no peer shares.

The ICFR adverse opinion is also unique. IP has a DS Smith material weakness, but that's inherited IT controls from a UK acquisition — standard integration risk. GPK's is "former senior management hid capex overruns from the Board." That's a different species of problem entirely.

The Insider Signal

Robbert Rietbroek has been CEO for about 30 days. He's read every document. He sat in the room when they negotiated the covenant waiver. He signed off on the ICFR disclosure. He knows the Waco cost overrun number, the inventory bloat, the bleached market collapse, the Pomerantz investigation.

He bought $501,000 of stock at $11.32.

Director Venturelli added $199,000 the same day. Both Form 4 code P — open market purchases, not grants, not exercises.

New CEO insider buying isn't unusual per se. But the size ($700K combined), the timing (two days after the worst 10-K the company has filed in years), and the price (3% of 52-week range) make this a high-information signal. He's not buying to signal alignment. He's buying because he sees the other side of the trough.

The Leverage Path

The math works if you trace it forward:

Year-end 2025: $5.6B debt / $1.54B EBITDA = 3.63x

2026 trough: $5.1B debt / $1.15B EBITDA = 4.43x (peak ≈4.8x in Q2)
  Covenant ceiling: 5.0x → uncomfortable but compliant

2027 recovery: $4.6B debt / $1.4B EBITDA = 3.3x
  Curtailments reverse (+$260M), Waco at run-rate
  Investment-grade path restored

The covenant waiver is a bridge, not a permanent condition. It steps to 4.75x by mid-2027, then back to 4.25x. The company is paying down $500 million of debt in 2026 from the FCF inflection. By 2027, if EBITDA recovers to pro forma levels ($1.35-1.45 billion), leverage returns to low-3x. Credit agencies gave stable outlooks — they're modeling the same recovery.

What Could Kill It

The bear case isn't complicated: Waco underperforms at full run-rate, paperboard pricing stays depressed through 2027, curtailments don't actually reverse, and leverage sticks at 5.0x. Goodwill impairment hits — PwC flagged $1.04 billion across Paperboard Manufacturing and International as Critical Audit Matters with slim headroom. The ICFR material weakness doesn't get remediated, eroding lender confidence. Stock goes to $9.

I'd put 30% on this. The structural capex reduction is physics, not hope — but the operating leverage works both ways, and a company that just disclosed its former management committed fraud hasn't exactly earned the benefit of the doubt.

Sizing the Setup

Current: $11.22. 7.5x trough P/E. RSI 32. 14% short interest. Options P/C ratio 0.14 (calls dominate). Street consensus: neutral. Zero strong buys. Wells Fargo underweight at $11.

EV-weighted target: $16.38. That's 46% upside over 18 months. Risk/reward is 3:1 at current price.

Forward alpha after stripping sector exposure (79% idio), applying 55% conviction (doorway state — the trough is confirmed but the recovery isn't): 10.8% annualized. Moderate. This isn't a screaming table-pound. It's a governance-reset-at-trough where the CEO is buying with perfect information and the structural FCF inflection is confirmed by hard capex numbers.

What Resolves It

• Q1 2026 earnings (April 30): Waco ramp progress, actual capex run-rate, CEO commentary on portfolio review
• ICFR remediation: Next 10-Q filing. If remediated, governance discount narrows
• Permanent CFO hire: Signals bench depth. Interim CFO is a yellow flag
• Texarkana: Any advisor retention or process announcement
• Proxy filing (April 2026): Rietbroek's comp structure — is he incentivized on ROIC/FCF or revenue growth?

The CEO has told you what he thinks with his wallet. The question is whether the market has already priced the trough (at 7.5x P/E and 59% down, arguably yes) or whether 2026 gets worse before it gets better (at 4.8x peak leverage with an ICFR adverse opinion, arguably also yes).

Doorway state. Both interpretations fit. The CEO's $500K bet says which side he's on.

Evidence