Purdue's Bankruptcy Wasn't Accountability. It Was Almost a Firewall.

On September 15, 2019, Purdue Pharma walked into a federal courthouse in White Plains, New York, and filed for bankruptcy.¹ To most people this looked like the end of the road — the company that built OxyContin into a billion-dollar drug finally facing the music. It was not the end of anything. It was the opening move of a six-year maneuver, and the goal of that maneuver was never to surrender. It was to take a family that had withdrawn roughly $11 billion from the company and wrap them, by way of a corporation's bankruptcy, in a shield the family had never qualified for itself.⁵

The official story is that Purdue's bankruptcy was the Sackler accountability moment. It was nearly the opposite — an attempt to convert a corporate filing into a permanent liability firewall for the owners, and it almost worked. What stopped it was not justice catching up. It was a single sentence of statutory reading, decided 5 to 4.

The drug that made the firewall worth building

Understand the size of what was being protected. OxyContin went from about $48 million in revenue in 1996 to $1.1 billion by 2000, and a contemporaneous SEC filing logged U.S. sales of $1.182 billion for the twelve months ending October 2003.⁸ This was not an incidental product. It was the engine, and for two decades the family that owned the engine kept the cash flowing outward. The Supreme Court would later note, citing the bankruptcy court's own record, that the Sacklers withdrew approximately $11 billion from Purdue before it filed.⁵ When you have moved that much money out of a company over many years, the company's eventual bankruptcy is not a threat to you. It is a tool — if you can find a way to make its protections your own.

The trick at the center: a shield the family hadn't earned

Here is the mechanism almost everyone misses. When a company files Chapter 11, the company gets a discharge — its debts are resolved and creditors can no longer chase it. That protection belongs to the entity in bankruptcy. The Sacklers, as individuals, were not in bankruptcy. They wanted out from under thousands of opioid lawsuits anyway. So the plan proposed something audacious: the family would return $4.325 billion to the estate, and in exchange every opioid claim against the family — including claims from people who never agreed to it — would be wiped out.⁵ That is the move. Return a fraction of what you took, and buy a release no court had ever required you to qualify for, paid for with a corporation's bankruptcy rather than your own. It is the equivalent of someone else declaring bankruptcy and you getting the fresh start.

Dissenting states blew up that first version, which had valued the family's contribution near $4.3 billion. So in 2022 the number was renegotiated upward to roughly $6 billion — more money, but the same architecture underneath. The dollar figure was the negotiation everyone watched. The structure was the thing that actually mattered, and the structure never changed: a nonconsensual release for non-bankrupt people. The objectors took the question all the way up.

The five votes that read the statute literally

On June 27, 2024, the Supreme Court ruled 5 to 4 in Harrington v. Purdue Pharma that the Bankruptcy Code does not authorize a Chapter 11 plan to discharge claims against a nondebtor without the consent of the affected claimants.⁴ Notice what the Court did not do. It did not rule on whether the Sacklers were monsters, whether $6 billion was too little, or whether the deal was fair. It answered a narrow plumbing question: does the statute give a court the power to do this? It does not. The entire firewall rested on a power the Bankruptcy Code never granted, and once five justices said so plainly, the second deal collapsed. The accountability everyone celebrated turned on a question of statutory authorization, not on moral reckoning.

What followed was not the end either — it was a third negotiation. All 55 attorneys general eventually signed a final $7.4 billion settlement, confirmed by the bankruptcy court on November 18, 2025 and effective May 1, 2026.⁶ The crucial difference is buried in the AG announcements and easy to skip: this settlement does not grant Sackler family members immunity from future opioid lawsuits.⁶ After six years, the one thing the family had built the entire bankruptcy to obtain — a permanent shield — is the one thing the final deal withholds.

Wasn't $7.4 billion still a real reckoning?

The honest objection is that $7.4 billion is an enormous number and that the Sacklers gave up far more than any normal defendant ever would. That is fair, and it is true. Purdue's own legal history is not trivial either: a Purdue affiliate pleaded guilty in 2007 and agreed to pay $634,515,475, and Purdue Pharma LP itself pleaded guilty to three federal felonies in 2020 under what the Justice Department called the largest penalties ever levied against a pharmaceutical manufacturer.²³ So the system did extract money and admissions. But weigh it against the design. The family entered the process holding around $11 billion in withdrawn cash, structured the entire bankruptcy around buying immunity for a fraction of it, and lost that immunity only because four years of litigation reached a court willing to read the statute literally.⁵ A reckoning that depends on a 5–4 vote breaking the right way is not a reckoning the system reliably delivers. It is one it nearly missed.

Purdue Pharma dissolved on May 1, 2026, and its machines now run under a new name, Knoa Pharma, watched by a monitor and forbidden to market the kind of drug that built the fortune in the first place.⁷ That is a genuine ending for the company. But the company was never the point. The point was the family that had already taken its money out, and the bankruptcy was their attempt to take the protection out too — to file the corporation and keep the cash. They got the dissolution they could survive. What they did not get, after six years and three deals, was the one prize the whole maneuver was built to win: a wall they could stand behind forever. They had nearly bought it. The statute, read plainly, simply wasn't for sale.