The Federal Court Ruling That Redacted Nothing

The numbers in one of the largest patent disputes in technology history were staggering even by Silicon Valley standards. The plaintiff sought billions in damages. The defendant countersued. The case touched nearly every smartphone sold on Earth and raised questions about the boundary between inspiration and infringement that would eventually reach the Supreme Court. It was, by any measure, the most consequential patent litigation of the decade — and one of the most closely watched trials in technology history. So when the U.S. District Court for the Northern District of California published a "redacted" ruling in the case, every journalist, analyst, and attorney in the industry opened the PDF. What they discovered was that the redactions did not work. The confidential terms that both companies had fought to keep sealed were sitting right there in the file, available to anyone who could operate a clipboard.

The Stakes Behind the Seal

Patent litigation between major technology companies is, at its core, a war over money and market position. But the courtroom proceedings generate a second category of sensitive information that has nothing to do with the patents themselves: confidential licensing terms, internal pricing strategies, market share analyses, profit margin data, and negotiation positions that the parties have disclosed under seal. This information, if made public, could reshape negotiations with other licensees, alert competitors to vulnerabilities, or embarrass the companies in front of investors and regulators.

Both companies had designated enormous volumes of material as confidential during the litigation. Licensing rates for specific patent portfolios. Revenue figures for individual product lines. Internal communications about competitive strategy. When the court issued rulings that referenced this material, the standard procedure was to file a redacted version for the public docket, with the confidential passages blacked out, and a sealed, unredacted version accessible only to the parties and the court. The system depends entirely on the redaction actually working.

The Failure

The court published a ruling with what appeared to be standard redactions — black bars over the confidential passages. To the eye, the document looked correct. The sensitive figures and terms were covered by dark, opaque-looking rectangles. The ruling appeared to comply with the protective order governing confidential information in the case.

But the black bars were cosmetic. They were visual overlays placed on top of the text, not true redactions that removed the underlying data from the PDF content stream. The confidential text — licensing rates, royalty calculations, negotiation positions — remained fully intact in the file. A reader could highlight the blacked-out area, copy the text, and paste it into any text editor. The entire content of the "redacted" passages was recoverable in seconds.

Reporters and legal bloggers discovered the failure almost immediately. Within hours, the confidential terms that two of the most powerful technology companies in the world had spent millions of dollars in legal fees to protect were circulating freely on the internet. The court was forced to take the extraordinary step of sealing the entire opinion after the fact — pulling from the public docket a ruling it had already published. The damage, of course, was already done. Information published on the internet is not easily retracted.

The Anatomy of a Federal Court's PDF Mistake

What makes this failure notable is not just that it happened, but where it happened. This was not a solo practitioner filing a motion in small claims court. This was the Northern District of California — the federal court that handles more technology cases than any other in the country. Its judges and staff process sealed filings in high-stakes intellectual property disputes as a matter of daily routine. Few courts in America handle more PDF redactions than this one.

The failure highlights a persistent gap in institutional knowledge around digital documents. The federal courts have published guidance on electronic filing. The Administrative Office of the U.S. Courts has issued privacy policies. The Federal Judicial Conference has adopted rules requiring redaction of personal identifiers. And yet the basic technical question — does a black rectangle over text actually remove the text from the file? — appears to have been either unknown or unverified by whoever prepared this document for the public docket.

The error could have been caught by a five-second test: select the blacked-out text, copy, paste. If words appear, the redaction failed. That test was either not performed or not understood to be necessary — a gap worth noting for a court that serves as the de facto forum for America's technology industry.

What Got Exposed

The exposed passages contained precisely the kind of information that protective orders exist to shield: specific licensing terms between the parties, the court's analysis of reasonable royalty rates, references to confidential settlement discussions, and internal financial data that both companies had produced under compulsion of discovery. In patent litigation, licensing rates are among the most sensitive categories of business information. A rate disclosed in one case can be used as a benchmark by every other potential licensee, fundamentally altering the economics of an entire patent portfolio.

For both companies, the leak was not just embarrassing — it was potentially costly. Licensing negotiations with other companies, already underway or anticipated, would now proceed with the counterparty armed with information about what rates these patents had actually commanded. The carefully constructed information asymmetry that makes patent licensing profitable was, in one afternoon, partially demolished by a PDF that did not know the difference between hiding text and deleting it.

A Systemic Problem in the Legal System

This patent case failure was not an isolated incident. In another high-profile matter, defense attorneys filed a "redacted" brief that revealed, via the same copy-paste technique, that their client had shared sensitive internal data with a foreign contact. A federal transportation agency published its full screening procedures manual under black rectangles. A military report on an overseas incident was "redacted" with the same non-functional method.

Each of these failures has an identical root cause: the person preparing the document believed that placing a visual overlay on text was the same as removing the text from the file. It is not. A PDF is not a piece of paper. A black rectangle drawn on top of text in a PDF is an annotation — a separate layer of the document that floats above the content. The text underneath remains in the content stream, fully intact, fully searchable, fully extractable. The rectangle hides nothing from anyone willing to look.

Proper redaction does something fundamentally different. It removes the underlying text from the PDF content stream, destroys the character data, and replaces the region with an opaque fill that has no text beneath it. After proper redaction, there is nothing to copy because the text no longer exists. This is not a subtle distinction. It is the entire distinction. And it remains poorly understood by courts, law firms, and government agencies that handle sealed and confidential materials every single day.

How to Get Redaction Right

The fix is not complicated. Use a tool that performs true redaction — one that removes text from the PDF content stream rather than drawing over it. Then verify the result. Select the redacted area. Try to copy text. Search for words that should have been removed. If any content survives, the redaction failed.

PDFb2's redaction tool performs true content removal. It strips the underlying text from the PDF content stream — not the visual layer, the actual data. And because it processes files entirely in your browser, the unredacted document never leaves your device. For anyone filing documents that contain sealed or confidential material, that combination of proper redaction and client-side processing addresses both the technical and the privacy requirements in a single step.

The Northern District of California handles cases involving the most valuable intellectual property on Earth. A court of that stature publishing a ruling with non-functional redactions — and then having to seal its own opinion after the fact — illustrates how widespread the misunderstanding remains. The text is still there. It has always been still there. The only variable is whether someone checks.