This article was AI-generated as part of an experimental historical-content project. The date reflects the period being analyzed rather than the date the article was originally written.
A criminal trial is supposed to be where a story gets settled. For reputation purposes, it usually works the other way around. The trial is where the story gets written in its durable form, often weeks before anyone knows the verdict.
The FTX case is a clear example in real time.
Two records, running in parallel
Sam Bankman-Fried’s trial opened on October 3 in federal court in Manhattan, starting with jury selection. As NPR summarized that morning, he has pleaded not guilty to seven criminal counts, including wire fraud, securities fraud, and conspiracy to launder money, and the trial was expected to last about six weeks. The same day, Michael Lewis’s book about him, Going Infinite, went on sale.
By the second week, the government’s cooperating witnesses were on the stand. On October 10, Caroline Ellison, the former chief executive of Alameda Research, told the jury that Bankman-Fried “directed me to commit these crimes,” as the BBC and AP reported. The defense has not yet presented its case. A verdict is still some way off.
What strikes me is that two records are being built at once, and they follow different rules.
The court record is organized around legal elements and admissible evidence. It moves slowly, it is precise, and it ends in a verdict.
The search record is organized around attention. It is made of headlines, quotes, explainers, and encyclopedia summaries, and it has no ending. It simply accumulates.
Why the search record sets early
The search version of a story tends to harden during a trial rather than after it, for mechanical reasons.
Coverage volume peaks while testimony is happening. Every major outlet publishes daily, and each new article links back to earlier coverage, which reinforces a handful of pages as the authoritative ones. The most quotable lines get repeated in headlines, then in explainers, then in the summary paragraphs that Wikipedia editors update as the case unfolds. Once a phrasing becomes the consensus phrasing, it becomes the version people keep encountering, and the version that chatbots trained or grounded on web text are likely to repeat.
A verdict, when it comes, will add a sentence at the top. It rarely replaces what has already settled underneath.
The blast radius is wider than the defendant
For finance professionals, the more practical issue is co-occurrence. A six-week trial generates thousands of articles, and every name that appears in them repeatedly starts to travel with the case.
That includes the obvious entities, FTX and Alameda. It also includes investors who backed FTX on the way to what NPR described as a $32 billion valuation, the celebrities who appeared alongside Bankman-Fried during the company’s peak, and companies that surface in testimony or legal argument only briefly. On October 11, for instance, the judge noted a pending government request to exclude evidence about the value of FTX’s investment in Anthropic, according to the trial transcript. Being mentioned in a case is not the same as being implicated in it. Search does not always make that distinction visible.
What this means for investor relations and financial communications
A few practical points follow for anyone whose firm sits near a case like this.
Do not wait for the verdict to look at the record. If coverage contains factual errors about your firm’s role, those are worth correcting now, while articles are still being updated and before the phrasing gets copied into summaries. This is about facts, not spin.
Map your adjacency. Know which articles mention your firm alongside the case, and how. Counterparties, limited partners, and prospective hires will search these names for years, and they will see the first page, not the docket.
Publish a plain, dated account of the relationship, if there is one. A short factual statement on your own site gives journalists, researchers, and editors something citable that is not a secondhand characterization.
Markets react to news and move on to the next thing. Search does not price events. It stores them. By the time the jury reaches a decision, much of what people will find about FTX, and about everyone who stood near it, will already be in place.