The claims, checked
"The Theranos technology worked"
It did not, at least not as sold. Theranos pitched a device that could run a broad menu of blood tests from a finger-prick sample. In reality, the proprietary Edison analyzer could reliably handle only a small number of tests, and the company ran many patient samples on modified commercial machines built by other manufacturers, sometimes with diluted samples.
You don't have to take a jury's word for this part. Federal lab regulators inspected the company's California lab, found serious deficiencies, and sanctioned the company. In 2016 Theranos voided two years of results produced on its own device. A company does not throw out two years of patient test results if the machine works.
"Holmes was convicted"
Yes, and it is worth being precise about what she was convicted of. In January 2022, after a months-long federal trial, the jury found Holmes guilty on four counts of defrauding investors: three counts of wire fraud and one count of conspiracy to commit wire fraud. She was acquitted on the counts tied to patients, and the jury did not reach a verdict on several others.
In November 2022 she was sentenced to more than 11 years in federal prison, plus restitution, and she began serving the sentence in 2023. Her appeals did not overturn the conviction. So when you see "convicted fraudster" next to her name, that is not spin. It is the record.
"She knew the tests were unreliable"
This is the claim that separates fraud from failure, and the trial dealt with it head-on. Former employees testified that they flagged accuracy problems inside the company. Validation data showed the device failing quality checks. Meanwhile, investors were being shown a very different picture of what the technology could do.
Here's why that matters legally: to convict on wire fraud, the jury had to find intent, meaning Holmes knowingly deceived investors rather than just being wrong. The guilty verdicts on the investor counts mean the jury found exactly that. Her defense argued she believed in the technology and was let down by others. The jury heard that argument in full and rejected it on four counts.
Why this case matters beyond Theranos
Holmes is the benchmark. When people ask whether some other founder is "the next Elizabeth Holmes," this is what the real thing looks like: a specific false claim about a product, evidence the founder knew it was false, charges, a trial, a conviction, a sentence. Most founder controversies never get past step one. Compare this record with a case where a company failed loudly but no charges ever followed, like Adam Neumann and WeWork, and the difference is obvious.
One more distinction worth keeping: the jury acquitted Holmes on the patient counts. Even in the clearest founder-fraud case of the decade, the verdict was narrower than the headlines. That is normal, and it is why we check claims one at a time instead of stamping a whole person.
Related checks
Sam Bankman-Fried
The other benchmark conviction: seven counts and a 25-year sentence over FTX.
Charlie Javice
Invented user numbers sold for $175 million. A smaller fraud with the same anatomy.