By Kevin Brenner | Friday, June 5, 2026 

Not Good is not an indictment of the companies it covers. It is a study of the mistakes made inside them by people prone to making them. 

_______________________________________________________ 

In June 2014, Fortune put Elizabeth Holmes on its cover and ran an admiring profile of her directors under the headline “A Singular Board at Theranos.” Singular was one word for it. Here is the roster: 

George Shultz. Henry Kissinger. William Perry. James Mattis. Sam Nunn. Bill Frist. William Foege. Gary Roughead. Richard Kovacevich. Riley Bechtel. Two former Secretaries of State. A former Secretary of Defense. A retired Marine general who would become one. Two former U.S. senators, one of them a heart-transplant surgeon and Senate Majority Leader. A retired four-star Navy admiral. A former director of the CDC. The former CEO of Wells Fargo. The former CEO of one of the largest construction firms on earth. 

Two of them were even physicians. Bill Frist was a heart-transplant surgeon; William Foege, the epidemiologist whose containment strategy helped eradicate smallpox. But a board with two doctors still had no one who understood diagnostic laboratory science, the one thing Theranos actually did. None of the directors had worked in clinical lab medicine or FDA-regulated diagnostic testing. 

Not one of them was equipped to answer the only question that mattered: Does the technology actually work? And not one of them tried to find out.  

What Theranos Said It Could Do 

Elizabeth Holmes dropped out of Stanford at 19 to build Theranos on a single claim: a small device called the Edison could run hundreds of clinical diagnostic tests from a finger-prick of blood, faster and cheaper than any lab, with no needle. 

She named the device after Thomas Edison, and she meant it as a symbol of perseverance. As John Carreyrou recounts in Bad Blood: Secrets and Lies in a Silicon Valley Startup, Holmes liked to quote the inventor’s line, “I have not failed. I’ve just found 10,000 ways that won’t work,” and told a graduating class in 2015 that “we assumed we’d have to fail 10,000 times to get it to work the ten-thousand-and-first. And we did.” They had not. The Edison failed in the ordinary sense of the word, the one with no triumphant sequel. The name fit better than she knew. Before it was the Edison, employees had called the prototype the “gluebot,” because it was cobbled together from a robot built to dispense glue. 

If the pitch had been true, it would change medicine. Routine testing would reach anyone, anywhere. Early detection of cancer, heart disease, and infection would no longer require a doctor’s order and a vein. She told investors the military was already using Theranos devices on medevac helicopters in Afghanistan. It wasn’t. She raised more than $700 million on that story. 

The Edison could not do what Holmes said it could do. Theranos knew this. For years. 

On the handful of tests it could run at all, the results were unreliable. For the rest, which was the vast majority of the menu it offered patients at Walgreens locations across Arizona and California, Theranos was quietly running samples on commercially available Siemens analyzers, diluting finger-prick samples to fit them, then reporting the results under its own branding. The proprietary device handled roughly a dozen of the more than two hundred tests on the published menu. Patients who thought they were getting a revolutionary finger-prick test were getting repurposed lab equipment, returning results that were sometimes dangerously wrong. 

One patient, tested at an Arizona Walgreens in 2015, got a false-positive HIV result. Others got results that triggered needless alarm and follow-up testing. One man’s test falsely flagged possible aggressive prostate cancer, and a pregnant woman who had suffered three prior miscarriages was wrongly told her hormone levels had crashed, suggesting another. Both results were debunked by tests run elsewhere. Some patients may have had real conditions the tests missed.  

The federal indictment charged that Theranos was not capable of consistently producing accurate results for a list of tests that included HIV, hCG, sodium, potassium, and calcium. The full scope of patient harm was never publicly quantified. Sunny Balwani, the company’s president and Holmes’s second-in-command, was convicted of defrauding patients, Holmes was acquitted on the same counts, and the restitution went to investors and corporate partners, not patients. The people who got wrong results never got an accounting. 

What the Board Was Built to Do 

Holmes understood that a credible board was a product. She was not assembling oversight. She was assembling proof of concept. The board’s job was to make the company look legitimate, and it worked.  

Investors took the roster as the diligence. The head of investments for one family that put in $100 million later admitted she never visited a Theranos testing site, never called a Walgreens executive, and never hired an outside expert to check the claims. Why would she? Two former Secretaries of State were on the board. Academics who study venture fraud have a name for this: proxy due diligence, treating the presence of other reputable people as a substitute for checking anything yourself. Holmes built her board to exploit exactly that. Marquee names create social proof, and social proof stops questions.  

It worked for a decade. 

None of the directors had the scientific background to evaluate whether the Edison’s claims were even plausible. That was not an accident. Holmes recruited from a world where reputation is the credential and prestige stands in for diligence. Shultz brought in much of the rest of the board. It grew by connections, not by expertise. The people on it were exactly who Holmes needed them to be: famous, trusted, and with no reason to ask the one question that would have ended the company. 

It wasn’t always this way. Theranos once had a director who asked hard questions. In 2006, Holmes recruited Avie Tevanian, once Steve Jobs’s right hand at Apple and the executive who had led the team that built Mac OS X. He did not last long. “I think what she didn’t expect was that I would actually ask a lot of questions,” Tevanian later said. The revenue projections never materialized. The pharmaceutical deals he asked about were always “under legal review.” The technology wasn’t improving. When he pressed, Holmes gave him “a non-answer or an evasive answer,” then began freezing him out and sent another director to deal with him. 

When Tevanian compiled the board’s own materials into a record of how much Holmes’s story kept changing and pressed the chairman to act, he was told to resign. Theranos threatened him with legal action, and he was gone by 2007, barely a year after joining, warning the other directors that crossing Holmes meant retribution.  

Then Holmes built the board the world would come to know. The diplomats, the generals, the statesmen. People who would lend their names and never go looking for the documents that had cost Tevanian his seat. 

The Moment the Board Had the Information 

In 2013, a young Stanford biology graduate named Tyler Shultz went to work at Theranos. He was George Shultz’s grandson. He believed in what Holmes said the company was building. 

He landed on the assay validation team, the people responsible for confirming that tests run on the Edison actually worked. They didn’t. His quality-control results did not match what the company was claiming. In March 2014, using an alias, he filed a complaint with New York State health regulators, the first known regulatory complaint about the company’s lab. Weeks later he emailed Holmes directly, drew a blistering reply from Balwani, and resigned the same day. A year after that, a Wall Street Journal reporter named John Carreyrou reached him through LinkedIn, and he became a confidential source for the investigation that would bring Theranos down. 

 

Through all of this, his grandfather gave him no cover. George Shultz did not believe him, tried instead to broker a deal between his grandson and the company, and kept hosting Holmes as a welcome guest at family gatherings. 

 

Once Holmes learned he was talking, the company turned its machinery on him. At a confrontation at his grandfather’s home, Theranos attorneys ambushed him and pressed him to sign a sworn statement recanting what he’d found and promising never to speak to the press, telling him that signing would “end a world of suffering.” He refused. His grandfather pressed him to stay quiet anyway, and kept treating Holmes as family. 

 

Over the next two years, Theranos came after him in earnest. Private investigators followed him. The company’s lawyers threatened him with litigation, pressed him repeatedly to sign affidavits naming exactly what he’d told the Journal, and pushed him to identify other whistleblowers, a demand he kept refusing. His family spent roughly $400,000 in legal fees on his defense. 

 

Before his death in 2021, George Shultz came to acknowledge his grandson was right. He never apologized, but the two reconciled, and he said Tyler had “made me proud” and shown “great moral character.” 

That is the Theranos board in one story. A director’s own grandchild tried to tell him the company was harming patients. The director’s response was to sit in a room with the CEO’s lawyers while they pressured the whistleblower to sign a document and disappear. 

The board was not deceived in that moment. It chose. 

That is the part worth sitting with, because oversight was the one job the board could not delegate back to the person it was supposed to be watching. Strip away the prestige and a director’s role is simple to state, and it is a real legal duty, not a courtesy: directors answer to everyone who isn’t in the room, and they are responsible for knowing what the company is actually doing and making sure it stays inside the rules that govern it. Compliance isn’t someone else’s department. It sits with the board. That duty doesn’t bend because the company is glamorous, or the founder is persuasive, or the other directors are household names. It is the job. It was the job at Theranos too. 

 

My Read 

The conventional Theranos narrative is that Holmes was a brilliant liar who fooled serious people. That framing lets the board off too easy. 

Holmes lied. She lied to investors, partners, regulators, and patients, often alongside Balwani. The verdicts and the record make that clear. But the board’s failure was not, at its core, that it was deceived. It was that the board never tried to find out. 

Nobody visited the lab. And the one time a sitting Vice President toured the Newark facility, the company built him a fake one to look at. Nobody hired an independent expert to review the Edison’s capabilities. Nobody asked why a company claiming hundreds of tests had almost no peer-reviewed data behind it. Nobody seemed to notice that the scientists who actually understood diagnostics were raising their hands. In February 2015, Stanford’s John Ioannidis pointed out in JAMA that Theranos had never published a shred of peer-reviewed validation, coining the term “stealth research” for exactly this. A few months later, a University of Toronto clinical biochemistry professor reviewed the public claims and concluded most of them were exaggerated. The company’s answer to the one critic who mattered was to send its general counsel to pressure him into recanting.  

 

The questions the board never asked were the ones federal inspectors answered for it. When the Centers for Medicare and Medicaid Services examined the Newark lab in late 2015, it found deficiencies posing immediate jeopardy to patient health and safety, and went on to revoke the lab’s certificate and bar Holmes from operating a clinical laboratory for two years. The FDA separately cited the company for shipping its blood-collection vial as an uncleared Class II medical device. 

 

The board had the stature to ask those questions. It had the resources to commission that review. What it lacked was any reason to. Holmes had structured the board so that asking hard questions was not the job. The job was lending the name. 

There is a version of this story where one director with laboratory-medicine experience asks a single question in 2011: Can you show me an independent validation of the Edison’s accuracy across the test menu? That question ends Theranos before it opens a single Walgreens location. Before one patient gets a wrong result. 

Nobody asked it. 

A board does not have to be in on the fraud to fail at its job. The Theranos directors described themselves as people who were deceived, as if that closes the matter. It doesn’t. Being deceived is what happens to a board that never built a way to check. The lie was the symptom. The missing oversight was the disease. 

And the warning signs that mattered weren’t scientific. You did not need to understand diagnostic chemistry to ask why a company claiming hundreds of tests had almost no published data. Or why the experts in diagnostics who looked closely kept reaching the same skeptical conclusion. Or why the company’s instinct, when challenged, was to send lawyers rather than evidence. None of that requires a lab coat to read. It requires a board paying attention and willing to pull the thread. 

That is the lesson worth taking. A board that settles for what management hands it is not a check on management. It is an audience. 

The Bill 

Holmes was convicted on January 3, 2022, on one count of conspiracy to commit wire fraud against investors and three counts of wire fraud, involving transfers of more than $140 million. The jury acquitted her on the patient-fraud counts and hung on three others. She was sentenced to 135 months in federal prison. The Ninth Circuit affirmed her conviction and sentence in February 2025, and the full court denied rehearing that May. 

Then, in March 2026, a federal judge trimmed her sentence to 123 months under a guidelines amendment for first-time offenders, finding that none of her investor-victims had actually suffered financial hardship from losing their money. The investors’ own paperwork, attesting that they could bear the risk, helped get her the reduction. She is currently projected for release in 2031, and has a clemency request pending with the Trump administration. 

Balwani, tried separately, was convicted on all 12 counts, investor fraud and patient fraud alike. He got 155 months. Nearly 13 years. Balwani drew more time because the jury convicted him on more counts, including defrauding patients, while it acquitted Holmes on every patient-related charge. 

The two were ordered jointly to pay $452 million in restitution: $397 million to twelve investor-victims, plus $54.5 million to Safeway and Walgreens. 

The SEC charged Theranos, Holmes, and Balwani in 2018 with raising more than $700 million through what it called a “massive fraud.” Holmes settled, agreeing to a 10-year bar from serving as an officer or director of a public company, a $500,000 penalty, the return of 18.9 million shares, and the surrender of her voting control. Theranos, by then, was worthless. 

Walgreens, which had partnered with Theranos to run wellness centers in about 40 Arizona stores and invested roughly $140 million across the relationship, sued in 2016 to recover that amount and reportedly settled for under $30 million. It later paid $44 million to resolve a consumer class action brought by patients who bought Theranos’s inaccurate blood tests at its stores.  

Safeway spent about $350 million building clinics in more than 800 supermarkets under a project it codenamed “T-Rex,” even after its own executives, tested at a trial clinic in Safeway’s headquarters, got results so questionable that one false cancer flag had to be walked back by another lab. The warning was inside the building. Safeway spent the money anyway, and when the tests never materialized, it severed ties in 2015, recovering little. 

The investors who sat on that board lost their investments. None of them faced enforcement action. 

How to Avoid Becoming the Next Cautionary Tale 

Build boards for oversight, not optics. A board whose entire value is reputational cannot perform its basic function. In a regulated, technical industry, at least some directors need the expertise to ask whether the product works and to understand the answer. That means someone in the room whose explicit job is to say “show me the data,” and who can actually read it when it arrives. The Theranos board had nobody who could, so the question was never asked.  

Oversight is a system, not an instinct. The reason boards miss disasters is rarely that the directors are foolish. It is that no one required the bad news to reach them. A functioning board doesn’t wait to be told. It demands a standing schedule of reports on what’s going wrong, and it treats certain things as automatic tripwires that have to surface the moment they happen: a whistleblower complaint, a regulator’s inspection or warning, an investigation, a test result the company had to walk back. At Theranos, employees were complaining and regulators were circling, and none of it ever reached the board. A board that hears only what management chooses to share has handed the oversight function back to the people it is supposed to be watching. 

Prestige is not diligence. If you are an investor looking at a company with a marquee board, the question is not who sits on it, but what they have actually reviewed. Has the board commissioned independent technical validation? Has anyone watched the operations run? If the answer is no, and the board’s value is its roster, you are looking at a due diligence gap dressed up as a governance asset. 

Treat NDA enforcement against employees as a red flag, not a management tool. When a company aggressively pursues the people who raise internal concerns, that pattern tells you what management is afraid of. When a compliance program punishes the people who flag problems, it stops protecting the company and starts protecting the misconduct.  

A board exists to verify, not to vouch. At Theranos, the most powerful boardroom in America did the opposite, and that, perhaps more than the lie itself, is the lesson worth keeping. 

Not Good is a short newsletter about corporate misconduct, enforcement actions, and the very expensive lessons hidden inside other companies’ worst decisions. Subscribe, share, and send me the bad acts you can’t believe happened. 

Click Here to Schedule a 30 minute Consultation with Kevin Brenner

This blog post is for informational purposes only and does not constitute legal advice. The discussion of this matter, including the conduct of any individuals involved, is based solely on publicly available information and court filings. Nothing in this post should be interpreted as a statement of fact about any person’s character, intentions, or actions beyond what has been reported in official sources.

The analysis provided reflects general legal principles and commentary and may not apply to any specific situation. Reading this post does not create an attorney-client relationship with the author or their firm. If you have questions about how these issues may affect your organization, you should consult qualified legal counsel.

The information provided on this website is for general informational purposes only and should not be considered legal advice. No attorney-client relationship is created by accessing or using this website. Please consult with a qualified attorney before making any legal decisions. Global Link Law is not liable for any reliance on the information provided. Prior results do not guarantee a similar outcome.