Theranos Didn't Fail Because of a Broken Machine. It Failed Because No One Could Fire Elizabeth Holmes.
Holmes engineered herself roughly 99% of Theranos's voting rights — and stacked a board of statesmen who literally could not remove her. The fraud lasted a decade not despite the controls, but because she had built a company with none.
Comes with a free Founder Doctrine Canvas template — plus a worked example for Theranos.
The board of Theranos had a former Secretary of State, a future Secretary of Defense, and enough Washington gravitas to fill a museum wing. It also had no way to fire the chief executive. Not a hard way, not a slow way — no way at all. Because by the time anyone might have wanted to remove Elizabeth Holmes, she had quietly engineered herself roughly 99% of the company's voting rights, and a director who pushed too hard could be replaced before the board ever met to push back.56 That is the part of the Theranos story that doesn't fit on a magazine cover. It wasn't the blood machine that made the fraud possible. It was the cap table.
The popular story is that Theranos was a brilliant idea undone by a machine that didn't work, run by a charismatic founder who fooled a star-struck board. A naive board got dazzled. The real story is colder and more instructive: the board was not naive — it was disarmed by design. A company with normal governance catches a fraud like this in eighteen months. Theranos ran for over a decade. The difference was structural, and Holmes built the structure on purpose.
She didn't just run the company. She owned the off switch.
Most founders give up control gradually. Each funding round dilutes them; each new investor wants a board seat; eventually the founder answers to people who can vote them out. That's not a flaw in the system — it's the system. It's the mechanism that converts other people's money into other people's oversight. Holmes severed that link at the root. Theranos used a dual-class structure that handed her 100 votes for every single vote held by anyone else, and she later multiplied the voting rights of her shares until she held approximately 99% of the company's total votes.5 Read that number again. Not a majority. Not a comfortable cushion. Near-total. The board could discover anything — that the Edison analyzer had accuracy and reliability problems, that the technology was never deployed by the Defense Department despite the legend, that 2014 operating revenue was a little over $100,000 against the hundreds of millions being raised2 — and it would not have mattered procedurally. They had found out. They could do nothing.
What a 'make-believe' board is actually for
If the board couldn't govern, why assemble such a famous one? Because its job was never governance. Journalist John Carreyrou called it a 'make-believe' board, and the phrase is exact: the directors weren't there to oversee Holmes, they were there to be displayed to investors.6 A name like Henry Kissinger or James Mattis on a letterhead does not check a single test result. What it does is launder credibility — it tells a wealthy newcomer that serious people have already looked under the hood, so you don't have to. The board's prestige was the very thing that suppressed the diligence it appeared to guarantee. And the legal reality underneath was blunt: with controlling votes, a founder can replace directors at will, which leaves a dissenting director exactly two options — get fired or resign.6 Oversight, in any meaningful sense, requires the power to remove. Strip that and a board is theater with a quorum.
“Holmes retained total control of the composition of the company's board... approximately 99% of total voting rights, making board oversight structurally impossible.”5
The one time a director pushed, and what happened to him
Theory is one thing; the structure had a live test. Board member Avie Tevanian — a former senior Apple engineer, not a man easily intimidated — objected to a share issuance to a Holmes-controlled foundation that would concentrate her voting power even further. He was the rare director who saw the machinery for what it was and pulled on the lever. The lever did not move. Instead, in a private meeting, the board chair suggested Tevanian consider resigning.8 That is the whole mechanism in a single scene: a director objects to the concentration of control, and the concentration of control is precisely what ensures he, not Holmes, is the one shown the door. The system doesn't punish dissent loudly. It just routes it out the building before it can become a vote.
| A normal venture-backed board | Theranos | |
|---|---|---|
| Can remove the CEO | Yes — that's the core function | No — Holmes held ~99% of votes |
| Who controls board seats | Shareholders, by vote | Holmes, unilaterally |
| What a dissenting director can do | Build a coalition, force a vote | Resign or be replaced |
| What the board's prestige signals | Real scrutiny applied | Scrutiny that was never possible |
Layer the legal intimidation of internal dissenters on top of the share structure and the architecture is complete. The people inside who saw the test failures couldn't safely speak; the people on the board who might have acted couldn't legally win. Holmes didn't need to fool everyone forever. She needed only to ensure that anyone who stopped being fooled had no standing to stop her. Control wasn't a symptom of the deception. Control was the wall the deception lived behind.
Isn't strong founder control how every great company gets built?
Here is the honest objection, and it's a real one: founder control is not inherently sinister. Some of the most valuable companies of the era run dual-class structures specifically to protect a long-term vision from short-term shareholders, and they have rewarded that protection handsomely. Concentrated control let founders ignore quarterly noise and build things that committees would have strangled. So 'Holmes had too much control' can't be the whole lesson — plenty of honest founders have just as much. The distinction is what the control is paired with. Visionary founder control is insulation from the market's impatience while the truth is still verifiable by results: ship the product, the numbers settle the argument. Theranos's control was insulation from verification itself — it was deployed precisely where the results could not settle the argument, because the results were being hidden. The same structure protects a contrarian genius and a confidence game; the difference is whether reality is allowed into the room. At Theranos it never was. The voting math guaranteed it never had to be.
When you evaluate a founder-controlled company, the question is not 'how much control?' but 'what can override the founder if they're wrong?' Healthy concentrated control still leaves a circuit breaker — an independent audit the founder can't fire, a board seat that survives a falling-out, results that arrive on a schedule the founder can't move. Theranos had none. The voting structure, the hand-picked replaceable board, and the silencing of dissent removed every mechanism by which being wrong could become being stopped. A founder who has engineered away every check has not just bought freedom to build. They've bought freedom from being caught. Treat the absence of an off switch as the most important number on the cap table.
When the SEC finally settled with her in 2018, look closely at one of the terms it extracted: Holmes agreed to relinquish majority voting control by converting her super-majority shares to ordinary Class A.1 Regulators understood the same thing the board never could act on — that the control structure was not a footnote to the fraud, it was the load-bearing beam. Strip it out and the whole thing could no longer stand. The blood test was the product Theranos sold to the world. The cap table was the product Holmes sold to herself: a company where being wrong, however completely, could never become being removed. That worked for over a decade. It stopped working the moment someone with subpoena power could finally do what no director ever could — make the truth count more than the votes.
Founder Doctrine Canvas
A one-page canvas for the operating system inside a founder's head: the principles they hold, the formative experiences that forged them, and the specific strategic moves each principle produces. Blank to make your own decision rules legible to the people who execute them; filled as the worked example showing why the story's company keeps making the bets it makes — because the founder can't make any others.
The worked example unlocks with a subscription. See plans →
Sources
Where this comes from — the filings, records, and reporting behind it.
- 1The SEC charged Theranos, Holmes, and Balwani with raising more than $700 million from investors through an elaborate, years-long fraud; Holmes agreed to pay a $500,000 penalty, return 18.9 million shares, relinquish majority voting control by converting her super-majority Class B Common shares to Class A, and be barred from serving as officer or director of a public company for 10 years.
- 2On June 14, 2018, a federal grand jury indicted Holmes and Balwani on two counts of conspiracy to commit wire fraud and nine counts of wire fraud; the indictment alleges they knew the Edison analyzer had accuracy and reliability problems, performed a limited number of tests, and could not compete with existing machines. Theranos technology was never deployed by the U.S. Department of Defense, and the company generated a little more than $100,000 in revenue from operations in 2014.
- 3On January 3, 2022, a federal jury convicted Holmes of one count of conspiracy to commit fraud on investors and three counts of wire fraud against individual investors involving wire transfers totaling more than $140 million; the jury acquitted Holmes of all patient-related fraud counts. Balwani was separately convicted on all 12 counts and sentenced to 155 months (12 years, 11 months).
- 4Holmes was sentenced to 135 months (11 years and 3 months) in prison on November 18, 2022, and began serving her sentence on May 30, 2023, at the Federal Prison Camp in Bryan, Texas. As of 2024, her projected release date is August 16, 2032, reflecting good-conduct reductions.
- 5Theranos had a dual-class shareholding structure where Holmes held 100 votes for every one vote of other shareholders, and she retained total control of the composition of the company's board. Holmes later multiplied the voting rights of her shares to give her approximately 99% of total voting rights, making board oversight structurally impossible.
- 6The Theranos board included prominent figures such as Henry Kissinger and James Mattis, but was described by journalist John Carreyrou as a 'make-believe' board due to Holmes's voting control. Legal experts confirmed that a founder with controlling interest can replace the board, leaving directors only the choice of being fired or resigning if they strongly disagree with the executive.
- 7Holmes founded Theranos in 2003 at age 19 after dropping out of Stanford University during her sophomore year. The company's original name was 'Real-Time Cures' before being renamed Theranos. Balwani, 19 years her senior, did not officially join Theranos until 2009 as COO, though he advised Holmes from the company's inception and the two were secretly romantically involved.
- 8Board member Avie Tevanian objected to a share issuance to a Holmes-controlled foundation that would further concentrate her voting power; he was subsequently pressured in a private meeting by then-board chair Don Lucas, who suggested Tevanian consider resigning. This episode illustrates the structural suppression of board-level dissent at Theranos.