Theranos Whistleblower's AI Warning: The Diligence Gap Founders Own

Shultz's Warning Is About Verification, Not Villains

Theranos whistleblower Tyler Shultz, in an interview published October 30, 2026, called the AI funding boom "a great opportunity to go deceive people," because AI products are unusually hard for investors to diligence. He made the remark in an interview with The San Francisco Standard, summarized by TechCrunch, which framed the moment as one in which companies that have never turned a profit raise capital at trillion-dollar valuations.

The Theranos case was not built on a charismatic founder. It was built on a measurable gap: according to the SEC's 2018 press release, Theranos claimed it would generate more than $100 million in revenue in 2014 and generated a little more than $100,000 from operations. According to TechCrunch, the company did not share scientific information or evidence that its machines worked with potential investors.

Shultz argues AI is really hard to do due diligence on. "It's really hard to tell the difference between what is something truly unique and new versus like, a ChatGPT wrapper," he said, as quoted by TechCrunch. Overstating what a model does, often called AI-washing, is the modern version of that gap.

What the SEC Actually Charged in Theranos

On March 14, 2018, the SEC charged Theranos, Elizabeth Holmes and Ramesh "Sunny" Balwani with raising more than $700 million from investors through an elaborate, years-long fraud. The complaints alleged false statements in investor presentations, product demonstrations, and media articles about a portable blood analyzer that, per the SEC, could complete only a small number of tests.

Theranos was a private company. That is the detail founders raising private rounds should sit with. The SEC's charges followed the pitch materials, not a public listing.

The price was personal. According to MLex's summary of the settlement, Holmes settled without admitting or denying the allegations and agreed to:

• Pay a $500,000 penalty.
• Accept a 10-year bar on serving as an officer or director of a public company.
• Return 18.9 million shares.
• Relinquish voting control of Theranos.

The criminal case went further. The Ninth Circuit affirmed her conviction and a restitution order of $452,047,268 owed jointly with Balwani to 12 victims, per the HHS Departmental Appeals Board decision.

The Capital Formation Distinction AI Founders Conflate

Two things get blurred in AI fundraising: a trade secret and an unverifiable claim. Theranos, per TechCrunch, did not share scientific information or evidence that its machines worked with potential investors, citing concerns about revealing trade secrets. Protecting model weights is legitimate. Refusing to substantiate what the model does is a disclosure problem.

The distinction matters because AI decks are full of capability language that sounds technical but is really a claim of fact: proprietary model, autonomous agent, accuracy rates, enterprise deployments. Each, if it cannot be substantiated, carries the same risk as "deployed by the Department of Defense," a Theranos claim the SEC said was false.

What Your Leadership Team Should Decide This Quarter

Build a claims file before the next round

First, inventory every AI capability statement. Pull the deck, data room, website, and investor updates, and list each claim about what your product does.

Second, tie each claim to evidence. Benchmark results, customer contracts, architecture notes showing what is proprietary versus a third-party model. If the evidence does not exist, the claim comes out.

Third, separate projections from facts. Revenue forecasts need stated assumptions. Theranos's projection of more than $100 million in 2014 revenue against a little more than $100,000 in revenue from operations is the cautionary benchmark.

Assign owners

• The CTO signs off on technical capability claims.
• The CFO signs off on revenue and customer metrics.
• The board receives the claims file before any priced round closes.

Ask your general counsel or outside counsel to review the file against what investors are actually being told.