All of This Has Happened Before
Olive AI, IBM Watson Health, Carbon Health, Babylon Health. Four failures, six patterns, and the questions nobody asked in time.
Tyrion Lannister famously said in ‘Game of Thrones’:
“I drink, and I know things.”
These days I don’t drink. But after 40+ years in the healthcare industry, you do begin to notice things. Patterns.
One sequence that I have seen often repeated now is the remarkable human capacity to believe that:
this time,
this technology,
this founder,
this solution to a problem.
is going to be the charm.
But the wisdom of 40 years now cautions me:
“All of this has happened before, and all of this will happen again.”
Book of Pythia (Battlestar Galactica)
Over the past year, here on Substack, I have written four long-form analyses of major healthcare technology failures: Olive AI, IBM Watson Health, Carbon Health, and now Babylon Health. Each of these is part of the content for my paid subscribers, as they are heavyweight, serious Strategic Insight case studies.
But when I laid them side by side, something emerged that felt too important not to share more widely.
If you enjoy this article would you care to…
Four stories that feel like one
All four companies were focused on revolutionizing healthcare delivery. Different founders, models, and problem-solution offerings.
And yet reading the four case studies together produces the same eerie feeling. Not quite a sense of deja vu, more that there’s a pattern here.
The commonalities become clear when you read the four case studies thoroughly, but to distill them to their core, I created the following table. Below this table, I call out the 6 common patterns I see.
Cross-case summary: Four companies, four market segments, six patterns of structural failure mode.
The six patterns. Every time.
Look across those rows long enough and six things just jump out to you.
1. The claims always came before the proof.
In all four cases, the core value proposition was marketed, sold, and funded before independent validation of its value and outcomes. The sequence was always: announce, raise, hype, scale, then — eventually — discover the ‘emperor’s new clothes’.
2. Borrowed legitimacy substituted for evidence.
Elite institutions, prominent investors, and even government endorsements did the work that clinical and financial outcomes data should have done. The halo effect from these endorsements was significant. When it finally, and predictably, cracked, there was nothing there to support the business claims.
3. The technology gap was hidden in plain sight.
RPA bots sold as AI. A deterministic rule-based system sold as cognitive computing. A custom EHR created from scratch that increased complexity rather than removing it. An app with multiple incompatible internal builds sold as an autonomous diagnostic engine. In each case, engineers inside the company knew. Leadership chose to position the desired narrative anyway.
4. Rolling Scope was used to defer accountability.
When one product line failed to deliver, there was always another initiative, another geography, another acquisition to point to. Breadth created the illusion of progress while the core proposition, the motivating idea, remained unproven. Olive had six product lines. Watson spanned four continents. Carbon ran clinics, an EHR, chronic care, and hardware simultaneously. Babylon operated across five countries and multiple initiatives.
5. Leadership acknowledged the failure only after it was irreversible.
Every CEO eventually said the quiet part out loud — but the timing is everything. It was always during a layoff announcement, at a divestiture, or in a bankruptcy filing. Never too early to change course. Parsa’s case stands alone: he said it while the company was still listed. The structure of the business had made the admission consequence-free.
6. The verification event always arrived. The question was whether anyone was paying enough attention.
An Axios exposé. A STAT News investigation. An MD Anderson audit. A Lancet peer review. Post-COVID financials that couldn’t be explained away.
In every case, external reality eventually forced the reckoning that the company had structured itself to avoid and run away from.
The gap between the claim and the proof was not closed by the companies themselves. It was closed by journalists, regulators, researchers, and market conditions. The question was never “if” — only how much capital, how many patients, and how much institutional trust would be consumed in the interval.
Who shares the blame?
None of these four failed on their own.
Each had a board stuffed with investor representation.
Each had industry analysts acting as ‘fan boys’ writing them up.
Each had a trade press happily soaking up the marketing messages to amplify the story.
Each had institutional customers who signed the contracts, and in some cases, even regulators and government agencies who nodded things through.
All of them had the opportunity to review the same information that the journalists eventually used to take the companies apart. They just chose not to look at it, motivated by their own agendas
The founder who overclaims is the visible failure. But the less visible failure is the number of serious, sophisticated parties who found the claim convenient enough not to interrogate too closely.
The sting in the tail (consequences)
What gets burned is not just the company’s capital.
It is the clinician’s patience with digital health innovations. That weary jaundice comes all too quickly.
It’s also patient trust in AI-assisted care.
And finally its the opportunity for the next company, the serious one, the slower one. The one with the outcomes data that isn’t as sexy for the multiples we were looking for, whose revenue projections seem unexciting for a 4-year exit for this fund. Who now has to spend the first 30 minutes of every pitch meeting explaining why they are not going to be the next Babylon or Olive AI.
Keep haverin!
The four Strategic Insight case studies — Olive AI, IBM Watson Health, Carbon Health, and Babylon Health — are available to paid subscribers here:
