What Happened to Babylon Health? From $4.2B SPAC to Collapse

Babylon Health collapsed in 2023 because it turned itself from a digital GP service into a US risk-bearing medical group and could not pay the medical bills. Revenue rose from about $80 million in 2020 to $1.11 billion in 2022, almost all from Medicaid and Medicare Advantage contracts in which Babylon was responsible for members' total cost of care. Losses grew with the revenue, the 2021 SPAC raised less cash than planned, and within two years of listing at a ~$4.2 billion valuation the equity was worth nothing.

The story Babylon told was about artificial intelligence. In June 2018 it announced at the Royal College of Physicians that its AI had scored 81% on a recreated version of the exam British GPs take. Founder Ali Parsa promised "an affordable and accessible health service in the hands of every person on Earth." The business it actually ran was much older and harder: taking a fixed payment per patient and hoping care cost less.

What was Babylon Health?

Parsa, a former investment banker who had co-founded the hospital operator Circle, founded Babylon in London in 2013. The app launched around 2014 as a private subscription: members paid a monthly fee (reported at about £5–7) for video or phone appointments with a GP, often within hours. Early backers included Sweden's Kinnevik, which became the largest shareholder.

Babylon then had three business models in sequence, each harder than the last.

How did GP at Hand work, and why was it controversial?

In November 2017 Babylon launched GP at Hand with a practice in Hammersmith and Fulham. It used NHS rules from 2015 that let patients register with a practice outside their area, and it grew to about 100,000 registered patients, England's largest single GP list (reported).

Under the NHS contract, a practice is paid a fixed annual sum per registered patient, adjusted for age, sex and deprivation. GP at Hand's registrants were overwhelmingly young adults who rarely need a GP. An independent 2019 evaluation found its patients younger, healthier and more affluent than average, satisfied with fast access, and prone to leaving when they needed ongoing care (reported). Critics, including the Royal College of GPs, called it cherry-picking, and the local commissioning group reported that the practice's growth pushed it toward a deficit (reported). NHS England adjusted payments for new registrants of digital practices.

The lesson for later events is that GP at Hand's economics depended on the mix of patients, not on the AI. Revenue per patient was below average, but cost per patient could be far lower still. That is a payer-formula gap, and the payer closed part of it.

What went wrong with Babylon's AI claims?

The June 2018 study compared Babylon's system with doctors on a small set of vignettes and a recreated MRCGP exam. A Lancet letter and other clinicians pointed out that it was not a peer-reviewed prospective trial, that it used a small number of test cases written by Babylon-linked doctors, and that it could not support the claim of performing on par with GPs (reported). Babylon also publicly identified an NHS consultant who had criticized its symptom checker under the name "Dr Murphy" (reported, 2020), which deepened clinicians' hostility.

The AI dispute mattered less to the company's finances than to its credibility. Babylon was valued as a technology platform. Its revenue came from human doctors and, later, from insurance-like risk. Chapter 14 of the book explains why AI claims need evidence matched to the job the AI does, and the industry study on digital health, telehealth and virtual care shows how few virtual-care models earned payer budgets.

Why did Babylon pivot to US value-based care?

After its $550 million Series C in August 2019, at a valuation above $2 billion, with Saudi Arabia's PIF ($200 million reported) and the US insurer Centene ($50 million reported) among the investors, Babylon decided the US was its main market. Full-risk primary care was the hottest model of 2020–21. Oak Street Health, Cano Health, VillageMD and agilon health were growing quickly, and investors valued them on members and revenue.

Babylon's argument was that cheap digital access and AI triage would catch problems early and keep members out of hospital, the source of savings in a risk contract. It bought Meritage Medical Network, a Northern California independent physician association with about 700 physicians (April 2021), and in January 2022 the kiosk company Higi and DayToDay Health. By January 2022 it had more than 275,000 US value-based care members (company).

The structure of the revenue explains both the growth and the collapse. In a capitated contract, a plan passes the provider a large share of the premium per member per month, and the provider books that as revenue. So revenue looks like an insurer's: large and fast-growing. The margin is the gap between that premium and medical costs. Babylon's own profile in the research gives a stylized illustration: 100,000 members at about $900 per member per month is about $1.08 billion of annual revenue, and a two-point miss on the medical cost ratio costs about $22 million. Medical claims arrive faster than year-end risk settlements, so cash runs out before the books show the problem.

How did the SPAC make it worse?

Babylon merged with Alkuri Global Acquisition Corp. and began trading on the NYSE on 22 October 2021 at an implied pro forma equity value of about $4.2 billion. The deal promised up to about $575 million of gross proceeds, roughly $345 million from the SPAC trust plus a $230 million PIPE, but trust redemptions were high, so the PIPE and existing holders supplied most of the cash actually received (reported). A SPAC let Babylon market projections of rapid US growth that a traditional IPO would not have allowed, and committed it to delivering them.

The net losses understate cash burn in places, because SPAC-related warrant and earn-out gains were booked as income. Babylon borrowed from AlbaCore Capital, exited its NHS hospital partnerships in Wolverhampton and Birmingham in 2022 to focus on the US, and in March 2023 guided to adjusted EBITDA profitability by mid-2024. Two months later it withdrew its outlook.

How did Babylon Health collapse?

The end came within four months. On 10 May 2023 Babylon announced a take-private framework with AlbaCore, reported at about $34.5 million of funding, and withdrew its 2023 outlook. It was delisted from the NYSE in June. A proposed merger with the Swiss company MindMaze collapsed in early August. The US business was shut that month, with about 94 Austin employees laid off (reported). The UK subsidiaries entered administration, and the clinical services, including GP at Hand, were sold to eMed Healthcare UK. Shareholders received nothing, and Kinnevik wrote its stake down to nearly zero (reported).

GP at Hand survived under eMed, which shows the UK practice had some durable value as a service even after the company around it failed.

Which factor of the Care Equation did Babylon fail?

The Care Equation holds that a healthcare business needs care that works, a payer with a renewing budget, an affordable route to that payer, and proof that survives contracting. Babylon's failure ran through the second and fourth factors together.

On the payer side, Babylon had willing payers: the NHS, governments and US health plans. But each was a powerful buyer writing its own contract. When the NHS limited out-of-area registrations and US plans pushed risk onto Babylon, the company had little leverage. In the US it also took risk without controlling the care. Many of its members kept their existing in-person doctors, so Babylon bore the cost of decisions it did not make. The book's chapter on value-based care lists "risk taken without controlling the care" as one of five failure modes, and names Babylon in it, alongside agilon, Bright Health and CareMax.

On proof, Babylon never published medical cost ratios showing its model lowered spending. Its AI claims were contested, and its US results, the evidence that actually mattered to payers, were never shown in a way that demonstrated the model worked. A risk-bearing provider whose only proof is revenue growth has proved only that it can sign contracts.

Capital, one of the book's two "weather" forces, set the timetable. Babylon pivoted into risk at the 2021 peak of digital-health funding, when US venture investment reached $29.1 billion (Rock Health), and needed more money in 2023, when that capital had disappeared.

What does Babylon teach about digital health and value-based care?

Know which business you are in. A company that markets AI but books capitated premium is a risk-bearing medical group. It should be valued, capitalized and managed like one.

Revenue growth in risk is a liability first. In insurance-like businesses, new members bring unknown risk and low initial risk scores. Growing 3.5 times in a year with thin capital, as Babylon did in 2022, is the opposite of how successful risk-bearers grow.

Control the care you are paid to manage. The full-risk companies that survived, such as Oak Street until its sale to CVS, owned their clinics and employed their doctors. Babylon's app sat on top of care it did not direct.

Publish the number the payer cares about. For a risk contract, that number is medical cost per member against a credible baseline, not app downloads or exam scores.

Babylon belongs to a cohort of 2021 listings that failed in similar ways, several of which appear on our list of the biggest healthcare bankruptcies. The mechanics of risk contracts are explained in our comparison of fee-for-service vs value-based care, and the wider system in how the healthcare industry makes money.

The takeaway

Babylon Health did not fail because its app was bad; patients who used GP at Hand liked it, and the practice still runs. It failed because it financed a fast pivot into full medical risk with marketing claims, a thin SPAC and borrowed money, and never proved its model lowered costs. Its revenue looked like a technology company's growth curve and behaved like an underpriced insurance book. For the economics of digital health and why so few models earned a durable payer, read Chapter 13, Digital Health and Virtual Care.