💥 Zero Bullshit
Chapitre 01

General anaesthetic

Loïc Brotons is 22 when he lands in Niaga, Senegal, as a fourth-year medical student. The clinic has no other doctor. He stays five months, alone, handling a cholera epidemic with whatever comes to hand: improvised suction, cauterising with heated scissors1. On the side, he sets up a taxi company in Dakar, four vehicles he runs for four years before giving them to his host family.

The man is a doctor, a real one, an anaesthetist and intensive care specialist registered with the French medical council2. After his studies, he works at Grenoble university hospital in ski trauma and intensive care, then at the Pierre Rouquès hospital, known as Les Bluets, in Paris’s 12th arrondissement, the historic maternity hospital founded by the CGT metalworkers’ union in 1937. He also spends five years in Sweden, where he discovers something that doesn’t exist in France: a shared national patient record. A single piece of software that centralises a patient’s medical data and that every care worker can consult, whatever the hospital. That’s the project he brings home.

Except that since 2012, Brotons has mostly been a tech entrepreneur. Three startups before Galeon, all liquidated: Squirlink, PocketTribe (2012-2013), MF Mayfly3 (2014-2015, an ephemeral social network based in Copenhagen4). None of them in healthcare. On those past failures:

He says they asked people whether they’d be interested, people said yeah, yeah, and the day it launched, nobody used it.

The pivot to healthcare comes in 2016, after those three failures. The team that forms around the project comes from the same world. Matthieu Gueniffey, the future COO5, dates it back:

More than ten years ago, he says, when he was trying to mine Bitcoin on his computer.

That’s how he met Brotons. The founding core came together over Bitcoin mining and social network projects, not in a hospital. Frédéric Da Ros and Peter Bartos are also on board from the start.

Because in France, it’s a desert. Each hospital has its own electronic patient record, or EPR6 (when it has one at all), often ageing, rarely compatible with the one next door. The market is held by a handful of long-established vendors, Softway Medical first among them. And the problem goes beyond sharing files. Around 80% of the data produced in hospitals is free text: dictated surgical reports, discharge letters, scanned handwritten notes. Unusable as it stands for research or for training artificial intelligence models. An EPR that structures data at the point of entry, that codes diagnoses, categorises procedures and standardises results, produces raw material you can actually use. And usable means: training an AI to detect breast cancer on anonymised mammograms, cross-checking millions of care pathways to spot drug interactions, predicting post-operative complications. That’s the promise. And it is a real, documented problem that everyone in the sector knows.

Since 2016, France’s 898 public hospitals have been organised into 136 regional hospital groups, known as GHTs. Purchasing decisions for information systems, including the choice of EPR, are taken at GHT level. And the market is held by players of a different order. Softway Medical, number one by client count, reports around €150M in revenue, 1,100 employees, a valuation above €1bn, and the backing of Bain Capital and Bpifrance, France’s public investment bank. Dedalus, the Italian-born European leader, posts more than €100M in consolidated revenue in France with a comparable headcount. Behind them come Maincare, Berger-Levrault (more than 2,000 employees at group level) and InterSystems (a US group whose TrakCare system is deployed in French university hospitals). The Ségur digital health plan, a government programme launched in 2021 with nearly €500M in state funding, has sped up modernisation and consolidation. A GHT buying an EPR signs up for a contract of 5 to 10 years, involving hundreds of care workers, migrations of critical data, and service continuity 24/7. The selection criteria always include the bidder’s financial strength. A vendor that goes bust halfway through a rollout means an entire hospital left without a system.

This is the world in which Brotons launches Galeon in 2016, initially specialising in obstetrics7. The path to legitimacy is solid: i-Lab 2017 winner, the French research ministry’s competition, in the medical technology category8, an HEC Paris programme in 2018, the Agoranov incubator in 2019, the European EIT Health competition in 20209, HDS (France’s health data hosting certification) and ISO 27001 certification in December 202310, and three awards at CES in Las Vegas in January 2024 (cybersecurity, Human Security, Web3 & Blockchain)11.

And behind the awards, real paying customers:

  • Caen university hospital rolls out the mother-and-child EPR on 2 May 2023, and 809 births and 5,472 pregnancy consultations go through the software between May and July12;
  • the Saint-Joseph hospital in Marseille votes unanimously for the first full multi-specialty rollout13 (it is an ESPIC14, a private non-profit hospital, not subject to the same tendering rules as a public GHT; a distinction that will matter later);
  • in total, Galeon claims 18 public hospitals equipped, around 3 million patients, and more than 10,000 care workers using it daily. We’ll see further down what those figures really cover.

The fact remains that the product is real. Brotons15 puts it this way:

They are building a cathedral, he says, a long-term project. Venture capital funds think in five or six years. Galeon thinks in twenty.

Except that the cathedral is going up opposite Softway Medical (€150M in revenue, 1,100 employees), with, by its own figures, around fifty staff and revenue of €3.3M targeted for 202616. Galeon is not among the vendors listed for wave 1 of the Ségur EPR programme. To win GHT contracts, you need size. To get size, you need GHT contracts. Galeon chose a third way out of the dead end: funding its growth with debt sold to private individuals.

The story is lovely, the product is probably useful, and the doctor is the real thing. That is exactly what makes the rest of this investigation so uncomfortable.