Wearable tech and AI are coming together to read the secrets of racehorses' hearts. The answers might help save lives.
Steam rises off the backs of freshly galloped horses at Fergal O'Brien's racing stables in the Cotswolds in the UK. It's raining on the morning I visit, and the sweet smell of hay mixes with mist and wet grass.
Over their doors peer the elegant heads of Miss Tillytucks, Cool Customer, Ignite the Night and Renegade Girl. Almost a hundred of these Thoroughbred athletes make up the yard, which has been playing host to a world-first experiment.
Inside one box, Max Kendrick is readying a young bay mare, Sióg Geal ("Bright Fairy"). The 31-year-old jockey turned assistant-trainer wraps a wide, elasticated band around the mare's ribs. A small red light flashes in time with her heart.
Under the band's wicking, rubberised electrodes sends an electrocardiogram (ECG) trace straight to Kendrick's phone, where it pulses like the graphs seen by hospital beds.
From Kentucky in the US to Melbourne in Australia, tens of thousands of horses are now wearing devices to measure their performance and their health. The aim is to help them win their races – but also, potentially, to save their lives.
Since he was a child, Kendrick has honed his senses to read horses; to "feel" their behaviour and pick up on their unspoken tells. But now many trainers like him are turning to technology – to flag changes that the human eye can't detect.
The device being trialled today is a lightweight version of one Kendrick tested in May 2025, when his ride, Ted's Friend, became the first in the world to wear a heart monitor in an official race.
If the new devices can be rolled out more widely on racecourses, their data could help unlock a "black box" of cardiological secrets, says Celia Marr, Rossdales Veterinary Surgeons' internal medicine specialist, who is working on the British Horseracing Authority 's (BHA) pilot scheme.
"Ultimately, this is not just about developing better technology to monitor horses," adds Kamalan Jeevaratnam, professor of clinical physiology and head of the School of Veterinary Medicine and at the University of Surrey, whose team is using the data to build AI systems that can screen for cardiac risk .
"It is about using technology to understand the heart in ways we have not previously been able."
The marriage of wearable tech and AI could improve veterinary screening in several areas, including musculoskeletal injury , the number one cause of race day fatality.
It is, however, a "particularly exciting" development for cardiology, says Jeevaratnam.
In the past, horses' hearts could only be monitored for relatively brief periods, often in highly controlled environments or clinical settings. Plus, there are only so many heartbeats of an ECG that the human eye can analyse.
Now, it's possible to monitor while in motion and for longer. This allows vets to detect subtle changes , not just obvious abnormalities, Jeevaratnam says.
"We are moving from taking occasional snapshots of the heart, to watching the heart's electrical behaviour almost like a continuous film."
As this understanding builds, there's hope of tackling some of horseracing's most devastating health risks, including sudden death.
Sudden death effects only around 0.03% of horses, but accounts for up to a quarter of racecourse deaths – while also increasing the risk to jockeys and spectators, as heart failure can send horses careening into crowds .
It is a considerable diagnostic challenge, since many victims have no obvious signs of cardiac abnormality, either before collapse or during a post-mortem examination . Yet progress is being made.
In the US, equine cardiologists at the universities of Minnesota and Pennsylvania have collected ECGs from racehorses who've gone on to die from the condition – and shown a connection to the irregular heartbeat seen in atrial fibrillation .
Irregularity is something that is already "relatively easy" to screen for before exercise by using a stethoscope and smartphone ECG device, says Sian Durward-Akhurst, an assistant professor in large animal cardiology at the University of Minnesota, Twin Cities.
Meanwhile, at the School of Veterinary Medicine at the University of Surrey, researchers have already developed two different AI-led approaches that can identity horses with irregular ECGs and help predict potential cardiac issues – one for atrial fibrillation and another for exercise-associated arrhythmias .
It will be a while though before any AI tool will be able to screen every horse who shows up at a racetrack and flag abnormalities to a human vet in real time, says Marr.
Challenges still include fine-tuning these AI tools, establishing which kinds of arrhythmias are harmless and which could be fatal, and getting the AI tools to read the ECGs in real-time .
But, ultimately, the aim is to help predict cardiac issues before they occur.
"Just like an Apple watch, which will flag if it thinks your heart rate is irregular," says Marr. "The horse would be on its way to the start line, and the AI would be sending a signal back to say not to let it run."
How can a rider or trainer tell when a horse feels good?
Wearable equine technology – or equine "fitness trackers" – are helping inform this ancient connection . By measuring everything from heart rate to stride length, they can use AI to build a picture of what "normal" looks like for an individual horse.
They're not a replacement for trainers or vets, says Kendrick, but an additional tool that can help them decide when to rest a horse and when to race. Similar technology is already used in human sports , he notes, citing the UK's Premier League football, New Zealand's All Blacks rugby team and the National Football League (NFL) in the US.
Tens of thousands are already using such devices at training yards, including the Arioneo Equimetre and Equimetrics V-Pro used in the BHA's heart monitoring scheme. In Dubai, even racing camels are wearing them. In the US, a sensor from StrideSafe is also being deployed at racetracks, to help predict musculoskeletal injury .
Until last year, though, the impact of a race on a horse's cardiac response was still "essentially guesswork", says Kendrick. So he reached out to Sally Taylor, head of equine regulation, safety and welfare for the BHA, to see if that could change.
"Raceday data provides another dimension," explains Jeevaratnam. "It allows us to examine what happens when the horse is performing at its physiological limits, under genuine competitive conditions."
The development of AI-interpreted ECGs also has potential to reach beyond horses.
Horses are particularly fascinating models for how human hearts respond to extreme exercise, says Jeevaratnam. They have "exceptionally powerful" hearts that can adapt remarkably when athletically trained.
The work likely won't translate directly into human treatments, he cautions, but could provide new insights. For example, in screening for exercise-related cardiac arrhythmias , and improving understanding of triggers.
"In all sports – human too – risk will exist," adds Marr. "This is about how we reduce that."
Out on the gallops at Ravenswell Farm, the yard's Irish "bossman", 54-year-old Fergal O'Brien, is an expert in managing risk. The horses are fresh and it's not long before one leaps sideways, throwing its jockey into the mud, then takes off at pace with the rest of the field.
Fergal rushes to check the rider is OK: the back of her body protector is smeared with mud but she's unharmed. Panic can be catching, so everyone – both riding and watching – is trained to stay calm in moments like these. Yet there's a palpable sense of relief when the unmanned horse is finally caught.
"I love a Monday," O'Brien says playfully of the horses' high spirits, while doing a little jig, as if to shake off any tension that might still be hanging in the mist.
How does he feel about the new tech? And is there a risk AI analysis could put trainers or vets out of a job?
"You just hope you get more right than wrong," he says of the constant calls a trainer must make about a horse's fitness and health. Anything that helps that is welcome.
• A painting went missing 20 years ago. Police tracked it down with the help of AI
• The medieval secrets being revealed by AI
• Scientists are defying physics to mimic bone for longer-lasting hip replacements
Some concerns do still linger. Marr points to the technical challenge of distinguishing heart signals from other forms of electrical "noise".
Jeevaratnam notes the more generalised risk of humans not feeling bold enough to sometimes reject AI's automated instructions. Trainers must be wary of becoming "over-dependent" on AI, he says.
But for now, at least, there's little sign of this technology replacing the "eye" of trainers and vets – even as it helps them see further into horses' hearts.
"The way she jumped, the height she was giving it: you can see she's got no pain anywhere," O'Brien says while watching Sióg Geal trial the new monitor.
For more science, technology, environment and health stories from the BBC, follow us on Facebook and Instagram .