Editor’s note: This year What If?, our annual collection of scenarios, considers the future of health. Each of these stories is fiction, but grounded in historical fact, current speculation and real science. They do not present a unified narrative but are set in different possible futures

“LET FOOD be thy medicine and medicine be thy food.” The diktat from Hippocrates, who defined the principles of medicine in ancient Greece, hovers in bright holographic characters over the main stage at the World Economic Forum in Davos. The central theme this year is how to make personalised nutrition more widely available to those unable to afford its benefits. Hot topics include whether metabo-watches, implants and other personal-nutrition trackers should be free for everyone (as they are now in some Nordic countries), why personalised nutrition is good for business and the perennial debate over how governments can best regulate corporate use of consumers’ personal data.

Amid the arguments, there is broad consensus that the rise of personalised nutrition has done a lot to promote healthy and environmentally friendly eating over the past decade. In 2031 the proportion of obese Americans fell for the first time in more than 20 years, and the rate of diabetes has fallen for three years in a row from its all-time high of 22%. Europeans are getting slimmer and healthier, too.

But progress has been slower than hoped, and in emerging markets obesity is still rising, hobbling economic growth. Environmentally sustainable eating, though increasingly popular in the rich world, is still not on track to reach the “planetary health diet” target set by scientists in 2019 in the Lancet, a medical journal. That target, which big food manufacturers and many other firms have pledged to support, called for a 50% worldwide cut in red meat and sugar consumption and a doubling of the consumption of nuts, fruits, vegetables and legumes between 2020 and 2050.

That personalised nutrition is the best way to drum up demand for healthier and more earth-friendly foods became clear in the mid-2020s. A decade earlier, scientists had begun to unravel why one-size dietary guidelines in the form of food pyramids, sugar and fat labels and so forth were not turning the tide on diabetes, obesity and other diseases caused by bad diets. Faddish regimens with catchy names like Keto or Paleo worked for some people but were useless for many, if not most, people who tried them. And people who lost weight often found it hard to sustain.

The diets that came and went until the 2020s required steely willpower and careful planning. The biggest problem, however, was their failure to recognise that people’s bodies react differently to the same foodstuffs. By the late 2010s mounting scientific evidence showed that meals that were perfectly healthy for one person could be another person’s fast-track path to diabetes, obesity or heart disease.

It turned out that even the same meal eaten by the same person at a different time of day could be metabolised in a more or less healthy way, depending on their other eating, sleeping and exercise patterns. The most crucial discovery was the role of the microbiome, the colony of 100trn microbes living in the human gut. The microbiome, it turned out, was the factory that converted food into the various substances the body needs to function—as well as those that cause poor health. And everyone’s microbiome is unique.

A landmark in the idea of personalised nutrition was a study published in 2015 by researchers at the Weizmann Institute in Israel. They devised an algorithm based on artificial intelligence that could accurately predict an individual’s response to any given food, measured by continuous blood-glucose monitoring with a small device attached to the upper arm. Spikes in blood glucose after meals are known markers for weight gain and a panoply of metabolic disorders. The algorithm used data on lifestyle, medical background and the composition of the microbiome. Within three years scientists in America, Britain and Germany had replicated the Israeli team’s work and the business of personalised nutrition entered a new era.

During the early 2020s the number of startups offering bespoke nutritional advice by algorithm soared. Some used mail-in samples of body fluids or continuous monitoring devices to track blood levels of glucose, lipids, vitamins and so on. A few, including DayTwo, Million Friends and Zoe, did microbiome mapping too (through genomic analysis of everything found in a person’s stool sample). Many firms did just the bare minimum: checking for a handful of genes that had been linked with certain reactions to various foods. This had limited utility. By the late 2020s the market had reached maturity after a brutal shake-out.

Soup-to-nuts service

A handful of firms have thrived and are now household names. EatLogic, the second-largest, agreed last month to be acquired by Google, subject to regulatory approval. The leaders all have essentially the same business model. Their apps and algorithms identify what people should eat and avoid, and keep track of what is in their cupboards, refrigerators and online shopping carts. AI-generated recipes use flavour combinations favoured by leading chefs. The apps also analyse restaurant menus and recommend which dishes to order—sometimes with minor tweaks, such as swapping a vegetable or changing a salad dressing. All this helps people make good food choices. Accuracy has steadily improved as the implants and wearable devices paired with these services have become smaller, cheaper and more capable.

Makers of kitchen appliances, such as Philips and Samsung, have been central to the personalised-nutrition ecosystem since the early 2020s. At Davos their chief executives talked about the challenges—and opportunities for public health—of developing cheaper models for emerging markets, where the number of middle-class households is growing fast. (Obesity is also most common in that demographic segment.) Industry bosses reckon that in countries like India and Kenya, about 20% of households can afford a smart fridge, though one with far fewer features than the models that are now standard in America. In 2034 just over half of American households had a smart fridge linked to a personal-nutrition account.

The food industry has also adapted surprisingly quickly to the personalised-nutrition revolution, given how slowly it moved to reduce salt and sugar in processed foods. Its transformation is evident on supermarket shelves, where processed foods are available in multiple variants, tuned for each of the main metabo-types identified by scientists. (Some variants are, for example, higher in fat and fibre but lower in protein.)

Artificial meat and fish grown from animal stem-cells—which in 2034 surpassed the traditional variety by sales volume—also come in metabo-type varieties that include different ratios of the fat, protein, minerals and vitamins found in “real” animal products. Restaurant menus, too, increasingly cater to the most prevalent metabo-types among their clientele.

One of the most contentious topics discussed at Davos was how to make personalised nutrition more affordable. The first-generation services, offered in the early 2020s, started at several hundred dollars for initial tests, and hefty monthly fees thereafter. Today’s most basic plans are about 80% cheaper, after adjusting for inflation. Users who let providers sell their personal data get hefty discounts, though some regulators are looking to curtail the practice. Employers, health insurers and governments are increasingly subsidising personalised-nutrition plans and offering vouchers and other perks to obedient users.

But cost is not the only hurdle to greater uptake. In England, the National Health Service offers a free plan to everyone, along with subsidised personal devices that can be paired with it. This helps explain why about 70% of adults in England now use a personalised-nutrition service, the highest rate in the world. Convincing the remaining 30%, which includes many of those who stand to benefit the most from changing their diets, will take a lot more than free gadgets. Many take a dim view of the whole idea, because of conspiracy theories that doctors are struggling to dispel.

In the final debate on the main stage at Davos, the majority of speakers were optimistic about the future potential of the technology, while others worried about the difficulty of expanding adoption within these more “hesitant” groups. The discussion ended on a bittersweet note. Personalised nutrition, it seems, is not to everyone’s taste.

Full contents of this What If?
Freedom to tinker, October 2029: What if biohackers injected themselves with mRNA?
The other epidemic, June 2025: What if America tackled its opioid crisis?
A tale of two cities, June 2041: What if a deadly heat wave hit India?
You are what you eat, January 2035: What if everyone’s nutrition was personalised?*
iHealthy, September 2028: What if smartphones become personal health assistants?
Mrs Chippy’s benediction, February 2055: What if marmosets lived on the Moon?
Novel treatments, August 2050: What if dementia was preventable and treatable?
Rage against the machine, December 2036: What if an AI wins the Nobel prize for medicine?
Germ of an idea: What if germ theory had caught on sooner?

This article appeared in the What If? section of the print edition under the headline “You are what you eat”