Last year, I used a continuous glucose monitor (CGM) for the first time, and it completely changed how I eat.
After a couple of weeks using an over-the-counter Stelo CGM, I learned that sugary snacks shoot my blood sugar into the stratosphere, salads and veggies keep it at a manageable level, and light exercise, even a short walk after a meal, helps bring it down almost instantly.
This information was so revelatory that I began to wonder whether putting CGMs and the information they provide into the hands of a broader set of people could help us better manage societal health over time. After all, if it changed the way I eat, could it do the same for millions of others?
One way to explore that question is to talk about the technology with smart people. Last week at CES, I did just that when I moderated a session titled “From Brainwaves to Blood Sugar: How Next-Gen Tech Shapes Diets” during The Spoon’s Food Tech Conference at CES. I purposely programmed the session with a mix of panelists to bring medical, startup, investor, and researcher perspectives to the conversation.
After sharing my experience with a Stelo CGM on stage, I asked Howard Zisser about the importance of over-the-counter CGMs. Zisser, a physician and longtime pioneer in diabetes technology who worked on some of the earliest CGM systems in the early 2000s, when the technology was designed almost exclusively for people with diabetes, said the value of these newer CGMs lies in the data and what you can do with it.
“Instead of one or two readings a day, you suddenly have 300, 500, 600 readings a day,” Zisser said. “You start to see trends. What happens when you fast, when you exercise, during a menstrual cycle. IYou get a rich data set that’s your data personally.”
When I talked about how jarring my own experience was in seeing my blood sugar spike, Zisser argued that the shock is part of the value. Unlike biomarkers such as cortisol, which are difficult to influence in real time, glucose is actionable.
“You see it, and you can change your behavior,” he said. “Next time, you make a different choice.”
He likened glucose feedback to learning how to drive with a speedometer. Without it, he said, you’re guessing. With it, you can learn how your actions translate into outcomes.
But not everyone will wear a sensor on their arm. Noosheen Hashemi, founder and CEO of January AI, argued that while hardware CGMs are powerful, they are not scalable to the hundreds of millions of people with undiagnosed prediabetes or metabolic dysfunction. She said technology like that developed by her company leverages machine learning models trained on years of CGM data to predict glucose responses without requiring buying a hardware sensor.
“Our claim to fame is creating the world’s first continuous glucose monitor with AI,” Hashemi said, explaining that the system can generate directionally accurate predictions using inputs such as age, weight, activity level, sleep, and food intake.
But for all the data and actionable insight these tools can provide, they do not guarantee lasting change. Sherry Frey, VP of Total Wellness at NielsenIQ, shared research showing that even after receiving a diagnosis and initially adjusting their diets, behavior often reverts within months.
“We actually see about nine months in that a lot of behavior reverted,” Frey said. “When people were maybe less engaged and a little tired of of having to eat differently.”
That drop-off highlights both the opportunity and the challenge for health technology. Sustained engagement requires more than numbers on a screen. It requires context, interpretation, and motivation.
Frey also noted that adoption of wearables and health-tracking technologies is expanding beyond affluent early adopters. One of the fastest-growing user groups, according to NielsenIQ data, includes consumers on SNAP benefits, many of whom are using these tools for chronic disease management rather than fitness optimization.
“The addressable market is much larger than people with diabetes,” Frey said.
As we discussed what makes behavior change stick, I asked whether giving consumers more data, as the Nest thermostat did starting a decade ago, would prompt lasting change. Peter Bodenheimer, U.S. venture partner at PeakBridge VC, said yes, but only if the insights are actionable.
“Insights that tell you, ‘if I do this, then something good or bad happens,’ tend to be the things that people respond to and maintain.”
The panel also acknowledged the downside of constant feedback. More data can mean more confusion, anxiety, and misinformation. Hashemi shared an example of a user who believed their glucose should never rise above 110, a misunderstanding fueled by social media rather than clinical reality.
“Metabolic fitness is how you go from fasted to fed efficiently,” she said. “It’s a preposterous idea to keep your blood sugar the same all the time. So yes, there’s a lot of misinformation.”
Zisser reinforced that interpretation depends heavily on individual context, goals, and physiology. The same glucose spike can mean very different things for a professional athlete, a person with diabetes, or someone trying to lose weight.
We also discussed other technologies that can help us understand what’s happening inside our bodies, such as neural implants and other next-generation sensors. Hashemi pointed to implantable sensors capable of reading multiple analytes for years at a time, as well as emerging efforts to continuously measure substances like lactate, ketones, alcohol, and eventually insulin.
“Yeah, there’s definitely implantables,” said Hashemi. “There’s one that reads 20 different analytes, including glucose. It lives, you have to inject it under your skin. It can live 900 days. And it’s still in animals. It’s not in humans yet. But these things are coming.”
As the number of measurable signals grows, so do concerns about privacy, trust, and data ownership. Frey noted that while many consumers want their health data integrated in one place, roughly half remain uncomfortable with embedded sensors and worry about how their information might be used by insurers, governments, or corporations.
Others felt that the benefits of these technologies may ultimately outweigh more abstract fears. When people see tangible improvements in sleep, energy, or focus, trust can follow.
“No government, no doctor can make somebody healthy,” Hashemi said. “The only person that can do that is yourself.”
As we wound down the session, we talked about personalized nutrition, a topic that has long been a point of heated discussion in the world of food and health. The panelists agreed that while personalized nutrition may never be perfectly precise, the combination of biological data, AI, and human context is moving the industry closer to that goal.
“The gold lives in the combination of data,” Hashemi said, suggesting that consumer-generated health data will increasingly merge with clinical care, especially as value-based healthcare models expand.
In the end, the promise of next-generation health tech may be less about perfect prediction and more about empowerment. One idea that Zisser suggested was possibly getting these types of technologies into the hands of young students as we are teaching them how to eat.
“When my dad taught me how to drive, he didn’t put me in a car without a speedometer, right? It’s like, have feedback, I have information. And so to give people that access to that, and not that they would need it all the time, but so they can learn how their choices impacts their glucose.”
Not a bad idea. I can only imagine what my long-term health outlook might be different if I’d had insight into the impact of certain foods on blood sugar when I was much younger.
If you want to hear my conversation with these smart people, just click play below.
