Design products for people, not ask people to adapt
In a San Francisco design lab, Samsung is quietly rewriting the terms of how technology learns to fit the human body. Rather than relying on the limited testimony of a few test subjects, the company's Design Innovation Center now consults hundreds of millions of anatomical data points and thousands of AI-driven simulations to ask a question that has long eluded the industry: what does comfort actually mean, at scale, across all of humanity? The shift from intuition to measurable engineering suggests that fit and form—once considered matters of taste—are becoming disciplines as rigorous as any other branch of applied science.
- Traditional wearable testing has always carried a quiet flaw: a small room of volunteers cannot represent the full spectrum of human bodies, leaving millions of real-world users to adapt to products designed for the few.
- Samsung's Design Innovation Center is dismantling that limitation by building proprietary databases of 3D and 4D anatomical scans from globally diverse populations, creating digital stand-ins that never tire and never run out.
- For the Galaxy Buds4 alone, the team ran over ten thousand simulations to refine details as precise as earbud rotation angle and body size—decisions that produced measurably better stability and comfort across a far wider range of ear shapes.
- The stakes extend beyond comfort: a wearable that shifts during wear produces less accurate health sensor data, meaning fit is now a medical precision problem as much as a design one.
- As Samsung's anatomical database expands with each new product cycle, the AI tools trained on it grow sharper—pointing toward a coming decade in which entirely new categories of wearables may become possible only because the data finally exists to design them.
Inside Samsung's San Francisco design lab, a fundamental rethinking of wearable development is underway. For decades, the industry leaned on small groups of testers to judge whether earbuds or watches felt right—a method that worked well enough in the testing room but struggled to account for the enormous variation of human bodies in the real world. Samsung's Design Innovation Center has chosen a different path: rather than testing more people, it analyzes hundreds of millions of anatomical data points and lets artificial intelligence carry the weight.
The method is called computational design. Led by executive vice president Federico Casalegno, the San Francisco lab begins by capturing three-dimensional and four-dimensional scans of ears and wrists from a globally diverse population. These become "digital twins"—precise anatomical models that can be tested endlessly without a single human volunteer present. From there, the team runs thousands of simulations, each probing how a design adjustment might affect comfort, stability, or sensor accuracy.
The Galaxy Buds4 offers a concrete example of what this looks like in practice. Samsung analyzed hundreds of millions of ear data points and ran more than ten thousand simulations to refine two seemingly small variables: the rotation angle of the earbud and the size of its body. The result was measurably better stability and comfort across a broader range of ear shapes than previous designs had achieved—and improved sensor precision, since an earbud that stays in place produces more reliable health readings.
What gives Samsung a durable advantage is the proprietary nature of its database and the custom AI tools trained on it. As both grow with each new product cycle, the simulations become more accurate and the insights more nuanced. Casalegno envisions this foundation enabling entirely new wearable categories within the coming decade—devices that could only be conceived once designers had enough data and computational power to optimize for human anatomy at this level of depth. The deeper promise, he suggests, is a partnership between human creativity and machine intelligence: designers freed from the constraints of traditional testing, empowered to be more inventive, with AI confirming whether their ideas hold up across the full range of human variation.
Inside Samsung's San Francisco design lab, a fundamental shift is underway in how wearables get built. For decades, companies have relied on a small group of testers to evaluate whether earbuds fit comfortably or watches sit right on the wrist. The problem was always the same: human bodies vary wildly, and what works for ten people in a testing room doesn't necessarily work for millions wearing the product in the real world. Samsung's Design Innovation Center has decided to solve this differently—not by testing more people, but by analyzing hundreds of millions of data points and letting artificial intelligence do the heavy lifting.
The approach is called computational design, and it represents a move away from intuition toward measurable engineering. Federico Casalegno, the executive vice president running the San Francisco lab, describes it as a way to design products for people rather than asking people to adapt to products. The philosophy sounds simple enough, but the execution is intricate. The process starts with capturing three-dimensional and four-dimensional scans of ears and wrists from a globally diverse population. These scans become "digital twins"—precise anatomical models that designers can test against endlessly without needing a single human volunteer in the room.
Once those digital twins exist, the real work begins. Samsung's team runs thousands of simulations, each one testing how a design change might affect comfort, stability, and sensor accuracy. For the Galaxy Buds4, the company analyzed hundreds of millions of data points from ears around the world and ran more than ten thousand simulations to refine what might seem like minor details: the angle at which the earbud rotates, the size of the main body. These weren't guesses. They were decisions backed by data that covered far more human variation than any traditional testing method could capture.
The payoff is tangible. Those small adjustments—a slightly smaller body, a refined rotation angle—resulted in measurably better stability and comfort across a much wider range of ear shapes than previous designs achieved. Casalegno notes that the process also improves sensor precision, which matters because an earbud that shifts even slightly during wear can produce less accurate health readings. A wearable that stays put is a wearable that works better.
What makes this approach proprietary to Samsung is the database itself. The company has built its own collection of anatomical data and developed custom artificial intelligence tools trained on that data. As that database grows—and it will, with every new product tested and refined—the simulations become more accurate and the insights deeper. Casalegno sees this as the foundation for the next decade of wearable innovation. He suggests that as the tools improve, Samsung may be able to design entirely new categories of wearables that don't yet exist, products that could only become possible once designers had enough data and computational power to optimize for human bodies at this level of precision.
The broader implication is a shift in how technology companies think about fit and comfort. These are no longer subjective qualities left to the judgment of a handful of designers and testers. They are engineering problems with measurable solutions. For consumers, this means wearables that work better for more people—not because the company tested more people, but because the company understood human anatomy more deeply. The real power, Casalegno suggests, lies in the partnership between human creativity and machine intelligence: designers freed from the constraints of traditional testing can be more inventive, knowing that artificial intelligence will help them validate whether their ideas actually work across the full spectrum of human variation.
Notable Quotes
Technology without humanity is simply perfection without purpose— Federico Casalegno, Executive Vice President, Samsung Design Innovation Center
The real power lies in the partnership between human creativity and machine intelligence, enabling designers to be more inventive while AI validates ideas across human variation— Federico Casalegno
The Hearth Conversation Another angle on the story
Why does fit matter so much for something like an earbud? It's just sitting in your ear.
An earbud that shifts during wear doesn't just feel uncomfortable—it actually degrades the sensors. If the device moves, the accelerometers and other sensors lose accuracy. You're measuring heart rate or movement from a slightly different position each time. The data becomes noise.
So computational design is really about solving two problems at once—comfort and measurement precision.
Exactly. Traditional design treated them separately. You'd test comfort with people, then hope the sensors worked. Computational design treats them as one problem: design a shape that stays stable on the body and delivers accurate readings.
How does analyzing millions of ear scans actually change what a designer does day to day?
Instead of building a prototype and testing it on ten people, a designer can run ten thousand simulations overnight. They see instantly how a small change in angle affects stability across different ear shapes. It's like having access to a library of human variation that would be impossible to gather physically.
But doesn't that risk making everything optimized for an average that doesn't really exist?
That's the opposite of what happens. By analyzing the full range of variation, you're actually designing for the extremes—the people whose ears are smallest, largest, most curved. The average takes care of itself.
What happens to all this data once the product ships?
It feeds back into the database. Every Galaxy Buds4 sold is another data point about how the design performs in the real world. That makes the next simulation more accurate. The system gets smarter with scale.
So in ten years, Samsung's wearables could look completely different because the AI will have learned things we can't even predict now.
That's the bet. Right now, we're optimizing within the category of earbuds and watches. But with enough data and computational power, you might design a wearable that nobody thought was possible because nobody had the tools to make it work for most people's bodies.