Within a few months, the same class of hardware was tested twice in ways that could not be less alike. In late December 2025 a creator bent a Galaxy Z TriFold until the frame gave way, on camera, once. In March 2026 Samsung Display filled an MWC booth with rigs that fire balls at folding panels. Both get called durability testing. Neither replaces the other, and knowing which question each answers is most of what you need to read either.
What a camera has that a bench does not
JerryRigEverything's TriFold run, published as Samsung TriFold Durability Test: We found the limit and covered by 9to5Google on 25 December 2025, went badly for the phone. By that coverage and Tom's Guide: the outer glass marked at level 6 on the Mohs scale, the inner folding panel marked at level 2, fine grit reached the hinge and left it grinding, and the frame gave way when bent against the direction it folds — the first Samsung foldable to break in that segment. The conclusion in that coverage was structural, not anecdotal: more fold points and a softer screen material give a trifold more ways to fail than a single-hinge foldable has.
Three properties make that worth your time. Nobody chose the procedure to flatter the device. The same procedure, by the same hands, has been run on model after model, which makes outcomes comparable across phones even though each is a single unit. And the result was published whatever it turned out to be. A manufacturer's test programme has none of the three, and not because the people running it are dishonest: a test you commission is a test you can decline to publish.
What a bench has that a camera cannot
Repetition, a written procedure, and somebody from outside the room. Samsung Display announced on 22 July 2025 that the panel used in the Galaxy Z Fold7 completed 500,000 folds across 13 days at 25°C and stayed fully functional, verified by the certification company Bureau Veritas. That is a real method and a real number: fixed conditions, a fixed count, an outside party, and a result you can run again on another unit.
It is also narrower than the headline sounds. Half a million folds is a fatigue result for one motion at one temperature. It says nothing about scratch hardness, nothing about grit reaching a hinge, and nothing about force arriving against the fold. Samsung Display's own extrapolation in that release — more than ten years for someone folding about 100 times a day — is arithmetic on top of the test, not something the test measured.
Vendor claims also arrive pre-framed, and most of them are not numbers at all. On the Fold8, Samsung describes a dual layer of titanium under the display and an Armor FlexHinge, and says the result reduces crease visibility while improving durability. There is no count behind that, no written procedure and nobody from outside: it is a description by the party that built the thing, against a baseline the same party chose.
A ball fired at a screen is a demonstration
MWC 2026 is where the categories blur most easily. Samsung Display's own release of 2 March 2026 describes a golf putting challenge in which a foldable phone serves as the hole, and a basketball backboard carrying 18 foldables while a robotic arm launches balls at them. Android Central, reporting from the booth, wrote that the newer panels carry a shock-dispersing structure with thicker glass and a titanium plate beneath the display.
Look at what that establishes. A demonstration is built so the device survives it; surviving is the exhibit. It does show a panel taking a class of impact repeatedly, which beats a slide. But there is no published pass criterion, no stated sample and nobody outside counting. Demo, test and certification are three different things dressed alike.
They disagree, and both can be right
A panel that survives 500,000 folds and a phone that dies on its first reverse bend are not contradicting each other. Fatigue along the axis a hinge was designed for and failure across that axis are separate modes, and passing one predicts nothing about the other. An ingress rating works the same way: it certifies particles up to a defined size under defined conditions and promises nothing about a hinge with sand in it. Every test is silent outside its own mode, and that silence is where readers quietly fill in confidence nobody earned. Which parts of a destructive test generalise at all is the question that follows this one.
If the foldable is already yours
Neither test reaches your unit, and neither describes it. What the two jointly justify is small and practical. A factory film over a panel that marks at fingernail hardness is doing structural work, so peeling it off is a decision. A day at the beach is a hardware event for a hinge, whatever the IP code says. Closing a folding phone on something left in the crease is the failure you can actually cause yourself, and it is the one the bend segment found.
When you are choosing, use them for different jobs. Vendor bench numbers are the fair way to compare fold endurance between models, because the procedure is written down. Creator tests are the fair way to compare materials and sealing, because the same hands do the same thing to everything. Take neither outside its lane, and treat a booth demonstration as marketing that happens to be physical.
The general rule is easier to apply once you have watched it break in a specific case: what a durability test proves about your own phone works through the TriFold results segment by segment, and an engineering test aimed squarely at a marketing claim is the format at its most useful. If the hardware is what brought you here, what actually improved in foldables this year is built on the manufacturers' published figures, and a blind camera test shows what a creator can control for that a spec sheet never will.
Last updated September 22, 2026