A blown fuse is designed to interrupt current, yet the question of whether it can start a fire is not as straightforward as a simple yes or no. In the vast majority of properly engineered installations, the fuse opening is the endpoint of a fault, not the beginning of a combustion event. But the real answer hinges on what happens inside the fuse holder during those few milliseconds when the element vaporises and an arc forms. That arc is a plasma column with temperatures that can exceed 5,000°C, and if the fuse is not properly contained, that heat can indeed ignite nearby materials.
Consider a standard 30A automotive blade fuse mounted in a plastic holder under the dashboard. Under a dead short, the element melts in roughly 5 ms at 300A, but the arc continues until the gap widens enough to extinguish-typically 1 to 2 ms longer. During that interval, the arc energy reaches about 15 Joules for a 12V system. Fifteen joules is enough to raise the temperature of a 1‑gram piece of PVC by roughly 30°C, but not enough to ignite it because the heat is concentrated in a tiny volume between the contacts. However, if the fuse holder has corroded terminals, the contact resistance adds extra heating before the fuse even blows. A corroded terminal with 50 mΩ of resistance dissipates 4.5 watts at 300A during the fault, which is negligible for milliseconds. The real fire risk emerges not from the arc itself but from the sustained heating that occurs when a fuse does not clear properly.
A fuse that blows but fails to fully separate-a phenomenon called "sputtering" or "partial clearing"-leaves a conductive bridge of molten metal that continues to carry current in an intermittent manner. That condition produces a series of micro-arcs that can last for seconds, each one ejecting hot metal particles into the surrounding cavity. One documented case in a residential solar combiner box involved a 15A DC fuse that had been subjected to repeated partial faults. After the third event, the fuse element had deposited enough metal vapour on the inner wall of the ceramic housing to create a tracking path. The next overcurrent caused a surface arc that ignited the plastic baseplate, even though the fuse's internal element had melted. That fire was traced to a fuse with an insufficient sand filling-the quartz was only 60% of the required volume, so the arc plasma escaped through the end caps.
The holder itself plays a critical role. A fuse is only as fire-safe as its mounting. A glass cartridge fuse in a spring-loaded clip can generate significant heat at the clip interface if the spring tension weakens over time. At 80% of rated current, that clip might reach 120°C under normal conditions-hot enough to discolour adjacent wires. When the fuse blows at 200% overload, the clip temperature spikes to 250°C for a few seconds, which is below the autoignition of most plastics (around 400°C), but if the clip is mounted against a polypropylene wire bundle, the sustained heat from the arc flash can raise the insulation to its softening point. I have examined a failed pump control panel where a 10A glass fuse had blown, but the arc had shot straight through the glass envelope-which had a hairline crack from prior handling-and impinged on a nylon cable tie. The tie melted and dripped onto the fuse clips, creating a carbonised path that later caused a secondary short. The original fuse had opened correctly, but its physical breach enabled the fire.
The voltage rating is another hidden variable. A 250V AC fuse applied in a 125V DC circuit might appear acceptable, but DC arcs are much harder to extinguish because there is no zero crossing. The arc duration in DC can be three to five times longer than in AC for the same current. A 20A fuse rated for 250V AC that clears an AC fault in 6 ms might take 18 ms to clear a DC fault of identical magnitude. That extra 12 ms of arcing deposits roughly three times more energy into the fuse body. In a tight enclosure with limited airflow, that energy can raise the ambient temperature inside the fuse holder by 80°C, pushing nearby insulation into a thermal runaway zone. A telecom rectifier site lost a string of batteries because a DC fuse took 22 ms to clear a short-the arc vented through the fuse's ceramic wall, which had a manufacturing porosity of 4%, and the ejected molten silver ignited the polyethylene battery tray.
There is also the matter of the fuse's interrupting rating. Every fuse carries a maximum fault current it can safely break. A 100A fuse with a 10 kA interrupting rating, when subjected to a 25 kA fault, will not clear the arc; instead, the element vaporises so violently that the fuse body shatters. The resulting arc then transitions to the surrounding air, sustained by the available fault current until some other device opens or the source impedance limits it. That condition is a genuine fire starter. In a low-voltage industrial panel, a 600A class J fuse was accidentally replaced with a 600A class H fuse that had only half the interrupting capacity. When a downstream welder shorted, the fault current reached 18 kA, the class H fuse exploded, and the arc flashed across the bus bars, igniting the dust accumulation on the panel floor. The fire was contained, but the lesson was clear: a blown fuse that blows apart is more dangerous than one that remains intact.
Thermal imaging of blown fuses reveals another nuance. After a fuse clears, its resistance rises to infinite, but the holder's contacts remain hot from the preceding overload. If the overload was gradual-say, 110% of rated current for several minutes before the fuse opened-the entire fuse assembly may have reached 180°C before the element melted. The fuse opens, but that stored heat does not vanish; it conducts into the wiring terminals and can raise the adjacent conductor temperature above its rated 90°C for many seconds after the fault is gone. In a test I reviewed, a 200A fuse that opened after a 4‑minute overload caused the terminal block temperature to peak at 195°C, which softened the PVC insulation on the outgoing cable to the point where the conductor shifted under vibration, eventually grounding against the enclosure. That ground caused a second fault, but the original fuse was already blown-so the second fault relied on a backup fuse, which also blew, but the heat from the first event had already compromised the cable.
The one scenario where a blown fuse undeniably starts a fire is when the fuse is not the actual protection device but a sacrificial component in a poorly designed circuit. Some aftermarket automotive accessories use a fuse as a fusible link in the ground path, expecting it to open on overcurrent. But a ground-side fuse, when it blows, leaves the entire load side floating at supply potential, and the open fuse holder can arc to the chassis if the gap is insufficient. That arcing is sustained not by the load current but by the leakage through the controller's internal capacitors. A car audio amplifier with a 50A ground fuse blew that fuse during a bass transient; the resulting arc between the fuse clip and the chassis lasted 40 ms and melted a hole in the floor carpet, which smouldered for several minutes before the driver noticed the smell. The fuse performed its primary function-it opened-but the arc path it created was external to the fuse itself.
So the direct answer is that a correctly specified, properly installed, and undamaged fuse will not start a fire when it blows. The fire risk comes from the system around it: holder degradation, voltage mismatch, insufficient interrupting capacity, or mechanical breakage that allows the arc to escape. A blown fuse is a symptom, not a cause, but if that symptom is misinterpreted-if a mechanic replaces a blown 15A fuse with a 30A one to avoid future trips-then the next fault will not blow the fuse at all. Instead, the wire will overheat, the insulation will char, and that charred insulation will ignite. In that case, the fuse did not blow; it was bypassed by an oversized replacement. The fire was not from a blown fuse but from the absence of a proper one. That distinction is everything in the field, and it is the reason my first check after any blown fuse is not the fuse itself but the fuse holder, the clips, and the wire gauge leading to it. Those components tell the real story.