It’s mostly a consequence of how common silicon CCD/CMOS sensors work, not just an arbitrary design choice.

Silicon sensors respond naturally across a broad range of wavelengths, including into the infrared, which is why cameras need IR-cut filters. Their response is not flat: sensitivity rises through much of the visible range and then falls off again at shorter wavelengths. Blue light has shorter wavelengths and higher photon energy, where silicon’s efficiency is already dropping; even shorter wavelengths such as UV are lower still.

On top of that, cameras use a Bayer color filter array to turn the sensor’s monochrome response into RGB. The blue-filtered channel typically has a lower peak response than red and green, so less signal reaches that channel. To produce a balanced image, the blue data often needs more gain, which also boosts noise.

Typical lighting makes this worse: sunlight and especially tungsten/incandescent light contain relatively less blue, so the blue channel starts with less signal before amplification.

So the lower blue sensitivity is largely rooted in the physics of silicon sensors and the color-filter system, though exact behavior varies by sensor and filter design.