The man wandering through the vineyard looks lost. He’s brandishing a handheld sensor that resembles a self-checkout scanner in the supermarket or perhaps a flashlight. He pauses as he walks down the rows of vines, holding the device carefully against bunches of grapes. There’s no flash or beep as he does so; inside the grapes, though, beams of shortwave light from the gadget agitate molecules of anthocyanin. The combination of wavelengths is tailor-made to resonate with these color compounds, which indicate a grape’s ripeness.
The pigments fluoresce under the beam of the gadget, allowing it to log their presence, like a teacher asking pupils to raise their hand. It’s a process that’s repeated weekly, starting as the fruit begins to turn in July, with the readings banked to a central server. There, artificial intelligence compares them with historic and theoretical data: Is a given bunch riper now—higher levels of anthocyanins—or is it past peak, with lower levels? How do its current readings compare with parallel data taken the previous week?