Eighty years from equation to silicon

The first mathematical model of a neuron as a threshold logic unit — the seed of every artificial network that followed.

Donald Hebb proposes that neurons that fire together wire together, the ancestor of spike-timing-dependent plasticity.

Equations describing how ion channels create the action potential, giving engineers a physical target to imitate.

Carver Mead builds analog VLSI circuits whose transistors mimic neural membranes, founding the whole field.

HP Labs demonstrates the missing fourth circuit element, offering a device that both stores and computes a weight.

One million digital neurons and 256 million synapses on a chip drawing about 70 milliwatts in real-time operation.

Manchester's machine wires up hundreds of thousands of ARM cores to simulate biological networks in real time.

128 neuromorphic cores with 130,000 neurons and programmable on-chip learning rules, including STDP.

A second generation with graded spikes and faster circuits, paired with the open-source Lava software framework.

Intel assembles 1,152 Loihi 2 chips into a 1.15-billion-neuron research system for large-scale brain-inspired AI.