Investing in proprietary optical fibre, and building the fusion splicer in-house (1980)

Owning the tool that joins them

The fusion-splicer research, begun by three people, appears to have been the part of this decision that told most. Optical fibre itself came out of joint research with the telegraph and telephone public corporation, in which Furukawa Electric and Sumitomo Electric took part alongside Fujikura; it was never Fujikura Densen’s alone. But no one owned the device that joins fibres in the field. Cable, like copper wire, is a product that drifts toward a contest of volume and price, and holding the tool that sits outside it left earnings that were not at the mercy of raw-material markets. Reading the inconvenience of splicing as something to sell turned out, in the event, to be right.

The effect of betting on light, however, was long in coming. Forty years passed between reaching 0.27 dB/km in 1979 and ultra-high-count cable taking its place at the centre of the data centre. In the meantime rivals caught up on thin, many-core designs — President Okada remarked in a 2022 interview that others have lately begun doing it too. A policy of leading on technology demands both the stamina not to cut research spending and the patience to wait until demand arrives. The conviction that “there is no way to compete but on technology” points less to the boldness of the choice than to the difficulty of sustaining it.