Electronics Weekly Electronics Design & Components Tech News
He writes:
After three decades in the industry I know that whatever I predict for the next six, I’ll be wrong. But it’s fun to contemplate.
Two mega-trends face us: at the macro‑level, the resurgence of nationalism and deglobalisation; at the micro – within the semiconductor industry – reaggregation of the value chain and resurgence of asics.
In both cases, pendulums swing. The ebb and flow between asics and general‑purpose devices, between custom and generic, is characterised by Makimoto’s wave with its ‘roughly a decade’ cadence. This coincides with a swing to vertical integration and away from industry layering and focus.
The geopolitical pendulum swings more slowly. Global trade intensity only recently passed its 1913 peak and is now falling back. Trade wars, rising nationalism and practical repercussions of Covid‑19 mean a move to more local sourcing.
Geopolitical trends are complex for business and bad for societal progress and the global community. But the ‘verticalisation’ of the semiconductor industry is more exciting, putting the focus on the system, on value and differentiation. The emphasis moves towards innovation in custom hardware (or hardware-software co‑design), further driven by applications that must be more security‑sensitive and safety‑critical.
In today’s environment there’s a need and a desire to do the hard stuff in‑house, under control. The incredible cost of deep sub-micron means fewer people can afford to build those devices, so bigger, richer companies dominate.
We move from a situation that looks like natural selection working on random variation, to an era of ‘intelligent design’.
That is both a promise of radical change and incredible ideas, and a concern that the loss of diversity, of experimentation and novelty can lead to consensus and corporate dominance. You get the paradigm shift in architectures but lose the Darwinian diversity of ideas.
In previous decades, vertical integration saw AT&T Bell Labs invent the transistor, Shannon’s Information theory, C and Unix. But it led to ‘innovation rationing’, where diversity was limited (think of the ‘plug‑compatible computer’ concept in the PC arena).
Which of those precedents will win? What will technology from Google, Apple, Facebook, Huawei, Alibaba and Baidu look like? How will that interact with a less global world? With American and Chinese ecosystems diverging in an unprecedented way, how will Europe, India and others navigate this less global world? A two-pendulum system is notoriously chaotic, so the future is as hard to predict as ever.
