Conclusion

Competition between "tube" and "solid-state" technologies has existed since the 1970s, and the boundary between them has been slowly evolving over time towards higher power and higher frequency, in favor of solid-state solutions. This progression has never been linear, but has been marked by the successive appearances of bipolar silicon, MOS silicon, GaAs and now GaN field-effect transistors, to name but the most important milestones. We are now entering the GaN era, which is likely to last one or two decades. Tubes, for their part, have made enormous progress, notably in terms of frequency and efficiency, but they remain an expensive technology whose use must be justified by specific needs.

In a few rare fields, tubes are the exclusive solution, due to the power involved (thermonuclear fusion, for example). But these are niche markets, and it is more interesting...