Already in the U.S., Verizon and AT&T have been vocal about their interest in deploying 5G trials that will use these new, higher frequencies (sometimes referred to as millimeter-wave bands), and other governments are quickly lining up with similar proposals. In December, the World Radiocommunication Conferences (WRC) approved the planning and ratification of the global use of these wavelengths for cellular by 2019. And in recent weeks, Japan and South Korea both announced plans to have multi-gigabit-per-second 5G cellular systems in place for the Olympic Games each country will be hosting—summer 2020 in Tokyo and winter 2018 in PyeongChang.
The future 5G at mmWave will require MIMO operation to support multiple independent data streams and enhance spectral efficiency. Ideally, the channel throughput can be increased proportional to the number of spatial beams. New hybrid MIMO architectures are being studied as an alternative for fully digital precoding, aiming at the possible reduction in the number of RF chains and ADCs/DACs. To our knowledge, there are only two 5G ready testbeds: one from 5G Innovation Centre at University of Surrey (UK) and the other a Massive MIMO testbed at Lund University (Sweden). While the former only allows for post-OFDM waveformes demonstration, the latter has been devised at sub-mmWave (below 6 GHz) and limited to 20 MHz of system bandwidth, despite implementing Massive MIMO without beamforming.
Researchers at IT-Leiria (Prof. Rafael Caldeirinha) and IT-Aveiro (Prof. Carlos Ribeiro and Pedro Cruz), in close collaboration with University of South Wales in the UK (Prof. Akram Hammodeh), are currently working on a hardware demonstrator of a complete MIMO framework (single- and multi-user/multi-beam) for 5G mmWave wireless communications, which will allow to select and test on-the-fly new post-OFDM waveforms and monitor in real-time relevant KPI of multi-gigabit data stream. The testbed is thought to be highly scalable to easily allow extension to Massive MIMO configurations and/or integration with other RF frontends.
The answer to the question is, supposedly by 2020.
There are multiple different trials ongoing or planned for 5G, in different parts of the world. I doubt anyone has pinned down exactly what techniques will become the standard(s), although some candidates often mentioned are massive MIMO and FBMC, as well as channels up in the multiple GHz and multiple 10s of GHz.
On the other hand, you also seem to have answered your own question?