The end of his tweet seems to indicate they're looking into the very low power Proof-of-Stake alternatives, potentially Ada/Cardano, Dot/Polkadot, or waiting for the Eth2.0 switch (Since the wheels are finally properly in motion for that and looking feasible)