Controlled thermonuclear fusion is potentially a major vast new energy source. A reactor based on nuclear fusion would be inherently safe, environmentally friendly, and fuels are cheap, abundant and widely available. The JET tokamak experiment has approached the plasma conditions needed in a thermonuclear reactor based on magnetic confinement concepts. In single deuterium discharges, breakeven has been achieved and, for the first time with deuterium-tritium fuels, ~1.7MW of fusion power was achieved in a 2s pulse. The total energy release was 2MJ. These results were obtained transiently, limited by an high impurity influx. For long pulse high power operation, plasma dilution has been identified as a major threat to a reactor. Improved impurity control in the pumped divertor configuration in a New Phase of JET (1992-1996) is envisaged. Experimental results support a plasma model based on a single phenomenon and MHD limits. Together, these are used to define the size and operating conditions of a reactor. A Next Step device would demonstrate the scientific feasibility of ignition under reactor conditions and this is discussed within the context of an international collaborative programme.