A detailed KrF amplifier model is first benchmarked against new data and then used to design higher-energy modules with segmented pumping. It is found that segmentation with unpumped regions does not carry with it any penalty in efficiency because the distributed geometry has reduced losses from amplified spontaneous emission (ASE) that counteract the fluorine absorption of unpumped regions. A 68-kJ module is designed, incorporating a new water line geometry and a combined switch/bushing. The electrical parameters of the module are calculated in detail. The effect of multiplexed beam-line energy on facility size is discussed, and an energy of 100 to 200 kJ is found to be optimal. Two 68-kJ modules are combined in a 136-kJ multiplexed beam line, incorporating incoherent spatial imaging, that fits within a compact beam tunnel. A total of 16 such beam lines are arranged on four floors to deliver 64 beams to a target; the net energy is 2.0 MJ. Detailed calculations of prepulse ASE energy are given, and the levels are designed to be low enough not to initiate a prepulse plasma. The basic geometrical uniformity of target illumination is shown to be better than 0.3% for a 64-beam illumination geometry that has a high degree of symmetry. A test of the 68-kJ module would be necessary to verify the projected specific laser energy and facilitate more detailed design of this fusion laser.