A major concern of present-day tokamak design is the uncertainty in the plasma energy confinement scaling. We present sensitivities of designs similar to the International Thermonuclear Engineering Reactor (ITER) to changes in the energy confinement H-factor (H) for cases designed for ignition (the ITER physics phase), long-pulse, current-driven operation (the ITER technology phase), and achievement of both phases in a single device (bimodal). For cases that require ignition, the cost increases sharply for H-factors below 2. For technology-phase cases, the costs are much less sensitive to H-factor variations. For bimodal cases with H{ge} 1.8, the ignition criteria dominate if the energy multiplication factor Q need only be 5 in the technology phase; if Q {ge} 10 is required, the current drive criteria dominate. The bimodal cases are at most {approx}10% more expensive than the more costly of the physics-only and technology-only cases. Thus, the present ITER scheme of replacing the blanket and shield of the device between phases may not be the most cost-effective way to accomplish the desired performance goals. Investigation of device sensitivity to the H-factor indicates that, as the H-factor decreases from 1.8 to 1.2 for a given device, the achievable Q drops from {approx}10 to {approx}5more » and the divertor heat load increases by 50%. 8 refs., 4 figs., 2 tabs.« less