This article investigates the welfare implications of the rapid innovation in central processing units (CPUs), and asks whether it results in inefficient elimination of basic personal computer (PC) products. I analyse a game in which firms make multiple discrete product choices, and tackle challenges such as partial identification and sample selection. Estimation results demonstrate that the demand for PCs is highly segmented, and that fixed costs consume a substantial portion of the per-unit producer profit. The estimated model implies that Intel's introduction of its Pentium M chip contributed significantly to the growth of the mobile PC segment and to consumer welfare. The lion's share of the benefits to consumers was enjoyed by the 20% least price-sensitive consumers. I also find that the Pentium M crowded out the Pentium III and Pentium 4 technologies, and that the benefits to consumers from keeping those older products on the shelf would have been comparable to the added fixed costs. While total welfare cannot be increased by keeping older technologies on the shelf, such a policy would have allowed the benefits from innovation to “trickle down” to price-sensitive households, improving their access to mobile computing.