Security from the ground up
Since Brian Krzanich took over Intel as CEO, he has spent approximately $33B on acquisitions. The largest acquisitions were as follows: $15.7B for MobilEye that had experienced more design-outs than new design wins, $16B for Altera–great company, but not a portfolio fit, $225M for SoftMachines, and over $300M for Nervana Systems.
I believe that for a company the size of Intel, an acquisition total of $33B within four years is not only acceptable but demonstrates a capacity to prepare for the future. However, Intel still needs to advance on some of its prior faults. The tick-tock strategy, involving alternating cycles in which the process and the architecture were updated, appears to have been phased out. New products aside from the usual x86-64 for desktops, laptops, and servers have not been observed. In fact, the n-th iteration of Larrabee (this time the Xeon Phi Knight’s Landing) has just combined Knight’s Corner with a set of Xeon cores. In the tradition of announcing groundbreaking new I/O such as OmniPath, it followed its predecessors into decline. QPI switching never materialized, and Omnipath will similarly not be productized.
Additionally, China challenged Intel with the world’s fastest supercomputer, based on homegrown processors that exceeded Intel by a factor of three in performance at roughly the same power consumption. Furthermore, AMD, who is counted out by most, has now out-Intel-ed Intel itself. The AMD Ryzen desktop product line competes favorably with Intel’s counterparts at lower prices. Now, AMD has shown precursory Threadripper server processors that appear to outperform Intel’s competing products, also at a lower cost. Spending $33B for a company the size of Intel is not a problem, but it has to produce results.