Dear Reader,

Our whole economy is based on the idea that competition leads to innovation. Interestingly, the cause of innovation is best served if no one ever wins that competition. A company that truly succeeds in defeating all its adversaries gets complacent, and innovation tends to dry up.

For much of the PC era, Intel and AMD dueled for market share in the competitive PC industry. Intel was always bigger – well, way bigger – but AMD found ways to stay competitive. A computer with an AMD processor cost just a little less, which gave a push to demand, and AMD could match Intel with some serious engineering chops, which kept the competition lively. Intel would come up with some new innovation, and AMD would find a way to do something similar, or at least close enough to achieve near-equivalent performance. And AMD always put the emphasis on economy, forcing Intel to stay real with pricing.

The competition continued for several years around the double axis of performance and price. But sometime a few years ago, AMD started to fall behind. A miscalculation? An economic downturn? Maybe Intel just got so big it sucked up all the attention? Or maybe the price for a PC processor got so low that it wasn't as easy for AMD to slide under the Intel price? For whatever reason, AMD stock dropped to $1.67 per share sometime in 2015 – a fraction of its previous value. Many of us who believed the presence of AMD was a very good thing for the industry started to wonder if AMD would have a way to keep marching, but the company did exactly what one would hope they would do at such a moment: they called their best minds together and came up with a plan.

That plan has been exciting to watch. AMD's Zen architecture arrived in 2016, and the Ryzen processor series debuted in 2017, which was, by all accounts, a very good year for AMD. Revenues are up, and a share of stock is up to $16 – 10 times the previous low. The Radeon Open Compute (ROCm) initiative unveiled a whole new open framework for GPU computing, which plays to AMD's strength in the graphics processor space.

I don't know how many high-tech companies have gotten to the edge of disaster in recent years and gone right on over the edge. Nokia CEO Stephen Elop wrote his famous "Burning Platform" memo in 2011 to rally the troops and save the company [1], but it didn't work: The recommended plunge into icy waters could not forestall the final combustion.

AMD and its CEO Lisa Su have my admiration for working out a plan and keeping their focus on bringing the company back to stability. Comebacks are easy to wish for and not so easy to accomplish in real life.

Meanwhile, though, a third contender has entered the ring. ARM processors are a common feature of cell phones and tablets, and now that phones and tablets claim an ever-larger share of work that was once performed on PCs, the ARM universe is getting ever-closer to direct competition with Intel and AMD.

And the ARM phenomenon is seriously spilling out of the once-isolated mobile space. At the end of 2017, Microsoft rolled out a new line of ARM-based Windows 10 laptops. Even the speed-conscious high-performance computing industry is starting to look seriously at ARM, and the first petascale ARM supercomputer went online earlier this year [2].

The best feature of ARM processors is their efficient power usage. Low power means that a smartphone can ride around in your pocket longer without a charge. It also means (at least theoretically) you can pack more processors into that supercomputer without it overheating or using all the available energy.

Intel and AMD never had to be serious about optimizing for power usage in past generations, but now it looks like they will have to. And if they do, a new triple axis of performance versus price versus power could lead to a new generation of better, less expensive processors.

Joe Casad, Editor in Chief