MapR's failure represents nothing.
The Age of ARM


Some people are just hitmakers. When it comes to microprocessors, Jim Keller, is a hitmaker. Having worked on the AMD Athlon he drove a sword into Itanium with the K8 x86-64 bit architecture before leaving AMD for new opportunities.

There was a route through PA-Semi (Apple’s processor acquisition) before a boomerang back to AMD to work on Zen. Zen and its successors are also hits, so much so that AMD is on track to reach the heady height of 20% desktop/notebook CPU marketshare. Something they last did two decades ago back in the Athlon days.

Unlike the Athlon, where PC enthusiasts and budget conscious consumers sought out those processors, the new new thing is the growing AMD server CPU business. Years ago AMD failed to crack the fortress built of market development funds that Intel placed around the tin benders of the server business. That crimped AMD's server growth even when they had a winning line of chips.

In the age of hyperscale providers, where offering multiple CPU types is seen as advantageous to consumers, this type of fortress is a remnant of a bygone age.

Intel’s server CPU business has long been a jealously guarded high margin treasure but now AMD have shown up to make off with some of it. AMD started at 1% server CPU market share in 2017 and have grown to 8% in two years. Could they slice off 10 or 20% of the server market? It’s possible.

One minor wrinkle in all this is after bouncing around Tesla for a split second Jim Keller took a new job.

At Intel.

One of the people who dynamited Intel’s most high profile failure is now working to restart their invention engine.

Lets see if a tired old chip maker has one more hit left in them.