You have to admire AMD. They know when and how to be agile, mainly when decks are stacked against them.

AMD recently announced they are teaming with a different chip designer – ARM, the current darling of mobile and energy efficient processors – to make ARM’s products 64-bit. Those with good memories will recall that AMD practically invented the modern 64-bit processor industry when they released their Opteron chip and caught Intel sitting on their laurels (which must have been uncomfortable). This eventually forced Intel to license AMD’s 64-bit instruction set (which must have been uncomfortable). I was leading SuSE Linux’s North American strategy at the time and saw Intel staff faces wince when SuSE and Microsoft took the stage at the Opteron release event (which was obviously uncomfortable).

AMD didn’t keep their momentum and have had a number of setbacks. With desktops dwindling in numbers as consumers buy slabs and smart phones, and with bulk web servers drifting to energy efficiency instead of raw horsepower, AMD saw big chunks of their market evaporating. Rather than fight the trend, they found the best, cheapest and fastest way to ride the wave by partnering with ARM.

You could say the market strong ARMed them, but that pun would be too corny even for me.

Competing in fast moving markets means not wasting time when shifts occur. Granted you don’t want to prematurely jump on a bandwagon being pulled by dead horses, because whipping them doesn’t make the wagon roll faster. But wait too long and the market changes again … you are cut out of any market relevance. AMD almost waited too long given how pervasive mobile devices and green data centers have become. But Intel has lagged on both fronts and hasn’t proffered anything to slow ARMs dominance. AMD knows the intricacies of 64-but multi-core processing and has a patent portfolio to prove it, but faced a steeper uphill climb than Intel. So they teamed with ARM to accelerate 64-bit low-cost servers and this technology will drift into mobile.

Which must be uncomfortable for Intel … again.

Whole product definitions are the key to creating relevant products for which people will hurl money at you. In high tech, nobody can deliver all pieces of a whole product. That’s why partnership are common in tech. AMD and ARM separately couldn’t create energy efficient or competitive mobile 64-bit processors in the short timelines markets demand. But together – shoving AMD’s technology into ARM’s market leading ecosystem – will get both of them to the finish line fast.

The marketing lesson is with each whole product definition, review the market requirements document and decide if you will build, buy, partner or ignore each feature. If building takes too much time, buying is too costly or there isn’t an acquisition candidate, and if you cannot ignore the feature … then you must partner. Do so fearlessly because the market doesn’t give the timid a second chance.