Email
PrintWhere are we in cloud adoption? A week ago with the cloud stocks falling apart, I started hearing we are at saturation level. Every company, every vertical, every geography has enough data centers to handle all the business to be done -- and we had lost one of the great secular growth areas of our time.
How did the saturation edict take place? I think it all started on Nov. 28, when Autodesk (ADSK) , which is a Software as a Service company based in the cloud for engineers and architects, hit the wall. The Autodesk bombshell was reminiscent of that day in February of 2016 when Tableau Software (DATA) and LinkedIn (LNKD) blew up, taking down everything cloud from Adobe Systems (ADBE) and Workday (WDAY) to Salesforce.com (CRM) and ServiceNow (NOW) . Yes, the red-hot Autodesk machine could have that impact when it shocked Wall Street with a sputter when we thought it was going at about 160 miles per hour.
It took four weeks for buyers to decide that the destruction of Tableau Data and LinkedIn on the same day wasn't the end of the world. In fact, it was a pause that refreshed, because LinkedIn soon got a bid from Microsoft Corp (MSFT) after being courted by Salesforce.com, and Tableau was reframed as more of a static data company than a real-time cloud machine.
Still, post-Autodesk, we are in fairly similar shape and I have taken to asking everyone -- including those down the food chain, like real estate investment trusts devoted to data centers, and data analytics companies like Alteryx (AYX) , a guest on Mad Money last night -- if there really is some sort of slowdown occurring we don't know about, some saturation that would kybosh one of the great growth stories of the era.
I am including in my canvas the big cloud companies, Alphabet (GOOGL) , Amazon (AMZN) , Microsoft and IBM (IBM) , as well as the companies that abet cloud use, like VMWare (VMW) , Adobe and Red Hat (RHT) , and even companies that are more infrastructure and Internet of Things, like Cisco (CSCO) and the semiconductor companies that are part of the big chain of cloud command for autonomous vehicles and machine learning.
Here's what I have come back with: We have about 10% cloud adoption in this country, with much of it in the financial and retail sectors, only a little bit in healthcare and a small portion of it as part of the Internet of Things.
If anything, given that the other form of data keeping, expensive on-premise data sites with little analytics attached to them, is still very much in use -- it isn't easy to flee from the entrenched systems -- I think we have years of growth ahead of us.
Even better, for those who think that 10% is too close to 100%, China, perhaps the second-most penetrated when it comes to cloud, is only about 2% to 3% penetrated, a ridiculously low figure considering the commerce that's burgeoning there by the minute.
Sure, there have been periodic glitches in those who provide hardware for the cloud.
For example, we know there is a pause in some data center build outs, as Nvidia (NVDA) starts to ship its faster, smaller, less heat-intensive data center chips -- a development you can learn about by going to Nvidia's website and looking up a new and informative paper called "Nvidia Deep Learning Inference Platform." It is possible that this new chip might have sent reverberations through those who select hardware for these centers. Both Advanced Micro Dynamics (AMD) and Intel (INTC) have competitive products, but I don't know anyone who would dispute that this new inference platform wouldn't be worth waiting for.
There's also the lingering theme that there might be too much flash in the system, the inventory overhang theory that Katy Huberty from Morgan Stanley put into our heads ten days ago that really resonated. How could there NOT be a slowdown in the data center if flash prices are coming down?
But after all my canvasing, I come back and say that there might be speed bumps in cloud adoption, but we're still at the beginning of the shift toward this method of storage that is growing faster than any other element of tech.
So I would use the weakness to buy everything in the food chain except flash, which is at the epicenter of the slowdown thesis. Why court danger in one element of theme when it is "all systems go" for the others?
