OpenPOWER: The First Fight Isn’t Intel, It’s ARM
I attended the OpenPOWER Summit this week, which was held in conjunction with the NVIDIA GPU Technology Conference (GTC). This event is part of IBM’s huge effort to provide an alternative to x86 on servers and it is using a strategy similar to what worked against Microsoft with Linux. In fact, Canonical (Ubuntu) was on stage sharing its experience with exactly how to do this. There was a lot of anti-Intel rhetoric, but it struck me that Intel really isn’t the first target. It is ARM, which is currently backed by HP in a product code-named Moonshot, in a vastly weaker effort to provide an alternative to Intel.
Interestingly, IBM effectively created Intel back when it worked with Microsoft to create the IBM PC. And early last decade, HP was Intel’s most powerful partner until Itanium, a joint effort between the two companies, failed. Now these two once powerful Intel allies are on different paths to create a strong alternative to Xeon and Intel’s dominance.
HP vs. IBM
IBM has been pared down, much like Apple was pared down in the 1990s by Steve Jobs into a very focused enterprise company. Its SoftLayer acquisition basically gives it three powerful pillars, each working together to hold the company up. Cloud services, on premise hardware and software, and traditional IT services create a comparatively focused set of products targeting well-defined enterprise-class companies and major industries.
Led by a CEO, Virginia Rometty, who was groomed throughout most of her career to lead the company, and a seasoned set of executives mostly equally trained and groomed at IBM, the firm is unique in the market and not only because it is one of a handful of companies over a century old.
IBM has pivoted on Power, having spun out its Intel-based x86-based server and client businesses to Lenovo, its efforts are tightly focused on creating a powerful alternative to Intel’s dominance.
In comparison, HP is a bit of a mess. HP has had a series of layoffs but its complexity largely remains the same decade over decade. It is going to attempt to spin out PCs and printers and along with that, most of its debt, but it is still uncertain whether investors will bless that move. HP not only continues to sell a massive number of x86 systems, but it hasn’t even been able to retire all of its Itanium efforts yet and it is planning to bring out yet another new technology, memristor, in the same time frame that it is attempting the ARM pivot. Doing one pivot at a time is incredibly difficult; doing two should be impossible even for a focused company.
Its executive leadership is all over the map, with a CEO who failed as a politician and bailed out of eBay while it was still relatively small, and a mismatched group of executives from a variety of companies with no apparent long-term connections to their CEO. It has little consistency, and the strongest unit is actually becoming the PC unit, which is the one it is saddling with debt and then dumping.
Add to this that ARM, while dominant on mobile devices, is basically nowhere on servers yet, compared to Power which, thanks largely to IBM, is actually a viable server platform today.
This imbalance is actually pretty pronounced, which is why at the OpenPOWER Summit, there was a huge showcase of actual products. The last time I saw Moonshot, it was basically a foam mockup of what an ARM server might look like, even though there are actual ARM servers in market. And the first working version used Intel’s Atom processor, which doesn’t give you a lot of confidence in the ARM version. This was similar to HP’s 3D printer effort at this year’s CES; every other product on stage and in the 3D printer showcase actually were products, while HP’s was a foam mockup of what it might look like. This was positioned as a manufacturing line level product, even though there is no way you’d be able to integrate something that used top loading on a manufacturing line. Kind of scary, really.
Wrapping Up: ARM Will Be Easy, Intel Hard
I don’t think OpenPOWER will have much trouble rolling over ARM on servers. There is simply too much of an imbalance and unless IBM dies laughing after looking at HP’s effort, it should easily trump HP’s pathetic effort. Personally, I think Meg Whitman has done what she did with politics and given up.
But Intel will be far tougher and the conditions that existed between UNIX and Microsoft and fueled Linux don’t exist between hardware makers and Intel at the moment. Intel is running a very tight ship these days. IBM needs to convince the ARM server faithful that their battle is over and rally them to the common cause of taking on Intel to truly get to critical mass. It made good progress this week and it has HP’s unintentional help, but I wonder if it shouldn’t be instead focusing on doing something even bolder, like flipping the market, which is how Apple eventually surprised Microsoft. I think NVIDIA may be key.
Let’s save that for next time. For now, IBM can revel in kicking HP’s butt down the block yet one more time as it trains to do the same to Intel, which will definitely kick back.
Rob Enderle is President and Principal Analyst of the Enderle Group, a forward-looking emerging technology advisory firm. With over 30 years’ experience in emerging technologies, he has provided regional and global companies with guidance in how to better target customer needs; create new business opportunities; anticipate technology changes; select vendors and products; and present their products in the best possible light. Rob covers the technology industry broadly. Before founding the Enderle Group, Rob was the Senior Research Fellow for Forrester Research and the Giga Information Group, and held senior positions at IBM and ROLM. Follow Rob on Twitter @enderle, on Facebook and on Google+
Read more on "Server Hardware Spotlight" »