AMD is coming out with its new Naples processor for servers. After being knocked out of the market due to missed commitments, getting back in would typically be ugly. Server buyers tend to be very conservative and would be slow to move to a new processor anyway. AMD is taking a very interesting path to market that pulls from its core differentiating feature and builds on its increasing financial success to overcome this shortcoming.
Let’s chat a bit about this because the process is unique and I think it might represent a best practice for how to speed an enterprise-class product to market and assure customer advocacy for it.
AMD’s Sustaining Advantage over Intel
If I were to sum up one of the top competitive advantages that AMD has had over Intel, it is that it asks OEM customers what they want and then tries to build it. Intel tells its customers what it’s going to build. Historically, there has been more tension between OEMs in all segments and Intel with regard to direction and policies, while AMD’s issues have historically been more about meeting roadmap promises.
In short, AMD, being smaller, has historically been more responsive and often favored in terms of relationships, but this has been offset by an historic problem with execution. Lisa Su, AMD’s current CEO, who came from IBM, has been addressing AMD’s execution issues with near amazing success. AMD’s steadily improving financial performance supports these results and if AMD can convince OEMs that they will get the relationship they want, with the reliability they need, getting design wins shouldn’t be a problem. Clearly this is being showcased in its game console, PC, and graphics car businesses.
But getting these systems into companies might still be problematic.
In general, new systems like this have a cart and horse problem, in that large companies don’t want to buy technology in this class until it has been proven in large companies, making it nearly impossible to get around an entrenched vendor. However, we have a new class of companies at scale that are far more open to new technology, largely because their performance and cost needs are typically well ahead of where technology is currently performing. These are the massive cloud providers that are building data centers the size of small towns with energy requirements that rival some cities. They will often take products that haven’t even completed a beta cycle of aggressive testing and sometimes even use them in order to address these critical needs. And given that they tend to have cutting-edge fail-over capability tied largely to their critical ability to shift loads of all sizes near instantly, their environments are far more tolerant of as-yet untested technology.
What AMD is doing is putting its parts in development into these environments under load to assure that they are reliable and that the parts are tuned to the unique requirements of these massive cloud providers. In theory, if the coming Naples offering is specifically designed to handle affordably (in power used and cost) the massive variety of loads of a major cloud provider, then it should not only be ideal for on-premises cloud offerings but for a wide variety of similar requirements from enterprise buyers.
In short, by the time Naples comes to market, there should be a number of high-visibility, high-load cloud companies that will be happy to advocate that it performs as promised, potentially effectively addressing that cart and horse problem.
Wrapping Up: Best Practice
Generally, I think it ideal, particularly for a new product, to partner with end customers in that product’s development because it should assure a far better match to the needs of the related market. I’m not sure this process should ever be broken; it used to be really hard to collaborate in this fashion because doing so securely across company lines, particularly through OEMs, was virtually impossible. But given that cloud providers often work directly with companies like AMD, and collaboration tools have massively improved over the last decade, suddenly there is an opportunity for even parts makers like AMD to build relationships with those that use their products. Through this, it can uniquely understand current and emerging needs and come up with products that can address those needs.
I also think this process was likely refined in AMD due to its unique position with Microsoft and Sony’s game console divisions, which have to work tightly with the parts providers because, although their solutions are very different from servers, they too have to push the performance envelope while keeping costs down.
It won’t be until next year around this time when we’ll know if this process is actually effective for Naples, AMD’s new server part. But it should be, because working closely during development with those who will use a new product has been proven in most professions going back hundreds of years. Apparently, we in tech are sometimes slow learners. So, what is strange about this path is that it is unusual. I think everyone should work with users to develop new products. Just saying.
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+