Earlier this week, in conjunction with ARM Holdings plc’s announcement of the upcoming Cortex A53 and A57, full 64-bit CPU implementations based on the ARM V8 specification, AMD also announced that it would be designing and selling SOC (System On a Chip) products based on this technology in 2014, roughly coinciding with availability of 64-bit parts from ARM and other partners.
This is a major event in the ARM ecosystem. AMD, while much smaller than Intel, is still a multi-billion-dollar enterprise, and for the second largest vendor of x86 chips to also throw its hat into the ARM ecosystem and potentially compete with its own mainstream server and desktop CPU business is an aggressive move on the part of AMD management that carries some risk and much potential advantage.
Reduced to its essentials, what AMD announced (and in some cases hinted at):
- Intention to produce A53/A57 SOC modules for multiple server segments. There was no formal statement of intentions regarding tablet/mobile devices, but it doesn’t take a rocket scientist to figure out that AMD wants a piece of this market, and ARM is a way to participate.
- The announcement is wider that just the SOC silicon. AMD also hinted at making a range of IP, including its fabric architecture from the SeaMicro architecture, available in the form of “reusable IP blocks.” My interpretation is that it intends to make the fabric, reference architectures, and various SOCs available to its hardware system partners.
- It seems like a good bet that one of the first reference architectures will be a SeaMicro implementation with a scaled out fabric. I am still having trouble convincing myself that SeaMicro will exist in the long term as a server product competing with AMD system partners, but it will certainly be an excellent platform from which AMD can design and prove its reference architectures and other IP for its partners.
The primary risk is revenue cannibalization and margins: It will be hard to get the same gross profit or margins on ARM-based products due to the highly competitive nature of the market. Pricing will not be protected by an Intel umbrella, but rather will be a competitive, in the true sense, market price that will be forced to respond to other ARM vendors' pricing.
Balanced against this risk are strong potential benefits for AMD, which has never come close to regaining the momentum it had in the server segment since the glory days of beating Intel to market with HyperTransport, 64-bit x86 extensions, and dual-core CPUs. Among the obvious benefits are:
- An opportunity to be an early mover in rapidly growing hyper-dense and energy-efficient server segments. Although there are others in this still nascent segment, such as Calxeda and Applied Micro, the entrance of AMD changes the potential for enterprise and large cloud service adoption radically. While AMD is a mere shadow of Intel in terms of financial resources, it still dwarfs the combined total of potential investment by other ARM server companies, and if it shares its IP, it will be in a position to accelerate the adoption of ARM servers.
- Competitive differentiation versus Intel. Intel could introduce an ARM SOC of its own. But my guess is that it will stay the course and drive its future x86 designs further in energy efficiency and price performance. If it does this, AMD will be able to stake out significant share in a much smaller but high-growth segment. AMD will probably be able to catelyse the development of a whole tier of system partners to produce white box ARM servers, including many of the “local hero” companies that are beginning to eat up market share from the global vendors in emerging markets.
Who gets hurt? This announcement obviously pressures early movers like Calxeda and means that any current ARM system partners such as Dell and HP must re-evaluate their strategies in light of AMD's plans. Companies like Calxeda are not doomed, and much depends on how good AMD’s ARM SOCs are, but the entrance of an established global competitor into their niche segment cannot be a cause for celebration.
Server buyers will have to wait until sometime in 2013 to get any clarification of AMD’s plans for what is positioned as a 2014 launch of full 64-bit ARM servers. In the meantime, both system vendors and users can continue to assess ARM technology with the current Cortex A9 and possible upcoming A15 system offerings, since the future 64-bit products are expected to be fully software compatible with current 32-bit products.
All in all, for server buyers, as Mr. Rogers might say, “It’s going to be a good day in the neighborhood,” with more choices for massive energy-efficient data centers emerging over the next 18 to 24 months.