With SeaMicro buy, AMD to double down on servers

3 Comments

SeaMicro's SM10000-64 server.


Updated
just confirmed the deal

Chipmaker AMD plans to announce its acquisition of low-power server maker SeaMicro, according to my sources. This would be a huge move for AMD (s amd), which has to double down in the server market since it has failed in the mobile market.

SeaMicro’s servers are aimed at the emerging webscale and cloud computing market, and are fundamentally different machines than those built by HP, Dell and IBM. The promise of the boxes is they cram a lot of compute into a tight space and consume less power. The demand for this type of product has been so great that Intel (s intc) actually designed a version of its Atom processor for the SeaMicro machines last year.

In January SeaMicro placed a traditional Intel Xeon server chip inside its boxes to enable the gear to take on a wider variety of data center applicaitons. At the time I wrote:

Tuesday’s announcement puts a traditional Xeon architecture chip inside SeaMicro’s boxes, and in doing so has remade the traditional server in SeaMicro’s vision. SeaMicro has whittled down a server into three component chips — the CPU, memory and SeaMicro’s proprietary ASIC that helps the hundreds of chips inside the box communicate. What’s notable here is how quickly and how powerfully web scale data center operators have turned the tables on the server and chip industry, which had long been dominated by the vendors delivering innovation at their pace.

So this buy gets AMD into the server business, even though SeaMicro’s core IP is around its customized chip that can handle the networking demands of more than 500 chips in a small space, and it also puts SeaMicro at odds with its former partner and chip supplier. And while SeaMicro CEO Andrew Feldman has told me on multiple occasions that any CPU could fit inside the boxes, this buy may force a quick change.

So far AMD and SeaMicro have not responded to my requests for comment. I will update the story later as I learn more.

3 Comments

Mark Lambert

I dont know… I think a very reasonable argument can be made that both AMD *and* Intel have failed in the mobile market. The space has clearly been ceded to tablet and smartphone form factors where anything but Intel and AMD dominate (ARM, Qualcom, Huawei, et al)

visitor

Please don’t say AMD failed in the mobile market, say never tried to enter into it or was late in thinking about entering mobile market… Words can make lot of meaning

Stacey Higginbotham

I would tend to agree with you, only it defined the mobile market and used to have a strategy all around netbooks. Then it kind of failed to evolve in mobile.

Comments are closed.