HPE Acquires SGI For $275 Million (venturebeat.com)
An anonymous reader writes: Hewlett Packard Enterprise has announced today that it has acquired SGI for $275 million in cash and debt. VentureBeat provides some backstory on the company that makes servers, storage, and software for high-end computing: "SGI (originally known as Silicon Graphics) was cofounded in 1981 by Jim Clark, who later cofounded Netscape with Marc Andreessen. It filed for Chapter 11 bankruptcy in 2009 after being de-listed from the New York Stock Exchange. In 2009 it was acquired by Rackable Systems, which later adopted the SGI branding. SGI's former campus in Mountain View, California, is now the site of the Googleplex. SGI, which is now based in Milpitas, California, brought in $533 million in revenue in its 2016 fiscal year and has 1,100 employees, according to the statement. HPE thinks buying SGI will be neutral in terms of its financial impact in the year after the deal is closed, which should happen in the first quarter of HPE's 2017 fiscal year, and later a catalyst for growth." HP split into two separate companies last year, betting that the smaller parts will be nimbler and more able to reverse four years of declining sales.
For of those in need of an SGI history lesson, the SGI currently in business is not quite the same SGI (Silicon Graphics) of old. Remember that Rackable Systems acquired Silicon Graphics back in 2009 for like $20M I think? And they turned around and renamed Rackable to SGI.
Karma: Bad
I can't disagree. They spent millions to develop their Altix and UV lines (big shared memory systems, thousands of cores).
While they sold clusters, it was half hearted at best. They eventually stopped updating the cluster hardware to "focus" on their behemoths. A lot of the Crayons (ex cray engineers) at SGI hated clusters with the passion of a thousand suns.
They were also amazingly overbuilt systems - custom designed fans instead of commodity fans. Their chassis were built like tanks - they used an enormous amount of steel, and weighed 4-5x more than anything else on the market.
A lot of the engineers were so blinded by their hatred of clusters that they forgot that price matters to a customer. If you're doing the same job, and one vendor costs 5-6x more for the EXACT SAME OUTPUT, you've got to expect that one vendor to lose most bids.