If you were with us in November at Structure, you heard a lot of people talking about ditching their data centers and embracing the cloud. But what about companies that are going the other way?
TechTarget has an interesting look at what I call "the data center tipping point," in which companies born on the cloud start to realize they can run more efficiently in their own data centers after achieving a certain scale.
AI IS SUPER HOT, SO WHERE ARE ALL THE CHIP FIRMS THAT AREN'T NAMED NVIDIA? Nvidia, which appeared at the aforementioned Future of AI Symposium,
put out a post this week highlighting the work it's doing on chips for artificial intelligence applications, which are used by AI stalwarts such as Baidu, Facebook, Google, Microsoft.
Fortune unpacks Nvidia's post and examines the general state of chips for AI, noting "when it comes to AI, hardware is a huge differentiator, and that gap is widening all the time."
A DISPATCH FROM CLOUD CITY - STATE OF THE UNION 2016 Adrian Cockcroft's state of the cloud is a hallmark of Structure each year, and analyst
Charles Fitzgerald weighed in with his own take this week on the cloud market heading into 2016. One interesting suggestion: if Amazon Web Services wants to convince customers it isn't bent on locking them into its services, it will have to become more transparent, and that could be a tough thing to embrace inside Amazon's famously guarded corporate culture.
AMD FINALLY READY TO CRACK THE DATA CENTER WITH ARM CHIPS We spent part of last week's newsletter talking about the failure of ARM data center chips to make any sort of headway against Intel (just look at Intel's fourth-quarter results,
as reported by Bloomberg), and AMD's newest ARM-based server processors are finally available. But this sounds like too little, too late, for both AMD and ARM: "The A1100 is not an exciting product," longtime chip analyst Linley Gwennapp
told Venturebeat.
YAHOO RELEASES THE LARGEST-EVER MACHINE LEARNING DATASET FOR RESEARCHERS Yahoo may be at a crossroads in its long history, but it is still one of the largest sites on the web with a treasure trove of user data.
It announced this week that it will donate 13.5 terabytes of data to the machine-learning research community based off user interactions with its home page, something we'll be sure to ask Yahoo's Ron Brachman about when he's on stage at Structure Data 2016.
LOOKER ANALYTICS PLATFORM SCORES $48 MILLION LED BY KLEINER PERKINS The sweet spot in data analytics tools? Creating something that non-technical users can comprehend while giving the technical people some powerful tools to play with. Looker appears to have found this balance, and now has another $48 million ahead of what's going to be an uncertain year for fund-raising to continue building its company,
according to TechCrunch.