Chip giant Intel (NASDAQ:INTC) recently announced a refresh of its Compute Stick lineup. The company refreshed its Atom-based models, moving from its rather dated Bay Trail architecture to the newer Cherry Trail architecture. The company also introduced Compute Stick models with its higher-end Core m-series of processors.
I think that the Core m-based Compute Sticks are quite good and everything that I could have reasonably expected and hoped for. However, the Atom-based model looks hugely disappointing for one key reason.
Still with the Cherry Trail?
A while back, an Intel Compute Stick product roadmap hit the Web (via Liliputing). According to that roadmap, the company had planned to roll out a version of its Compute Stick using its next-generation Atom processor code-named Broxton.
Broxton, according to one Intel employee's LinkedIn profile, delivers a "completely new architecture" compared to the prior-generation Cherry Trail chip. The profile also states that Broxton "brings major improvements in power consumption, performance, and platform security features" from the prior-generation chip.
Indeed, Intel had said back in 2013 that Broxton would be ready for production by mid-2015, although this timeline was pushed out to 2016 at Intel's 2014 investor meeting.
At any rate, it's very disappointing to see that the company is using this now 1-year-old architecture in this new Compute Stick, especially since Broxton could have delivered a lot of much-needed user experience enhancements.
What kinds of enhancements are we talking about?
If you think about the Compute Stick, the main use case for such a product is as a computer plugged into a monitor for what is basically a low-cost, small-footprint, all-in-one PC replacement.
Such a device isn't going to be used for computing- or graphics-intensive tasks, but would instead likely be used as a general-purpose machine. Typical application use cases here would be Web surfing, video playback, and casual/light gaming.
Broxton is expected to bring a significant jump in CPU performance over Cherry Trail, which would have certainly improved the web browsing experience of the Compute Stick. Broxton is also expected to feature a derivative of the Gen. 9 media and graphics engine found on the company's latest Skylake processors, which would have helped with both video playback (Gen. 9 supports more advanced codecs in hardware) as well as with games.
At any rate, a move to Broxton would have put the Compute Stick at least in the same ballpark as today's top mobile devices in terms of performance and features. With Cherry Trail inside, it's not even close.
When should we expect the Broxton-based compute stick, then?
It's not clear if Intel would actually refresh its Atom-based Compute Stick at some point in the middle of the year, which is when I would expect Broxton to start hitting the market. I believe that doing so would be a good move, particularly as it would be nice to refresh the product line for the back-to-school season.
Additionally, Intel had said previously that Broxton would drive a significant platform bill of materials cost reduction compared to its Bay Trail platform. Although Intel probably managed to bring that cost down significantly with Cherry Trail, I suspect that Broxton will bring a further bill of materials cost reduction.
This means that Compute Sticks based on the Broxton platform might not only deliver significantly better performance/features, but they could actually be cheaper to build. A move from Cherry Trail to Broxton in the low-cost Compute Stick could allow Intel to either pass those savings on to the consumer (lower prices could potentially drive volumes up), or it could keep those savings as additional margin, helping the bottom line a bit.