The competition
We moved on to the competition, and NVIDIA inevitably came up. King-Smith mentioned that he had seen our recent interview with Mike Rayfield, the GM of NVIDIA's Tegra business unit, so we asked him what he thought of Rayfield's assessment of Intel's low power processor strategy, which he summarised as "dehydrating a notebook".
"With some of Intel's earlier efforts in embedded it may have been regarded that that's what it was trying to do, but I would say that what we know of Intel, and the back-to-basics work they've done on Atom and the underlying technology, is impressive. It's a damn sight more than just shrinking a desktop PC.
"And we've been saying pretty much the same thing about [NVIDIA]; they're coming from this premium, high-end PC space, and using very traditional technology to try and come down into this mobile, embedded space.
"But Tegra isn't a competitor of ours, it's a competitor of our licensees. The difference is that the graphics on Tegra is NVIDIA's own, and we'd be amazed if they used anything else given their heritage. We're helping a lot of our licensees to compete against Tegra and it's so far so good as far as we can see.
"From an IP point of view, ARM is certainly one of the people we see. They did used to license our technology and a number of MBX licenses were initially sold through ARM. But those licensees ended up talking directly to us because we're the guys who actually understand how it works. So when we brought SGX out we decided it was a lot better for us to just do it our own way.
"We feel we compete very well against our direct competitors because we have the history and track record that they don't.
"A lot of our top licensees and most successful chips incorporate ARM based CPUs and we're very happy with that. But our bigger picture is, if you look at these chips from the outside you might think it's got this thumping great processor and then a bit of graphics and other stuff, but that's not actually what's happening. The CPU is becoming a smaller part of the equation as the graphics, comms, memory, etc become bigger in terms of where the silicon real estate goes."