FF
I think there are a lot of expectations at NVIDIA. Expectations that are unlikely to be shared at the same level of enthusiasm by the ARM staffers.
Let's take a brief look at the history of hardware best running CNN's. Very brief: NVIDIA with its GPU's designed for high end gaming experiences, got lucky when the same architecture proved best-in-class for running large CNN's..
NVIDIA know they got lucky, being able to add the surging volume of CNN-related data centre and cloud business to their core gaming technology. They also know they need to do something dramatic to stay pre-eminent in their new AI marketplace. And they have buckets of money. So who is doing nimble things in very large numbers, away from the data centre?
“Unites NVIDIA’s leadership in artificial intelligence with Arm’s vast computing ecosystem to drive innovation for all customers”
“The combination brings together NVIDIA’s leading AI computing platform with Arm’s vast ecosystem to create the premier computing company for the age of artificial intelligence”
“Uniting NVIDIA’s AI computing capabilities with the vast ecosystem of Arm’s CPU, we can advance computing from the cloud, smartphones, PCs, self-driving cars and robotics, to edge IoT,”
“ ...the acquisition takes Nvidia from top-tier status as a GPU maker for data center, automotive and gaming, and it instantly cements the company as a top-tier chip designer with scale ...Arm and its intellectual property (IP) are already included in billions of chips ...this includes a large volume of smartphone devices that use Arm-based microcontrollers, but also a huge portion of the overall revolution of smartphone devices that have been built leveraging Arm architectures. “
In short, NVIDIA have been able to afford to buy a very attractive provider of IOT technology, who boasts much lower power consumption for its devices, than NVIDIA can.
But ARM does not do neuromorphic. Rather it will try to meld memory and other technologies to ramp down the power consumed at the Edge.
I think that could be called a reasonable transitioning strategy, but NVIDIA's strategy has much more lofty designs than just hanging on.
To wrap, NVIDIA talks of “NVIDIA’s leadership in artificial intelligence”, but it seems it is relying on ARM staff to pull an Edgy rabbit (they don't have) out of a hat.
If you have a one trick pony, but your master insists on new tricks ...and now,..where do you go?
One haven could be to join someone who a former colleague tells you knows how to do lots of relevant and current Edge+ AI tech tricks?
(20min delay)