In mid September Nvidia (NVDA) announced that it was acquiring chip designer Arm from SoftBank for $40 billion, SoftBank paid $31 billion in 2016 (well done SoftBank). I think this deal is worth paying attention to – that’s not a unique view. The semiconductor sector is changing and those changes will likely impact other sectors. So what does this transaction mean? Only time will tell. For now, in my view, better understanding these companies and this sector is going to be generally useful. We’ll start with some basics today.
Nvidia was founded in 1993 and is the premier GPU (graphics processing unit) maker. What is a GPU and how is it different from a CPU (Central Processing Unit)?
The giant of the CPU market that you have likely heard of is Intel Corp (INTC), founded in 1968 by Robert Noyce and Gordon Moore (Moore’s Law).
Some basic differences between CPUs and GPUs:
| CPU | GPU |
| Central Processing Unit | Graphics Processing Unit |
| Several cores | Many cores |
| Low latency | High throughput |
| Good for serial processing | Good for parallel processing |
| Can do a handful of operations at once | Can do thousands of operations at once |
Applications for GPUs have traditionally been in gaming and that remains a significant part of the business. However, the use environments for GPUs have grown considerably, two areas of note are Machine Learning (ML) and Artificial Intelligence (AI). ML and AI are being deployed across sectors, hence, a deeper understanding of their use cases is important.
Further down the field form GPUs are FPGAs, the recently reported potential AMD (AMD) – Xilinx (XLNX) deal is in part a demonstration of their relevance (more on this in another blog). What are FPGAs?
“Field Programmable Gate Arrays (FPGAs) are semiconductor devices that are based around a matrix of configurable logic blocks (CLBs) connected via programmable interconnects. FPGAs can be reprogrammed to desired application or functionality requirements after manufacturing. This feature distinguishes FPGAs from Application Specific Integrated Circuits (ASICs), which are custom manufactured for specific design tasks. Although one-time programmable (OTP) FPGAs are available, the dominant types are SRAM based which can be reprogrammed as the design evolves.” (source: Xilinx.com)
Now back to the Nvidia (NVDA) – Arm deal, and it’s not a done deal, to be completed it will need regulatory approval across multiple jurisdictions. As of now that approval seems to be more likely than not, however, not certain. We’ll continue to pay attention to this transaction and generally to what’s happening in the semiconductor sector. These companies enable many other companies across industries to grow and respond to ever changing customer needs – considering autonomous driving as an example. Thank you for reading and have a great day.
Please join us this Thursday October 15th at 6:30 pm EDT for the GWII Show – Live with Rodney Lake.
All the best,
Rodney
October 14, 2020