Our starting point is wanting to help our clients make better strategic and operational decisions.
Feb 5, 2018 |
Artificial Intelligence (AI) is set to reshape the world in a similar way that the Internet did two decades ago and this will have a major impact on the semiconductor industry. AI’s influence coupled with the rapid rise of cloud computing and the slowing of Moore’s Law is leading to a fragmentation of computer architecture. CPU designs are no longer keeping pace with improvements predicted by Moore’s Law (i.e. double CPU performance for half the price every 18-24 months). CPUs are serial devices and many of today’s AI applications can be more efficiently executed on parallel devices such as GPUs. As a result, GPUs and a range of AI-specific processors have the potential to displace the traditional CPU in the next few years with profound implications for industry stalwarts such as Intel. This PowerPoint presentation compares and contrasts alternative processor architectures being developed for both the server market and edge devices (automotive, mobile and IoT) and profiles key players.
Table of Content:
- Key Trends Driving the AI Chip Market
- AI Training and Inference
- Key Players Profiled
- Promising Start-ups
- Applications and Approaches - Software vs Hardware
- Conclusions - For the Cloud & for the Edge