AI Hardware Summit | The event which defines the industry for the year ahead Sponsored Content by AI Hardware Summit
Nearly 40 companies, including hyperscalers, semiconductor companies, device manufacturers and the critical mass of AI hardware start-ups from across the globe, will present their approaches to processing machine learning training and inference workloads in both server and client computing.
A year further into the nascent AI chip industry, the AI Hardware Summit returns to the Computer History Museum in Mt. View, September 17-18. Last year’s event was the first time that the AI chip industry had come together, and the show witnessed the world premiere of Habana Labs, alongside visionary presentations from hardware innovators working on silicon and systems for processing deep learning, neural networks and computer vision.
This year’s event will feature an opening luminary keynote from John L. Hennessy, Chairman of Alphabet, Inc. and Turing Award laureate, titled ‘The End of Moore's Law, CPUs (as we know them), and the Rise of Domain Specific Architectures’, as well as presentations from investment luminary keynote speaker Lip-Bu Tan, CEO Cadence Design Systems and Chairman of Walden International and a featured keynote from Naveen Rao , CVP & GM, Artificial Intelligence, Products Group, Intel.
The AI Chip Value Chain
Hardware is a gigantic bottleneck for deep learning research – as new hardware is developed, new AI workloads will be enabled, generating further demand for new hardware. The appetite is there, but how this hardware will be consumed is in question.
Will data-driven enterprises buy on-prem servers optimized for AI, or consume AI chips through the cloud? Will there be a specialized chip for every AI application?
Industrial and financial analysts, such as Karl Freund from Moor Insights & Strategy, will examine the formation of the AI chip value chain, and track the development of this fast-paced industry.
Innovation in the AI Chip Start Up Ecosystem
2018 was the year that companies announced AI chip projects, this year the summit focuses on seeing silicon being taped out, and early customer success stories. With new innovations entering the market, it’s likely some start-ups may disappear, and the landscape may look very different by early 2020.
Following last year’s launch with Habana Labs, the AI Hardware Summit guarantees industry-defining announcements from start-ups changing the face of the AI hardware ecosystem.
Training & Inference at Hyperscale
Hyperscalers need to drive training run times down from several weeks to several hours and are investigating solutions for increasing throughput, and lowering latency while overcoming limiting factors of power distribution, density and cooling.
Hyperscaler representatives such as Eric Chung, co-founder of Microsoft Brainwave and Mikhail Smeliyanskiy, manager of Facebook’s AI system co-design group, will discuss their respective companies’ hardware requirements for training and inference at hyperscale. In addition, they will unpack what the software infrastructure needs to look like to make capabilities enabled by new AI hardware available to the masses.
Applications for AI Accelerators in Edge Devices
Approaches to inference for edge computing, IoT and consumer devices are as diverse as the applications that they target. In autonomous vehicles, for example, new accelerators must be extremely performant and accurate, while consuming minimum power and being as durable as ASIL standards dictate. In other use cases, such as consumer devices, key constraints are cost and size.
A number of speakers including CEO of Panasonic’s Arimo, Christopher Nguyen, will examine from a silicon- and system-level perspective the challenges of developing highly specialized edge inferencing systems, how innovative AI hardware can give smart devices a competitive edge, and what device manufacturers are looking for from new AI chips.
This year, the summit can accommodate 500 attendees and there’s fewer than 300 remaining. As headline media partner for the event, we’re delighted to offer Enterprise AI readers a 10% discount to attend. Simply quote ENTAI10 when booking online at www.aihardwaresummit.com
Press contact for AI Hardware Summit:
Head of Marketing, Kisaco Research
+44 (0)203 897 6804