The cogs of NVIDIA's AI powerhouse machine #3
NVIDIA's role in AI hardware, education, and startups
Issue #3 of the
Infinite Waves
Newsletter
It is easy to think of Google, Amazon, Microsoft and IBM as being the major players in Artificial Intelligence. They get a lot of the news and are firmly cemented in the public consciousness. But NVIDIA, probably best known by most for its role in reshaping the gaming industry with its Graphical Processor Units (GPUs), has been transitioning over the last few years and it seems clear they see their future based around being a products and services company with a strong focus on Artificial Intelligence.
A brief history 👾
In 1999 NVIDIA released the world’s first GPU, the GeForce 256. The company defined it as:
“A single-chip processor with integrated transform, lighting, triangle setup/clipping, and rendering engines that is capable of processing a minimum of 10 million polygons per second.”
In the same year NVIDIA announced its initial public offering at $12 per share.
Over the next two decades NVIDIA continued to develop their GPUs and they found uses in gaming consoles, PCs, laptops, tablets, car navigation and entertainment systems, and supercomputers.
In 2015 NVIDIA began to channel some of its attention towards AI with a focus on Deep Learning (a method of Machine Learning, a subset of AI). They released the GeForce GTX TITAN X, ‘the most powerful processor ever built for training deep neural networks.”
AI focused 👀
NVIDIA has continued its investment in AI specific chips up to the present day, also branching out into the world of virtual reality, and setting benchmarks as some of the most powerful AI hardware technologies for robotics, drones , and autonomous vehicles.
Will Ramey, Global Head of Developer Programs at NVIDIA, explains in episode 113 of NVIDIA’s AI podcast that NVIDIA started seeing scientific papers for Deep Learning around 2015 in which their GPUs were being used for the Machine Learning technique. They began to offer training courses in Deep Learning and their participation numbers quickly eclipsed their other proprietary training courses. This was the point at which they doubled down and subsequently formed the Deep Learning Institute(DLi). The DLI since trained over 200,000 people worldwide, and counting.
NVIDIA Deep Learning Institute 💻
https://www.nvidia.com/en-us/training/
The NVIDIA Deep Learning Institute offers courses on a whole range of AI related topics aimed at a broad spectrum of abilities and backgrounds, from students to scientists. They provide introductions to industry-standard tools and frameworks as well as fully-configured GPU-accelerated cloud servers to complete hands-on exercises included in the training. 🏋️♂️
It all comes in the form of self-paced online learning. A smart move, not only empowering professionals with hardware, but providing an ecosystem to train new potential customers for their products. This is like Ferrari providing you with a car and then getting their lead Formula 1 driver to train you on how to get the most out of it when hitting those winding roads at (probably) the legal speed limit, of course. 🏁 🚗
NVIDIA Inception 🏤
No, NVIDIA isn’t planning on going on some multi-level dream adventures with Tom Hardy.
NVIDIA Inception is an AI and Data Science incubator for startups.
As Jeff Herbst, Vice President of Business Development at NVIDIA, stated:
“Inception is basically our Virtual acceleration platform for AI startups… our goal is to help you develop AI applications.”
(Episode 128 of ‘The AI Podcast’ by NVIDIA)
They have over 7,000 startups enrolled in a wide range of fields, from healthcare to Fintech.
Not a new kid on the startup block
For most of its life, even as early as 2001 (a company called Keyhole that ended up becoming a little thing called Google Earth), NVIDIA has been working with startups. NVIDIA Inception is a natural progression combining their knowledge of working with startups, data-driven insights, and not to mention that they’re packing hardware to make even the most hardcore DIY PC-builder nerd feel extrememly inadequate.
The inception of Inception
“NVIDIA has created a whole new programming paradigm, a programming model, called accelerated computing, or GPU computing.
“We basically have to go out and teach the world how to do this. And because it’s a new technology, startups are leading the way.”
Jeff Herbst.
Around 2008 when NVIDIA was building out their developer platform and working with developers and researchers they noticed that the teams would start to attract funding and startups would form.
At a precursor event to what is now their GPU Technology conference NVIDIA staged what’s called an emerging company summit. It was an experiment. They invited a whole bunch of startups, analysts, customers, reporters, and venture capitalists (VCs).
It was a success. It seemed analysts, customers, and VCs really cared about what startups were doing, according to Herbst.
NVIDIA continued to train developers and startups on their technologies, as well as helping with public relations. Big companies and VCs started to take notice.
This is how NVIDIA Inception was born, from this converging set of circumstances and ideas.
Educate, Nurture and Support
NVIDIA Inception’s mission is to “educate, nurture and support the global ecosystem of AI startups and the venture capital that funds them.” (Jeff Herbst)
Fundamentally they provide guidance, resources, and (occasionally) funding.
The startups also benefit from training and receive free credits for cloud hosting providers such as Amazon Web Services and Microsoft Azure. NVIDIA also offers discounts on GPUs, boosts in marketing, and introductions to customers and VCs.
AI changing the game like NVIDIA changed gaming
“I think think every piece of software, meaningful piece of software in the world, is going to be re-written over the next decade to take into account AI-driven programming techniques.”
Jeff Herbst.
Whilst there has been vast progress in all areas of AI and a high level of maturity in the hardware, research, and products, this is still only the beginning.
And it seems NVIDIA will be right there in all capacities helping at the intersection of these technolgies and techniques. Combining hardware and software to achieve unparalleled productivity and innovation in Artifical Intelligence.
NVDIA for Indie Developers
For Indie Developers the NVIDIA ecosystem then could be a great place to explore and study Deep Learning, with the aim of creating the next wave of solutions and startups in AI.
Sources:
‘The AI Podcast’ by NVIDIA
Ep. 1: Deep Learning 101 - Will Ramey, NVIDIA Senior Manager for GPU Computing
Demystifying AI with NVIDIA’s Will Ramey - Ep. 113
Exploring the AI Startup Ecosystem with NVIDIA Inception’s Jeff Herbst - Ep. 128
Sites