In this series of ai.nl, we feature the most promising AI companies in the world: where do they come from, what have they achieved and what are their plans for the future. In this episode, we are looking at NVIDIA, a Santa Clara, California-based computing platform company driving some of the leading edge innovation in the field of graphics, HPC, and AI.
In April last year, NVIDIA CEO Jensen Huang, wearing his trademark leather jacket, presented a GTC keynote from his kitchen. Having broadcast several announcements from his kitchen, nothing seemed amiss about this event until NVIDIA explained how part of the keynote featured a Huang who was far from real.
The moment showed NVIDIA’s vision to not just be a company building graphics processors powering high performance computers to mining bitcoin. Whether you call it a computer-generated clone, a deepfake or a digital double, NVIDIA made sure every aspect of Huang looked real, including his leather jacket.
A pioneer in graphics computing
NVIDIA’s success is essentially the story of three computing pioneers seeing the need for a new form of computing platform and starting a company to deliver on that need. NVIDIA was founded by Jensen Huang, Chris Malachowsky, and Curtis Priem in 1993, the same year when the term “millenial” was coined.
At the very beginning, Huang and his co-founders were convinced that the PC would become a commercial device and would eventually be used to enjoy multimedia and video games. They saw the PC as a modern device at a time when it existed in the form of a clunky device requiring a dedicated space. Their vision also stretched as far as to imagine a graphics platform that differed from graphics chip companies that were already there. NVIDIA set out to build an advanced specialised chip capable of faster and realistic graphics for video games.
“There was no market in 1993, but we saw a wave coming,” Malachowsky told Forbes in 2016. “There’s a California surfing competition that happens in a five-month window every year. When they see some type of wave phenomenon or storm in Japan, they tell all the surfers to show up in California, because there’s going to be a wave in two days. That’s what it was. We were at the beginning.”
This ability to be at the beginning of the new wave of graphics processors helped NVIDIA become one of the defining computing platform companies of this generation. It may not be as old as Intel but at a market valuation of $455B, it has had a transformational impact on how people use their computers.
The modern day PC is not just designed to get things done. It is designed to be portable, designed to be used as a media consumption device, and importantly, designed to be used as a gaming device. NVIDIA has single handedly pushed PC vendors towards that direction with its advanced graphics processor and it has also pushed cloud vendors to deliver high-performance computing on demand.
NVIDIA: key members
- Jensen Huang (Founder, President and CEO)
- Chris A. Malachowsky (Founder and NVIDIA Fellow)
- Colette Kress (EVP and Chief Financial Officer)
- Jay Puri (EVP, Worldwide Field Operations)
- Debora Shoquist (EVP, Operations)
- Tim Teter (EVP, General Counsel and Secretary)
NVIDIA: timeline of major events since its founding
- April 1993: NVIDIA was founded by Jensen Huang, Chris Malachowsky, and Curtis Priem
- 1994: NVIDIA signs first strategic partnership with SGS-Thompson
- 1995: NVIDIA launches its first product called NV1
- 1996: NVIDIA introduces its first Microsoft DirectX drivers with support for Direct3D
- 1998: NVIDIA signs multi-year strategic partnership with Taiwan Semiconductor Manufacturing Company (TSMC)
- 1999: NVIDIA goes public on January 22
- 1999: NVIDIA introduces its Quadro GPU aimed at professional use
- 2001: NVIDIA becomes the fastest semiconductor company to reach $1B in revenue
- 2002: NVIDIA ships its 100 millionth graphics processor
- 2005: NVIDIA builds processor for Sony PS3 and acquires Taiwan-based ULi Electronics
- 2007: NVIDIA registers first quarter with $1B in revenue
- 2009: NVIDIA introduces Fermi Architecture at the inaugural GPU technology conference
- 2014: NVIDIA debuts Tegra K1 with the aim of delivering best gaming experience on Android
- 2016: NVIDIA begins its AI transition with the launch of AI-powered Pascal, DGX-1, and Drive PX2
- 2018: NVIDIA brings real-time ray tracing to GPUs with its Turing Architecture
- 2020: NVIDIA successfully acquires Mellanox and announces $40B ARM acquisition
- 2021: UK government begins scrutiny of the deal while tech giants express their concerns
- 2022: NVIDIA and SoftBank announce termination of ARM deal
NVIDIA’s desire to be more than a chipmaker
NVIDIA needs no introduction to anyone who has followed the world of computing. The chip giant has single handedly transformed the PC gaming landscape with its GeForce graphics processors and the recent RTX graphics processor lineup. It has also changed the professional graphics industry with its Quadro lineup.
In the past decade alone, NVIDIA has progressed at such a rapid pace, introducing and evolving its architecture so consistently, that it has put rivals Intel and AMD on high alert. For millennials, NVIDIA might seem like another computing platform company that sells graphics processors used in everything from desktop gaming computers, gaming laptops, consoles, crypto mining, cloud computing, to self-driving cars but it aspires to be more than a chipmaker..
The graphics processors made by NVIDIA are essentially a means for the company to fulfil its bigger mission and that is to become a platform player. It not only wants to sell a graphics processor but also a platform that can be built on top of that processor. This theme played out in a big way at the GTC 2022 conference held in March.
The company introduced its Hopper H100 GPU, a GPU packed with a whopping 80 billion transistors and built on TSMC’s 4nm process, to scale AI data centres. It also expanded its AI Enterprise Stack at the event as an effort to drive digital transformation. The AI Enterprise Stack from NVIDIA is a multi-layer model with the bottom layer being represented by systems such as DGX, HGX, EGX, and others built using NVIDIA’s GPUs and DPUs.
The middle layer includes all the necessary software and programs required for developers to work with the hardware, including CUDA, TAO, RAPIDS, Triton Inference Server, TensorFlow, and other software. The top layer is a set of pre-built AI systems that help developers address specific use cases.
With AI Enterprise Stack, NVIDIA is arguing that developers need not go anywhere when building their AI apps. From acceleration to development to deployment, it has got a suite of services for AI developers. If Apple is transforming itself from a hardware company to a services company where hardware serves as the means to experience that service, NVIDIA is going through a similar transformation.
While it has a proven track record of delivering high-performance compute platforms necessary for multimedia, gaming, cloud computing, and accelerating AI applications, it is now looking to offer the tools necessary to build on top of this compute platform. In a nutshell, NVIDIA wants to be an indispensable platform for data scientists, machine learning engineers. The key thing to watch is how NVIDIA manages to do that without eliminating its partners.
What to read next?
- 🤗Hugging Face wants to be GitHub of machine learning: a look at its history, key members, and major achievements
- 🧑🏫 OpenAI wants to be the first to build artificial general intelligence: a look at its history, key members, and major achievements like GPT-3
- 🕶️ Darktrace wants to be the torchbearer for AI-powered cyber defence: a look at its history, key members, and major achievements