It’s the stock market success story of the century – and one that will shape the rest of our lives.
Since going public on the eve of the millennium, computer chip designer Nvidia has come from nowhere to become the largest company in the world.
The remarkable milestone was passed yesterday as the share price soared to a new high, overtaking tech giant Microsoft and valuing the Silicon Valley pioneer at $3.34 trillion (£2.6 trillion).
To put that into context, Nvidia’s stock market valuation now exceeds the annual output of the entire UK economy.
It’s been a truly stratospheric rise.
In February, Nvidia became the fastest company ever to go from $1 trillion to $2 trillion. Amazingly, it only lasted eight months.
Jensen Huang, 61, the Taiwan-born electrical engineering graduate who founded Nvidia along with microchip designers Chris Malachowsky and Curtis Priem
Nvidia originally made computer chips for video game software, but has since expanded its horizons to dominate the AI sector
The latest surge means the company – whose chips have accelerated the meteoric rise of artificial intelligence (AI) – has seen its share price rise from less than $4 in June 2019 to almost $136 today. That’s an incredible increase of 3,477 percent in five years.
How could a start-up without a business plan, founded in a restaurant in California, win everything? And can UK private investors get a piece of the action?
Nvidia is the brainchild of Jensen Huang – a Taiwan-born electrical engineering graduate whose parents sent him to the US as a child – and two microchip designers, Chris Malachowsky and Curtis Priem.
They founded Nvidia in 1993 in a Denny’s restaurant in San Jose, in the heart of Silicon Valley.
The plan was to call their company NVision – until they discovered that name came from a toilet paper manufacturer. Huang, who once worked as a waiter and dishwasher at a Denny’s outlet for $2.65 an hour, suggested Nvidia instead, based on the Latin word “invidia” meaning “envy.”
Since then, he has run the company and become one of the richest people in the world.
Nvidia’s main product is a graphics processing unit (GPU) – a wafer-thin circuit board with a powerful microchip at its core. These processors allow lightweight, energy-efficient personal computers and laptops to perform a large number of calculations at high speed.
For decades, Intel was the largest manufacturer of microchips, but Nvidia differs from its rival in several key ways.
Intel and others make industry-standard general-purpose chips, known as “central processing units” (CPUs), that handle all the main functions of a computer and produce one mathematical calculation at a time.
But Nvidia’s GPU can complete complex and repetitive tasks much faster, breaking them down into smaller components before processing them in parallel.
If CPUs are delivery vans delivering one package at a time, Nvidia’s GPUs are more like a fleet of high-speed motorcycles spreading through a city. That made them the perfect processors to power the dawning AI revolution.
Unlike Intel, Nvidia does not make its own chips – they are mainly outsourced to the Taiwan Semiconductor Manufacturing Company. Crucially, however, Nvidia not only designs the hardware, but also the software they run on.
“What Nvidia does for a living is not building the chip – we build the entire supercomputer, from the chip to the system, to the interconnects, the NVLinks, the networks, but very importantly the software,” says Huang, 61 .
This secret sauce software package is called Cuda. Nvidia’s chips were originally intended to improve computer graphics used by video gamers. The creation of Cuda in 2006 also allowed other universal applications to run on Nvidia chips.
Initially, AI was not one of them. In the early 2010s, AI was still a technological backwater where progress in areas such as speech and image recognition was slow.
Even less fashionable were ‘neural networks’: computer structures that mimic the functioning of the human brain.
Nvidia’s breakthrough came when the Cuda platform was championed by British-Canadian computer scientist and cognitive psychologist Geoffrey Hinton, dubbed the “godfather of AI.”
Two of his students trained a neural network to identify videos of cats using just two Nvidia boards. Google experts needed 16,000 CPUs to perform this feat. Machine learning had arrived.
The stunning results prompted Huang to go all-in on AI in 2013.
Nvidia’s GPUs were soon found in everything from smart cars to robotics and data centers, with customers ranging from Tesla to Microsoft and Amazon. One setback was a failed £31 billion bid to buy Cambridge-based chip designer Arm Holdings.
But Nvidia really came of age a year ago with the news that ChatGPT – Open AI’s chatbot – was powered by its supercomputers. This sent the stock into orbit.
The news sparked a frenzy among major tech companies and AI startups for Nvidia’s processors, leading to shortages that could last into next year.
On May 23, stellar results that exceeded market expectations increased the company’s value by $200 billion in just one day.
To meet the insatiable demand, Nvidia plans to launch a new generation of AI chips — codenamed Blackwell — later this year, each costing more than $30,000.
It can charge so much because of its stranglehold on the AI chip market, with a market share of more than 80 percent.
This near-monopoly has turned Nvidia into a huge money machine, potentially fueling massive future share price gains.
Some links in this article may be affiliate links. If you click on it, we may earn a small commission. That helps us fund This Is Money and keep it free to use. We do not write articles to promote products. We do not allow a commercial relationship to compromise our editorial independence.