Buy the future with Nvidia

Nvidia (NVDA) specialises in graphics processing units (GPU), which are computer chips that were originally designed to render three-dimensional visual effects onto a two-dimensional digital surface. It has pioneered the use of GPUs in the video games industry, enabling players to immerse themselves in a myriad of virtual landscapes. But the demand for ever-more realistic and responsive images has led to the development of increasingly powerful processors with applications far beyond the group’s gaming roots. With a rapidly expanding footprint in data centres and artificial intelligence, Nvidia offers investors a front row seat to the future of technology.

That’s not to say that gaming won’t still be significant to Nvidia’s earnings picture – it accounts for around half of the group’s total revenue. Nvidia is the number one player in GPUs for PCs and serves more than 200m gamers. It has benefited from the Covid-19-driven surge in gaming this year, with revenue from gaming jumping by close to two-fifths year on year in the three months to 25 October, to a record $2.3bn (£1.7bn).

This growth was thanks to demand for chips for Nintendo’s (JP:7974) hugely popular Switch console as well as Nvidia’s latest GPUs for laptops and PCs. The group recently launched a new generation of graphics cards – the GeForce RTX 30 series – that offers a superior performance at an attractive price point. This could spur a lucrative upgrade cycle among gamers.

The limiting factor here is supply rather than demand as pandemic disruption has created industry-wide capacity constraints. Chief financial officer Collette Kress says that “it may take a few more months for product availability to catch up with demand”.

Powering the future

Nvidia’s data centre business looks like it has even greater growth potential. These chips deliver the computing power required to process large volumes of information at pace. Nvidia's A100 data centre GPU is already used by major server vendors and cloud service providers, including the three big hitters – Amazon Web Services, Microsoft Azure and Alibaba Cloud.

Data centre sales are growing much more quickly than gaming, with revenue almost tripling on the year to $1.9bn in the third quarter. This was boosted by the $7bn acquisition of Mellanox Technologies in April, which is a specialist in data centre networking equipment.

Mellanox did see a large order from Huawei pulled forward, meaning sales are likely to fall in the fourth quarter. That spooked investors following the release of the group’s third-quarter numbers and the shares dipped despite Nvidia beating both its own guidance and analysts’ expectations. But contrasting the market’s more lukewarm reaction, brokers have once again upgraded their forecasts.  

The thinking computer's choice

As we shift ever closer to the ‘Internet of Things’ (IoT), this will require the use of artificial intelligence (AI) to make rapid decisions without human input and Nvidia’s GPUs can be used to train neural networks for machine learning.

While it may sound outlandish right now, the group is paving the way for the age of AI-based computing – something that will have broad applications across industries, from self-driving cars to smart factories.

In healthcare, Nvidia has already developed a platform to apply AI to medical imaging and genome analysis, which will be at the forefront of developing treatments for cancer. It has entered into a partnership with GlaxoSmithKline (GSK) that will see its computing power use AI to help develop new drugs and vaccines.

AI underpins many of the ‘megatrends’ we are likely to see unfold over the next few decades and that potential is a big reason why Nvidia is a popular fund holding. “Its graphics processing units (GPUs) have evolved into a computerised brain, straddling the exciting intersection of virtual reality, high-performance computing and artificial intelligence,” says Stewart Heggie, investment specialist at Scottish Mortgage Investment Trust (SMT), where Nvidia is a top 15 holding. “GPUs are the single most important items in developing AI applications and they are in demand across the globe.”

Ben Rogoff, lead investor of Polar Capital Technology Trust (PCT), is similarly enthusiastic, telling Investors Chronicle in August that Nvidia is “one of the core ways we play the artificial intelligence theme”. Mr Rogoff points to the group’s high barriers to entry, namely its experience in GPUs and the creation of proprietary programming software.  

Such product innovation has helped underpin robust margins, which are higher than those of chipmaking peers Intel (US:INTC) and AMD (US:AMD) (see table). Nvidia’s adjusted operating profit margin came in at 40.4 per cent in the first nine months of this financial year.

It is also spending heavily to safeguard its competitive position. It invested $2.8bn into research and development over the same time period – equivalent to just under a quarter of revenue – and also keeps its operating expenses relatively low by outsourcing manufacturing to the likes of Taiwan Semiconductor Manufacturing (US:TSM). Margins should expand as higher-value data centre products form a larger proportion of total revenue. Nvidia’s high margins have translated to a robust return on equity (RoE), which is a key measure of a company’s quality and profitability. According to FactSet, Nvidia’s average RoE over the past five years was 34 per cent.

Armed for future growth

The group isn’t short of growth drivers, but momentum could be turbocharged if the $40bn purchase of chipmaker Arm from Japanese conglomerate SoftBank (JP:9984) comes to fruition. Nvidia struck an agreement for a mixture of cash and shares in September and this would be its biggest ever acquisition. Arm designs chips and licenses out the intellectual property (IP), providing a steady stream of high-margin royalty revenue. Its technology is used by 70 per cent of the world’s population – in everything from smartphones to cars – and its customer roster includes the likes of Amazon (US:AMZN), which has now launched its own Arm-based server processors.

Nvidia would be able to leverage Arm’s client list to cross-sell its graphics processors and also use its own technology to expand Arm’s IP portfolio. The group says the takeover would boost its adjusted gross margin and EPS from the get-go, but this purchase is more about the long-term picture.

Nvidia is looking to create “the premier computing company for the age of AI”, combining its GPU and Arm’s central processor unit (CPU) capabilities. It estimates that the combined business would have an addressable market worth $250bn by 2023.

While the deal is guided to complete in 18 months, it will likely face tough antitrust scrutiny as Nvidia would gain control of the designs used by many of its competitors. Regulators in China could prove particularly difficult to please. Arm is currently seen as a neutral blueprint provider, but American ownership could leave it vulnerable to US government intervention to limit Chinese access to chip supplies. There is also an additional obstacle amid a standoff with the head of Arm’s Chinese business, Allen Wu. He remains in situ despite the board voting for his removal earlier this year and could complicate efforts to secure regulatory approval in China.

Meanwhile, as the UK considers granting ministers greater powers to intervene in foreign takeovers, the Nvidia-Arm tie-up could prove a test case. It remains to be seen whether Nvidia can escape with similar pledges made by SoftBank when it bought Arm back in 2016. Chief executive Jensen Huang has gone on a charm offensive, writing in the Cambridge Independent that Arm will remain based in Cambridge, its engineers will be retained and that its intellectual property will be registered in the UK.

If the Arm deal were blocked, it would be a disappointment for Nvidia, but it wouldn’t halt the growth story. The balance sheet isn’t likely to become overly strained if the takeover does go through, either. Nvidia generated $4.3bn of free cash flow last year and JPMorgan predicts it will produce a further $11.6bn across its current financial year and the next one. That means the $12bn cash portion of the Arm deal should be replenished fairly quickly. Nvidia has a long-standing track record of net cash, although this did dip from $8.3bn at the start of the year to $2.5bn at the end of October due to the Mellanox acquisition.

Just getting started?

Nvidia’s shares have demonstrated some serious momentum in 2020, more than doubling in value so far this year. Reaching an all-time high earlier this month, they are currently trading at trading at an eye-watering 43 times consensus 2022 earnings, with an enterprise value-to-operating profits ratio of 40. That begs the question of whether the group’s potential is now fully priced in. We’re inclined to think not.

Nvidia still has a long runway for growth as we are only at the beginning of the journey towards the IoT and widespread deployment of AI. With or without the addition of Arm, the group is set to play a pivotal role in the future of technology, and indeed how the world operates. While some investors may baulk at the lofty valuation, it may well end up looking cheap in retrospect. Mr Huang recently said that Nvidia is “firing on all cylinders” and we think it’s got plenty of fuel left in the tank.