After dominating stock market news for several years, coverage of Nvidia (NASDAQ:NVDA) seems to have cooled. I know the company remains one of the key players in the AI trade, but the likes of Micron and SpaceX have overshadowed it.
So I decided to find out how much weight it still holds, and whether it’s still a good AI investment. Naturally, I asked one of the most popular AI chatbots, ChatGPT.
The answer?
It came back with a tentative ‘yes’. It said:
I still think Nvidia is arguably the strongest pure-play AI stock, although I wouldn’t say it’s automatically the best AI investment for every investor.
It went on to note how Nvidia’s size ($5trn+) is unmatched but with competition increasing, expectations are also “enormous“. Microsoft, Amazon, Alphabet and Meta are all developing their own AI chips, while AMD is quickly becoming a strong GPU competitor.
That’s likely to put Nvidia’s gross margin under pressure, particularly when accounting for higher component and memory costs.
Here’s a beakdown of the core companies competing for the AI crown, and their main strengths:
That’s a strong line-up of contenders, many of which could soon offer cheaper — and potentially better — AI infrastructure. But Nvidia still has one trick up its sleeve: CUDA.
The GPU system
That’s Compute Unified Device Architecture, the software bridge that lets AI applications use Nvidia GPUs efficiently. Essentially, it removes the need for AI developers to use a third-party product to optimise GPU usage. That drives revenue away from competitors and directly into Nvidia’s pockets.
This is particularly relevant when it comes to AI, because the models perform vast numbers of repetitive calculations. Sending computations directly to the GPU can make training and running AI models dramatically faster.
Included in the system are bespoke AI libraries designed to automate low-level repetitive operations. That cuts out huge chunks of work for developers, making it the obvious choice for any AI business.
Until a competitor develops a similar system, Nvidia remains in a strong position to charge premium prices.
The bottom line
Nvidia aside, the real question is whether the enormous sums being spent on AI infrastructure are ever going to turn a profit. More than $1trn has reportedly already gone into AI infrastructure since 2024, with another huge wave of spending expected. Anybody who was alive during the dotcom boom might recognise a pattern here.
