Why the tech bubble seems safe - at least for now

For all the trillions being spent in the pursuit of artificial intelligence (AI) it still cannot answer the existential question: is AI a truly transformative technology, akin to the railways, or a bubble about to burst, taking global markets with it? In the absence (for now) of a definitive digitally-generated answer, the performance of chip-maker Nvidia is a good proxy.

The world's largest company by value has released quarterly results that give clues to the future, as well as evidence of current performance. Nvidia's growth is fundamentally linked to AI.

It made its name developing the graphic processing chips (GPUs) that powered gaming and by 2022 was turning over $9bn. A tidy sum, but soon to be eclipsed as CEO Jenson Huang spotted the potential of his GPUs in the nascent AI technology and threw the entire company behind its development.

Money latest: Airline makes plus-sized travellers buy two seats That really was transformational. After ChatGPT provided public proof-of-concept for large language AI models (LLMs) in early 2023, Nvidia took off.

Its valuation trebled last year and has risen 35% already this year on the way to a market value of $4.4trn. That's double the value of every company in the FTSE100 combined and getting on for 25% larger than total UK GDP.

With stratospheric valuations come similarly lofty expectations, however, which helps explain why, despite reporting $46.7bn in second quarter revenue and $26.4bn in profits, both ahead of expectations and up more than 50% year-on-year, markets were initially unimpressed and Nvidia shares dipped 3% in post-close trading. That sentiment may not last, but could be explained by questions over Nvidia's access to Chinese markets, blocked until recently by Donald Trump.

Huang has done a deal with the US president to hand over 15% of its revenues from sales to Beijing, but until it is inked, Nvidia is neither booking or forecasting Chinese revenue. The vibes around AI have also shifted in recent months.

A recent report from the Massachusetts Institute of Technology found scant evidence of meaningful returns on AI investment despite huge spending across the corporate landscape. And after OpenAI's ChatGPT-5 launched to lukewarm reviews, its founder Sam Altman, one of the biggest beneficiaries of the hype, wondered out loud whether the market was overheated.

Despite the noise, Nvidia believes it has plenty of road left to run, and with good reason. Tech giants Google, Microsoft, Meta and Amazon and many others are all chasing the potential of AI, perhaps uncertain what it might ultimately deliver but desperate to be first to find out.

That impetus is driving enormous investment in hardware. More than $41bn of Nvidia's Q2 revenues came from the sale of data centre equipment, and it predicts another $3tn-$4tn will be invested in AI infrastructure by the end of the decade.

If anything on that scale is realised, then Nvidia and Jenson Huang will keep winning. Like the railroad pioneers who transported prospectors west during the California gold rush, it matters less that they strike lucky as long as they keep trying..