Despite mounting fears about an ‘AI bubble’, Nvidia (NASDAQ:NVDA) is on course to deliver a market-beating stock price return this year. As I write (29 December), the share is up 40% versus a 17.8% rise for the S&P 500.
However, the pound has strengthened against the US dollar in the past year, meaning UK investors would have seen some currency drag.
Nevertheless, barring a spectacular end-of-year meltdown, Nvidia will produce a positive return for the sixth time in seven years. But the question now is, what might 2026 bring?
Latest broker target
Personally, I think making exact share price predictions is a bit pointless. There are just too many variables to ever make this an exact science.
But experts are paid to come up with them, so what are they forecasting? Well, the average 12-month price target from 56 analysts is $261. This is roughly 37% above Nvidia’s current share price.
In other words, £5,000 invested today could become £6,850 by next Christmas, assuming this forecast proves correct (which it might not). This also assumes there are no currency movements, which (again) is impossible to predict.
Sustainability concerns
Unless one has been living under a rock, it has been hard to escape debates about an AI bubble. For those who think there is one, the concerns seem to centre around three main things.
First, the hyperscalers (Microsoft, Google, Meta, and Amazon) are set to spend over $400bn on AI infrastructure in 2025. However, direct AI revenue is currently estimated at only a fraction of that. So critics argue these tech giants are overspending/overbuilding capacity.
Linked to this are concerns about how the buildout is being funded. Traditionally, Big Tech companies have funded capital expenditures through their own copious cash flows. However, the likes of Meta and Oracle are issuing debt to pay for AI data centres.
Finally, there’s the issue of circular financing. Microsoft and Amazon are investing billions into AI startups like OpenAI and Anthropic. Those companies then use the cash to buy cloud compute power from Microsoft or Amazon, who in turn buy more chips from Nvidia.
Stepping back then, I can appreciate why some investors are worried.
Valuation
Historically, every new, groundbreaking technology has usually produced a speculative bubble (the internet, for example). AI is unlikely to be any different, in my opinion.
Therefore, I’m avoiding loss-making OpenAI if it goes public in the second half of 2026 at a $1trn valuation (as reported by Reuters). And AI cloud computing firm CoreWeave, as well as highly-valued SMR nuclear startups like Oklo.
But at its current valuation of 24.9 times forward earnings, I don’t see Nvidia stock as ridiculously overvalued. This is because the firm enters 2026 with an unprecedented $500bn order backlog for its Blackwell and upcoming Rubin systems.
Even if new demand totally disappeared tomorrow (which is unlikely), Nvidia will be busy for a while clearing the orders already on its books.
Then the next-generation Feynman architecture is coming in 2028. These systems will be significantly more energy efficient. For hyperscalers, I think that will make these systems very attractive.
That said, I see AI bubble fears weighing on Nvidia’s share price in 2026. I don’t envision another 40% rise.
Therefore, I’m personally looking at other opportunities, including AI application companies where the technology can make them more profitable.
