Where does the AI boom go from here?

Anyone who is pessimistic about the ability of the AI boom to sustain earnings may want to look away now. Nvidia, whose GPUs are used for training the large language models that dominate the AI boom, outstripped even Wall Street’s already lofty expectations last week.

The Silicon Valley-headquartered company reported revenues of $13.51bn for the second quarter, up from $6.7bn during the same quarter last year and is forecasting sales of $16bn during this quarter, comfortably ahead of the average $12.61bn predicted by analysts that were polled by Refinitiv.

Nvidia is of course a special case because of the stranglehold it has on the chips used to train these AI models, although it is not the only tech giant to have soared this year as the AI boom has lifted the share prices of the so-called ‘Magnificent Seven’ – Nvidia plus Apple, Alphabet, Microsoft, Amazon, Meta and Tesla.

While there is much debate to be had about whether tech stocks can rally any further, especially after a more difficult period lately, market observers FSA spoke with were unanimous in their views that the recent repricing was more than merited.

“What I think has happened is with the large language models, there’s been a recalibration in the market, not so much about the impact of AI but rather the timing,” said Leon Goldfeld, Asia Pacific head of multi-asset solutions at JP Morgan Asset Management.

“If you thought the cash flow was going to be big 10 years from now, but you now think it’s going to be big five years from now, that does improve your valuations. The question is should they have gone up as much as they have? That’s a much harder discussion. But should they have gone up? I think the answer is unambiguously yes.”

‘Picks and shovels’

Just as during the 1848 California gold rush, many people seeking to make money looked to sell the tools and equipment used by prospectors rather than trying to strike gold themselves, a lot of the interest has been in the technology that will be used to fuel the AI revolution, often referred to as the ‘picks and shovels’ providers.

Yash Patodia, global industry analyst at Wellington Management, takes this concept a step further. He divides the current investment opportunities into three buckets essentially: the picks and shovel providers or enablers, the platforms and the applications.

He noted that the obvious winners were the enablers, particularly companies in the semiconductor and hardware space like Nvidia, while for platforms there was some clarity about who the winners would be, although even here it’s not entirely certain such as whether Meta’s open-source model will be successful, and for applications, it was still unclear.

These views were echoed by Richard Clode, portfolio manager at Janus Henderson and co-manager of the global technology leaders and sustainable future technologies strategies.

“I think we’re pretty clear as to who the infrastructure winners will be. I think we’re pretty clear as to who some of the initial platform winners will be. But I think TBC’d is will all of those software providers be able to provide the incremental productivity gains that mean that people will be willing to pay $30 per enterprise user in two years time or does that become a bit more commoditised and you can’t charge for it?” he said.

Patodia pointed to Microsoft as an example as the company recently announced that it intends to charge about $30 for its AI tool. Based on the fact that the company has around 350 million Office subscribers, that represents potentially a $10bn market opportunity, noted Patodia.

Although, generally, market observers FSA spoke with noted the fact that most of the bets so far in the public markets had been with the enablers, rather than who the winners on the application side will be, which is more of the preserve of venture funding.

“It’s not like the market is buying a lot of unprofitable companies who are developing software that may run using AI and that will disrupt business model A or business model B in different industries. The market’s not really pricing that. And even venture I think has only just started looking for those opportunities,” said Goldfeld.

Is it sustainable?

The extent to which share prices have rallied since OpenAI introduced the ChatGPT large language AI model to the world has been stunning. Nvidia is the most obvious example as its share price has more than trebled this year, although Meta’s has also doubled, while Apple, Alphabet, Microsoft, Amazon and Tesla have rallied too.

This performance is a far cry from last year when many market observers were calling time on the decade long tech rally, saying that its valuations could no longer be sustained in an era of higher interest rates and subsequently higher discount rates. Meta even dropped into the Russell 1000 Value index last year.

Given how heavily tech stocks sold off last year, market observers think that the valuations are currently justifiable, but more importantly noted that the technology behind a lot of these large language is proven, explaining why a lot of the enthusiasm is not ill-founded.

“I think the thing that makes us really excited about AI today is it’s not a new technology. We’ve literally had this conversation with Alphabet six years ago when they said AI was finally going to take off after many winters of discontent. In hindsight, the reason they were so excited is they had just invented the transformer model, which is obviously the technology that sits behind ChatGPT,” said Clode.

“What makes us excited about this is this isn’t like the metaverse where it’s a technology that hasn’t been invented yet. AI is a technology that is ready now. It’s been ready for a few years. It’s taken OpenAI to let the genie out of the bottle. And it can proliferate very quickly. We think this technology will proliferate in the next 12 months, not the next 12 years.”

The question of whether tech stocks can rally further is a trickier issue though, especially as they have fared less well over the last month and most of their earnings growth, Nvidia notwithstanding, has been a result of cost discipline rather than seeing any benefit from the AI revolution.

“Our thinking is that a lot of the rally is probably done. The whole initial repricing has played out over the last four or five months is probably more or less done. This catalyst, of the realisation that it’s developing a lot faster, that I think is more or less priced in the market already,” said Goldfeld.

Patodia has a different view.

“Do we think that stocks can rally further from here? In my view, they can. Stocks to me are a function of earnings power and I think there are few really differentiated companies, which won’t in my mind get commoditised over time and can actually extend their lead, which is where I see opportunity for value creation,” he said.