Don't confuse the AI rally with the dot-com bust

Enthusiasm around artificial intelligence (AI) has manifested in some truly eye-watering valuations. Nvidia, of course, leads the pack with 206% year to date. 

Not surprisingly, therefore, many are treating the tech sector as if it's cresting the valuation mountain. 

But what if it isn't? 

As Iain Fulton, Portfolio Manager, Yarra Global Share Fund, points out, AI as an industry is still in the formative stages of development. 

If we're only a few years into that build-out, it’s possible we're still early in this rollout phase and so valuations are justifiable when we look at the future growth and return on capital for those businesses.

So, who will the winners and losers be out of the trend? Iain Fulton has some ideas, which you'll find below.

Note: this interview was recorded on September 7, 2023.

Edited Transcript

LW: Mega caps have carried markets, but you’re underweight. Why?

We underestimated the impact firstly of the reallocation of CapEx from some of those mega-cap companies, particularly in digital advertising such as Meta, where they shifted their capital expenditures away from things like the Metaverse into focusing on their core business and reinvigorating their platform for social media, making it more interesting to advertisers and to consumers. 

They also cut costs, including a 20% cut in their workforce, so they've moved ahead of the rest of the economy and now they're seeing a slow-down in pressure coming through on profits. The tech sector moved ahead of those parts of the economy, shedding labour and cutting costs quicker than everyone else and that's without even mentioning AI.

LW: Is there still value to be found in tech?

I think it's interesting to think about tech in terms of the evolution of computing platforms. 

Where are we with AI and what does it mean for us? If you look back to the 1950s and onwards, every 15 years, we've seen a new computing platform emerge. 

IBM captured 85% of the value chain of mainframes, digital equipment and mini computers in the '50s and '60s. Microsoft and Intel captured 85% of the value chain of PCs, and then Nokia for cellphones, followed by Apple taking 85% of the value chain for smartphones.

Today, we think we're at the build-out phase and the hardware infrastructure build for an Internet of Things, of distributed computing, with an AI-enabled data centre at its core. 

There's probably going to be one company that captures 85% of that value chain, and that looks like Nvidia (NYSE: NVDA) today. 

If we're only a few years into that build-out, we're still early in this rollout phase and so valuations are justifiable when we look at the future growth and return on capital for those businesses. 

When we compare the rollout of the internet in the late 1990s that was financed and built out by heavily indebted telecom companies, versus today, the infrastructure rollout of AI is now financed by cash-generative, cash-rich businesses with zero debt and very large CapEx budgets. 

So you have $180 billion of CapEx between those top seven companies, plus private capital, that will come in to advance that hardware infrastructure rollout. So it's early days, but it could potentially make some of these stocks quite good value.

LW: How will we identify the winners and losers from AI?

The standout trend has been the advance of those stocks in 2023. We must remember how far they fell in 2022, so it's arguably been a reversal from an oversold position. But I think now, with the visibility on hardware infrastructure rollout to enable an AI-powered computing platform, we have visibility on those hardware companies and the success within that value chain. 

I think where we have less visibility is what the killer applications will be. Back to 1999, we had the merger of AOL-Time Warner and they were the largest internet access company and one of the largest media companies at the time, and their killer application was going to be buying goods live from TV shows in the comfort of your own living room.

Now, we buy goods online and we watch a lot of TV, but we don't necessarily watch it in our living room, and we buy things from our phone. We have the sum of human knowledge in our pockets. 

The applications of that technology at the infrastructure rollout phase are unknown, and they do impact and penetrate more areas and more sectors of the economy relative to the previous cycle. We think exactly the same will happen in AI. 

It's too early to tell who the winners from the applications will be, but we know that the infrastructure is going to be built out and there's good visibility in the revenue cashflow and profitability in that effort.

LW: The Netflix turnaround has been impressive. Is it too good to be true?

No, I don't think Netflix (NASDAQ: NFLX) is too good to be true. There is a transition in the business model away from subscription only to a mixture of subscription and advertising revenue, and we think that expands the available revenue opportunity for Netflix. 

So traditional TV networks, for example, were down $150 per annum per viewer in advertising revenue. If Netflix can earn $150 per annum per viewer in advertising revenue, plus $60 in annual subscription for their basic ad-supported tier of revenue, then you have $210 per viewer versus about 120 or $140 today. So if they're successful in that transition, then it's very much an expanding revenue opportunity with not much additional cost, which means better cash flow, better return on capital, and all that, a pretty compelling valuation.

Managed Fund

Yarra Global Share Fund

Global Shares

View