The Picks & Shovels of the Generative AI Gold Rush

Look around the city of Melbourne: it is filled with buildings that hark back to the gold rush era of the 1850-80s. Timber buildings in the city were rebuilt in solid stone and brick with artisanal flair, providing displays of abundant wealth. The city was built with gold, and this era became the beachhead for Melbourne, transforming into the industrial capital of Australia. It was one of the richest cities in the world, second only to London.

In boom times, there are often many indirect economic beneficiaries. You could say that today, on a global scale, we've now entered a generative AI gold rush.

As always, those selling the picks and shovels will do better than many selling the new generative AI tools themselves. It's not surprising that, as of yesterday, there were over 7,500 generative AI tools on the market, performing over 2,000 different types of tasks.

When it comes to the generative AI boom, astute investors should play the zone and not just the ball – by playing the zone, it is easier to pick winners, and the opportunities are plentiful.

During times of technology revolution, I like to look at the economic beneficiaries in three clear categories:

• those who provide the actual services and are at the center of things, in this case, generative AI and the infrastructure needed to create it (GPUs and chips).
• those who will gain operational efficiencies via employing the newly available technology; and the most exciting part of all
• the unexpected use cases that emerge. I call these the Unlocks.

In this piece, I'll go through each and where to look for them.

The Tools

The tremendous run of Nvidia in the past 12 months is one of the best examples of picks and shovels we've seen in the modern investing era. In the past 12 months, the stock has risen an incredible 338 per cent and now has a market capitalization of $US1.2 trillion. The chase for chips powering the AI boom seems as though it is unlikely to end anytime soon.

With this statement, we must remember that a vast majority of corporations and small businesses globally have yet to implement any form of generative AI within their organizations, other than some employees tinkering with ChatGPT. It feels much like it did in 1995 for the internet; we are really just getting started.

Another opportunity is semiconductor company Arm Holdings. A failed takeover attempt by Nvidia to purchase it for around $US40 billion in 2020 highlights the company's market importance. Backed by parent SoftBank, it has filed for an IPO, with an expected post-float valuation of $US60-70 billion, marking it one of the most significant in recent times.

Arm's financials indicate a healthy status, with a 16.2 per cent three-year earnings compound annual growth rate (CAGR) and a seriously impressive average gross margin of 95 per cent. The company's unique position comes from its licensing business model, as it designs but doesn't manufacture processors, which are ubiquitous in devices from Apple, Samsung, and Qualcomm. Interestingly, many of Arm's major customers are also anchor investors for the IPO, reportedly including tech giants like Alphabet, Samsung, Apple, Microsoft, and Nvidia themselves. The IPO lands on 13 September.

And while it's easy to think we might be done with the great AI chip run, recent developments in the startup world spell otherwise. Most startups that are receiving funding and experiencing growth are operating in the AI space. However, many of these startups face challenges in accessing the chips they require for their operations. Interestingly, the startup pitching process to venture capitalists has evolved to include discussions about access to 'compute' resources, making it a factor in investment decisions so much so that a new 'Chip Broker' market has also emerged, aimed at helping startups secure the access they need to production resources.

The situation could become more complicated as geopolitical tensions lead to a decoupling between the West and East in critical tech infrastructure and chips. Given these factors, demand for such resources is likely to continue rising.

Newfound Efficiency

The AI upside here is doing what we did yesterday more efficiently, the oldest trick in the technology book: productivity - doing more with less. And the 'less' in this case will quite simply be people. While I don't think AI will result in fewer jobs, it will result in fewer jobs in certain places. The clue is in the name: Generative AI. Whatever people create, write, think, do, and say, is now all up for grabs by AI.

In many ways, it's the same players who were major beneficiaries of the PC and internet revolutions who will benefit from lower costs. Media, creative industries, companies with a heavy administrative burden such as banking, finance, and legal work. My most recent Eureka piece is a classic example of what we should expect. Corporate layoffs can often be a clue to newfound productivity. The risk, of course, is that these newfound efficiencies can open the door to disruption via nimble players with access to low and no-cost generative AI tools themselves.

The AI Unlock

What's not so easy to see during technology revolutions is what I refer to as an Unlock. This is what you might refer to as an unexpected use case that creates entirely new corporate typologies and industry verticals. This is where the biggest investing upside most often resides – going from non-existent to giant. Before we get to what this looks like for AI – it's instructive to review those we've seen during previous tech shifts.

Social media unlocked User-Generated Content. This was something traditional media failed to embrace, the possibility of interactive media. This was unsurprising given it compromised their integrity and expertise as the authority, and so they simply put newspapers on screens. The same was true of linear TV ignoring the possibility of content creators that YouTube leverages. The net result is that legacy media organizations are now rounding errors compared to those who usurped them.

More recently, the launch of the smartphone was telling. A key unlock was the opportunity provided by GPS. It enabled businesses like Uber, Airbnb, dating apps, and social media with geotagging all providing hundreds of billions in newfound market value. These use cases were largely a surprise and unexpected. It's notable that GPS was barely mentioned in the inaugural address of the iPhone given by Steve Jobs.

Likewise, the real hits for AI have yet to arrive, and will likely be something we are yet to see, let alone invest in. The likely candidates in my view are climate and energy-based technology, and medical breakthroughs. Both are industries that can benefit significantly from AI trawling through incredibly large global data sets, and cross-referencing insights and finding solutions which no human or organization could conceivably find.

As investors, all we need to do is pay attention and be nimble to get on these unlocks once they arrive.