Conventional wisdom is a byproduct of groupthink that presents solutions good enough for the average person while simultaneously not being right for any individual. You follow it at your peril. The more different you are from the person that defined a rule the less you should follow the rule. Each Monday I will challenge the investing norms that just may be holding you back from living the life you want.

The not so inevitable investment case for AI 

“There must have been moments even that afternoon when Daisy tumbled short of his dreams -- not through her own fault, but because of the colossal vitality of his illusion. It had gone beyond her, beyond everything. He had thrown himself into it with a creative passion, adding to it all the time, decking it out with every bright feather that drifted his way.”

- F Scott Fitzgerald

In 2009 an experiment was conducted by Rob Walker and Joshua Glenn called the Significant Objects Project. The premise was simple. 100 near worthless items were purchased at thrift stores and garage sales for a total of $128.74. A writer created a compelling narrative about each item and it was put up for sale on EBAY using a picture of the item and the story. The 100 items were sold for $3,612.51. The lesson is simple. A good story goes a long way.

Narratives are powerful and we see them throughout the investing world. Narratives about the overall market and individual companies shape investor behaviour. Narratives are used by the investment industry to sell funds and ETFs.

We create narratives to justify our own poor behaviour even when we know we aren’t doing the right thing. How many investors have justified selling in a falling by market by reassuring themselves that they will ‘get back in at the bottom when the market seems safer’. Our narratives about the past reinforce our sense of self-worth. They ascribe great skill and foresight when things have gone well and unavoidable bad luck when things have gone poorly.

Here is a narrative on an investment opportunity. Does this seem like an attractive investment opportunity?

There is an emerging technology that is early in the adoption cycle. This new technology will reshape the world and within 20 years the 150 million current users of the technology will grow to 4.5 billion users. These people will use the technology daily and it will play such a large role in their lives that a movement will start to limit their use of the technology. It won’t work. The technology will make them more productive, change the way they buy and sell goods and services and alter the way they interact with the world and their fellow humans.

There are several ways you can invest in this technology. You can invest in the build out of the infrastructure that supports the adoption of the technology. You can invest in the companies that provide the technology to users. Or you can invest in the companies that will sell goods and services using the technology. More on my narrative later.

The preordained AI roadmap

The investing world has been focused on DeepSeek. I woke up as surprised as most investors that DeepSeek has created an AI model that is open source, just as accurate as other proprietary AI models, about as fast as other AI models and uses 95 percent less chips than other AI models. Will this all turn out to be true? There is a lot of skepticism.

Part of this focus on DeepSeek is the fact that the last few years the market has been driven by the AI narrative. To be successful a narrative doesn’t have to be coherent. It can link various strands together in a dubious fashion and still drive behaviour.

Conventional wisdom suggests that AI will transform the way we do everything while being immensely profitable for a disparate set of companies all along the AI value chain. In a simplistic value chain there are suppliers of raw materials, companies that create goods and services out of the raw materials and the consumers of the goods and services.

The narrative around AI suggests great riches for the suppliers of chips and data centres that are needed to support the AI models. Nvidia NVDA is the poster child for the suppliers, but local data centre provider Goodman Group GMG is another example. The prevailing narrative is that there will be no disruption to this inevitable outcome. Nvidia will prosper, data centre providers will prosper, and power suppliers will prosper. Innovation won’t disrupt this pre-ordained course.

The narrative is less clear when we turn to the creators of the AI models and the consumers of AI. Yet the market insists this too will be great for investors. We can take the case of Alphabet GOOGL and Meta META as an example. They are spending billions on AI. For example, Meta just pledged to spend $65 billion US on AI in 2025. Yet it is unclear exactly how they will monetise their AI models. The market seems unconcerned. Alphabet and Meta will prosper along with Amazon, Apple and everyone else that could possibly use AI. Even Telstra is touting their AI credentials - but to be fair it isn’t working for them.

The genius of Alphabet and Meta’s business models are that they offer great free products to attract huge audiences which allows them to sell advertising. Will AI expand their share of advertising? Unlikely as their market share is already so high they are facing anti-trust concerns. Can they charge more for better targeted advertising? Seems doubtful if everyone has AI targeted advertising and overall marketing spending doesn’t increase.

Could this spending be defensive and are Alphabet and Meta just in an AI arms race to maintain the status quo? Maybe. Meta and Alphabet dominated advertising because they could target ads better than anyone else. AI threatens this moat because it might be able to do search and social better and target advertising better. Winning this arms race might be necessary to maintain their moats. But investors might have been better off if AI didn’t exist.

The prevailing narrative gives no credence to the inevitable losers from AI. Displaced workers and the populist politicians will get a say. Regulators who are already worried about the power of big-Tech will get a say. The creators of the content and intellectual property that AI steals will get a say. Can any of them derail AI? Probably not completely but they can disrupt and influence the pre-ordained path that many investors see for AI. Nothing in life is inevitable.

All these details just get in the way of the compelling narrative. Every new technology is an opportunity and a threat to some existing business model. It often takes a while to figure out the implications. A lot of very smart people are trying to figure out the investment implications of AI. Many of them are financially incentivised to further the hype as long as possible. When the hype machine grinds to a halt it is often everyday investors who have bought into the narrative hook line and sinker that are left holding the bag.

The exquisite corpse will drink the new wine

Deep in the post-World War 1 malaise a group of French surrealists were sitting around Montparnasse. They were playing a game that involved writing down a single word on a piece of paper and passing it to the next person to add another word. According to legend the first time the game was played the collaboration resulted in the phrase The exquisite corpse will drink the new wine. Not too shabby.

I think about this story when I read the marketing for investment products with exposure to AI. A spin through the wepages of AI ETFs tells me it is a ‘transformational technological megatrend.’ Apparently ‘we are at the dawn of an intelligence revolution, more profound than all past industrial revolutions.’ AI has the ‘potential to change how a huge portion of the economy operates.’

Like the line from the surrealists the AI marketing sounds good but doesn’t mean anything. But there is a difference between art designed to please the aesthetic and marketing designed to influence behaviour. In the former the lack of meaning can add to the allure. In the later our bubble detectors should be triggered.

Morningstar recently conducted research on thematic funds and ETFs. A thematic fund or ETF is an investment product based on a compelling narrative that many investors believe is highly probable to eventuate and - in a fit of linear thinking - will lead to high returns. AI is a perfect example of an all-encompassing thematic designed to titillate investors.

There were 2,776 thematic funds globally at the end of June 2024. Chances are most of them won’t be around for the long-term. In the 12 months prior to June 2024 18% of thematic funds survived and outperformed their Morningstar Global Target Market Exposure Index. If we go back 15 years, the success rate drops to 9%. Over 60% of the total thematic funds available to investors 15 years ago have since closed.

Investing in a thematic may seem smart. Yet it is hubris personified to think great wealth is so easily obtained. Just like the Significant Objects Project what is being sold is nothing more than a story. In the real world we call that a gimmick.

Final thoughts

The emerging technology I described earlier in the article was the internet. I described the prevailing narrative being sold to investors in the late 1990s. Everything I described was true. It all happened. The internet did transform the world. In 2020 there were 4.5 billion people that used it daily.

Infrastructure was needed to support the expansion of the internet. Yet internet infrastructure investments performed terribly. Global Crossing and Worldcom who laid the fiber to connect the world went out of business. Cisco which provided the networking equipment traded at such an inflated valuation the share price is still below the high from 25 years ago. Microsoft took close to 15 years to reach a new high after the bubble burst.

The companies that provided access to the internet through search have all been crushed by Google which didn’t go public until 2004. Almost every early internet retailer went out of business and Amazon emerged from the wreckage.

The transformative vision of the technology came to fruition. The internet was all it was cracked up to be. The investment returns weren’t. Optimistic investors crafted a narrative ot justify high valuations, Investors jumped in and got burned.

Whatever happens with AI I think the DeepSeek episode put some doubt in the prevailing narrative. The ferocity of the sell-off before facts were known may indicate that some people on the AI bandwagon know how precarious a ‘too good to be true’ narrative can be. They know they are buying a story but don’t care because they are making money. They also know it can’t last forever.

Just remember that every narrative needs a universal truth that is used to justify all sorts of fantastical outcomes. The snake oil salesmen promoting the narrative can point to that truth to bludgeon any naysayers. AI can change the world as we know it but that doesn’t mean the current investment environment will lead to riches. Take caution next time you hear about a can’t miss thematic.

Please share any thoughts or topic suggestions with me at [email protected]  

Get Mark's insights in your inbox

Read more of Mark's articles

Previous editions of Unconventional wisdom:

What I've been eating

Pilu is one of my favourite restaurants in Sydney. It is hard to beat the combination of the view of Freshwater Beach, amazing Sardinian food and great service. Below is spaghetti alla chitarra with vongole from my latest trip out to the Northern Beaches. Chitara is guitar in Italian and refers to the preparation which involves the pasta being shaped by pushing it through a set of wires strung between a wooden frame. 

Vongole