Is South Korea the Canary in the Coal Mine? - The AI Market Warning.

Ah, yet another big week in AI. This is The AI Exhale, hosted by Vibe CFO with Jez and Cam. Five minutes, one thing that matters, or maybe two. We ignore the rest so you can too.

Jez: All right, so we're gonna run, like we often do, a news item. Is Korea the canary in the coal mine? Quick shout out to canaries.

I actually looked up this saying, 'cause I always question them. Do you know they took canaries down coal mines and they died first because of the toxic gas, and then the miners got out in time? I just never really looked into it. So anyway, that's a little random fun fact. I might do those ongoing, maybe not. Back to the news story. So South Korea, their share market more than doubled in the first half of this year, quite like the US, and we'll pin to that in a moment.

The driver, obviously, for them regarding the AI industry is memory chips. Then in late July, it fell 44% from its June peak. That is the worst since the GFC. It's around $2.18 trillion gone across two sessions. Everyday investors there lost something like $39 billion. It has rebounded, like any dip, by about 18%. But it really just shows, you know, and Cam, you can speak to this, there is hype in terms of valuation going on. And do you agree with the sentiment of this episode title? Is this a warning for the rest of us, or is Korea just a special case?

Cam: That's the connection that people are making in the market. And these stories can always age poorly or nicely depending on the outcomes that flow. We don't know, but people are extrapolating that the market in South Korea is propped up by Samsung and SK, who's a chip provider. And so it's very AI held. AI has a strong control of what happens there. So this is a big dip really connected to the market.

And then they're connecting it to America, saying, look, even though it's a much smaller scale, there's a direct connection to the US market, with a few more companies, but Anthropic and OpenAI and Google and Amazon and Microsoft, I mean, they're connected, right?

Jez: And I'll read a couple of stats around this, and this ROI that we'll often bring into, I guess, every session, 'cause, you know, money talks. MIT found 95% of AI pilots delivered no measurable impact on profit. A survey of 639 senior AI leaders last month found 57% say their returns still don't outpace what they've spent, same as last year, while 93% in that same survey say their capability improved.

Cam: Where something goes wrong, where the return on investment with Anthropic goes wrong, this is like a sneeze that everyone catches. Because they're all selling to each other and have got this connection, if we have something with AI where the return isn't there, with this massive spend, it can flow on like this. That's the connection they're getting at.

Jez: So, you know, we're seeing that the tech is strong. The tech continues to prove itself as, as we like to say, an absolute productivity beast. But yeah, clearly there's a bit of disbelief around the future. Because business, as it is, is to make money.

Cam: That's right. I think the takeaway here is that AI is real, just like the internet was, just like computers were, just like the industrial revolution was real. It's just that stock valuations are disconnected from economic reality. Markets are pricing certainty on AI, and that is a risky choice. We know, and I think most people agree, that AI is pretty influential, but that doesn't mean it's certain. People are pricing in certainty with this, and that's the danger I think people are worried about with the overinflated markets.

Jez: And to your point around the internet, soon after that we saw real early versions of social media, and things like AOL Messenger. Who even knows where that went? Yahoo, they're still bumping along. But it's about picking, from an investor's perspective, and we're not here to provide advice, it's always about picking the right horse in the race. But of course at the moment with AI, we do have that tiny handful of big tech giants leading that charge and obviously having the most money poured into them.

Cam: Yeah. And one of the arguments that I'm seeing as well is that they're valuing them like SaaS providers. And that's the question we don't quite understand yet, because the SaaS model was absolutely gold, right? As soon as you hit product market fit, the cost to put on a new customer was bugger all, so it scaled enormously well to get major profits once it hit off. And that's what we've been running over the past 20 years. Now AI is different. Well, it's shaping up to be not exactly like a SaaS model, because the AI tokens come with it per user. So the incremental cost to bring on you, Jez, as a new customer, we carry the token cost with it. So the margins are squeezed. It's not like traditional SaaS. Now, things can change. That's the current argument right now.

Lots of things can change that redirect this and make it a lot more profitable, but that's the basic argument.

Jez: That's right. Like you said, the cost to serve per client, it's not even like it used to be, an accounting exercise, if you will. It's literally the business model now. So then are the markets over-inflating it, and valuing these companies as SaaS instead of as AI companies?

And that's your five-ish minutes back. Vibe CFO, Australian founded. An AI CFO for small and medium businesses, out of the box, governed and secure. vibecfo.ai. Now you know.