What CSL can teach us about the AI bubble

I have been involved in financial markets for more than 40 years, experiencing my first crash in 1987. It was a bubble just waiting to be burst.

There is a phrase I often use: "If I could do with hindsight what I could do with foresight, I'd be a darn sight better off by a long sight."

And so it has been with CSL, certainly one of the most loved retail investor stocks, held by some investors since it first floated back in the 1990s.

Many have been dyed-in-the-wool CSL fans ever since, and no amount of price volatility could change their minds.

For many years, though, CSL underwhelmed. Then it became a fallen angel, with the share price decimated by wave after wave of bad news that eventually culminated in a massive write-down.

However, that write-down did seem to signal the bottom. Some called it a kitchen-sink moment - everything chucked out there.

Since those bleak days, CSL has recovered dramatically. It has almost doubled from the lows and it is not so much about the numbers as the sentiment surrounding the stock that has changed.

In simple terms, it just got too cheap to ignore. It has now turned what was an annus horribilis into an annus reset.

The power of sentiment

What CSL highlights is the extraordinary power of sentiment - that moment when something changes in the eyes of investors.

What has become increasingly apparent in modern markets is that stocks, themes and ideas have a huge propensity to overshoot. That applies to the downside and the upside, and at $90 CSL had overshot dramatically.

Every broker still seemed to have a buy recommendation or an outperform rating on the stock, yet it continued to be humbled by short sellers and hedge funds looking for an easy target. Then something changed.

It was that reset, the big write-down and a steadying of the ship under interim CEO Gordon Naylor. For years, CSL had been a market darling, trading on a PE multiple that was increasingly difficult to justify when growth appeared to be slowing.

Then it crashed to earth with one of the lowest PE multiples it had seen for years.

The renaissance started with the August results, with the stock rallying hard and fast, and it has continued ever since. The market had gone from pricing in a business that was fundamentally broken to pricing in a business that offered significant upside now that it had been stabilised.

The story unfolds

A number of factors helped the bulls nail their courage to the sticking place.

Guidance was better than expected, with net profit after tax (NPAT) growth of around 5% forecast for FY27. NPAT was ahead of consensus, the cost-cutting program was beginning to deliver, and the $1.1 billion buyback certainly helped.

It has also, no doubt, been helped by defensive buying. In a world of risk, CSL suddenly stood out as a company where much of the bad news appeared to have already been priced in.

There is no doubt in my mind that CSL paid too much for Vifor. That damage is now known and has been priced in accordingly. The core business was still intact. CSL Seqirus has stabilised, while CSL Behring has been helped by improvements in plasma collection costs and manufacturing efficiency.

The tale of CSL, though, is more than just one company: the stabilisation of a business and a recovery from massively oversold levels. It is a story about how sentiment and a narrative take hold, get magnified and push out any of the contrary arguments.

Sentiment and narratives can change quite quickly. It just takes a catalyst.

The AI comparison

When we look at the markets at present, it seems that everything is still about AI. The narrative is crushing all before it, and we are seeing a substantial part of US GDP being driven by the AI boom. Some analysts have said around 20% of US GDP growth is AI- and data centre-driven.

The same could be said, to a lesser extent, in Australia, as we see data centres pull in more investment dollars and commentators fear we could be left behind if we do not embrace this new paradigm.

But we know that all things must end. History tells us that.

What we do know, if we take CSL as the example, is that a trend eventually changes and can go aggressively in the other direction.

The issue is not whether that will happen, it's what that catalyst will be. It is often said that it is not about 'timing the market', but 'time in the market' that counts.

If you had been a holder of CSL at the beginning of this year and didn't check your portfolio, you might imagine that nothing much had happened, because the shares are now around where they started the year.

Missed opportunities

But that misses the point - there was a massive opportunity in between. 'Timing the market', or at least trying to 'time the market', would have increased your returns dramatically if, of course, you had got it right.

And that is the problem, human emotions get in the way. At $90, your stop losses, fear and need to preserve capital would have been overwhelmingly in favour of selling the stock. The 'brave' investor, or perhaps I should say the 'lucky' investor, may have doubled down and bought more.

Some would call that dollar-cost averaging. Others would call it averaging down. Some might even call it throwing good money after bad. But once the train started to leave the station, there were certainly signs that it was worth getting back on board.

The problem with timing the market remains our emotions. There are, of course, plenty of trading programs and techniques that endeavour to take the emotion out of the investment journey. Some impose that discipline on us, and that can certainly be worthwhile.

Others just amplify our fears. And as they say, "As humans are programming these techniques, garbage in, garbage out."

As I wrote at the start of this article, hindsight is a wonderful thing. With CSL, hindsight makes it blindingly obvious that the love for the company was simmering below the surface and only needed a spark to reignite the flame.

Similarly, when we look at the AI frenzy at the moment, we can see fear and nervousness, also simmering beneath the surface, just looking for an excuse to hammer the sector.

We have seen that on several occasions already, although it has generally been nullified by the 'buy-the-dip' mentality that has become so deeply embedded, particularly in the US market.

I have been involved in financial markets for more than 40 years, starting my career in London and experiencing my first crash in 1987. Once again, with hindsight, the reasons and catalyst for the 1987 crash seem obvious.

It was a bubble just waiting to be burst. At the time, we were too busy making money to notice the tsunami of doubt heading our way. And how far we have become divorced from reality.

I do not doubt that when the AI bubble eventually bursts, we will all look back at the glaringly obvious catalyst and the glaringly obvious change in sentiment, much as has happened with CSL.

I may not be able to time or predict what that catalyst will be, but a very profitable exercise is to recognise a change in sentiment and ride that change.

My own timing in CSL was pretty much spot on. I turned bullish on the stock around the $95 level. But I am, by far and away, no investment genius. My target for the bounce was back towards $110 to $120. Way off the mark there.

Although I picked the turning point, and the fact that it had simply become too cheap to ignore, I jumped off the train before it reached its destination. Perhaps that is the real lesson from CSL.

Something I need to learn - 40 years and still learning from my mistakes. Sometimes you can identify the turning point, but not maximise your timing. Sometimes you just have to close your eyes, hold on and stay on the train until the sentiment changes again.

One thing is certain: markets change, but human emotions remain. Staying humble and learning from your mistakes never goes out of fashion. I may be an old dog, but there are plenty of new tricks out there. Hindsight is a wonderful thing.

Get stories like this in our newsletters.