3 risks you won't see coming

In a Trump world where facts are flexible and bad news has become white noise, and nothing matters anymore, just follow momentum, right? 

Honestly, looking at the price action, you'd be forgiven for agreeing, so let's think about what used to matter and what the market says doesn't matter anymore.

• Valuations? Multiple reliable valuation metrics are well above historical highs…. Market: This time it's different
• Fundamentals? mythical forward PE and TAM projections justify the capex-to-revenue gap in AI that is widening by the quarter…… Market: Nobody cares (or checks)
• Geopolitics? Iran, Ukraine, the Strait of Hormuz, tariffs (remember them? They are still in place)….. Market: it will pass
• Economic data? Inflation, rate hikes, spiralling Govt debt…. Market: all priced in, transitory or irrelevant.

The market has decided that in this new world of AI and nationalism, none of these matters 

When the world ignores risks, the next leg lower is almost always caused by complacency, not catalysts. 

The Trinity of Complacency

You can ignore geopolitics. You can ignore valuations. You can ignore economic data, apparently for months at a time. But complacency breeds lazy risk management, which is always dealt with harshly eventually. 

While being over leveraged, over allocated and not bothering with downside insurance can be ignored in isolation, they are a dangerous when mixed together. You can't ignore a margin calls when you’ve allocated all of your cash, and you can’t buy insurance when the house is already burning.

Record margin debt. FINRA margin debt hit $1.42 trillion in May 2026, up 66% from $850 billion just thirteen months earlier. It jumped $112 billion in that single month alone. That's the highest it has ever been. Higher than the dot-com peak. Higher than the pre-GFC high. Higher than anything. When markets are this leveraged, forced liquidations on any decent dip become automatic. The selling feeds on itself, because every margin call creates the next one.

Record Low Fund manager cash levels: No cash left to buy the dip. The BofA Global Fund Manager Survey shows cash at 3.9% of AUM, down from a brief 4.3% rebuild in April. That was the biggest one-month drop in cash since February 2024. The reading crossed BofA's own contrarian sell-signal threshold, which triggers below 4%. Fund managers have less dry powder right now than at any point since the survey began in 1999. Lower than the dot-com peak. Lower than the pre-GFC bond bubble. Lower than anything. The old "buy the dip" mantra needs buyers. There aren't any left.

No hedges. The S&P put-call skew just collapsed to 0.71, against a historical average of 12.13 and a high of 33.68. That's not low. That's the floor. Almost no one is paying for downside protection. Nearly half of fund managers in the BofA survey are running with zero hedges against a sharp drawdown. The insurance has been sold off cheap right before the fire.

In plain English: everyone is leveraged to the eyeballs, no one has cash, and no one has insurance. That's not a market pricing risk. That's a market that's stopped bothering to.

Corporate Debt Market is already telling you

The guys writing the AI cheques (hyperscalers) are looking at being cash flow negative for the first time in years, and are raising debt at a rate of knots, and the market is looking full, judging by Amazon's (NASDAQ: AMZN ) most recent bond raise where it had to sweeten the deal with 18-21 basis points of extra yield to get it over the line.

So why do they need so much money? The AI race got expensive…. really, really expensive. Bloomberg's relative performance chart tells the story in one picture: since mid-May, the AI semiconductor names, the "check receivers," have surged to 260.46, which is great if you are a receiver like NASDAQ: MU or Global X Semiconductor ETF, but in 2 weeks, we are from hearing from the “check writers” and it's unlikely to be pretty. (NASDAQ: MSFT NASDAQ: META  NASDAQ: GOOG ) 

How we are playing it

In short, taking profits in hot money sectors like semiconductors, taking advantage of cheap defensives like healthcare, buying some underpriced oil majors, and most importantly, maintaining high levels of cash.

• Take Profit. The AI story is real, but the SOX index is the most crowded momentum trade in the market. When credit markets are pushing back on hyperscaler funding, capex has to drop and the last place you want to be overweight is the sector that will feel it first.
• Healthcare as defence. ASX: CSL CSL (ASX: CSL) and ResMed NYSE: RMD (ASX: RMD) are non-cyclical earnings stories with pricing power that doesn't care about tariffs or oil prices. Healthcare is where you hide when the cycle turns, and these two are the best quality defensive names on the ASX.
• Oil majors for the oil shock. as much as Trump keeps saying Hormuz is open, the stats tell a different story and if Hormuz stays disrupted, you want the companies that benefit from higher prices, not the ones that get crushed. Chevron NYSE: CVX  (NYSE: CVX) for scale and dividend yield, and Woodside Energy ASX: WDS  (ASX: WDS) for direct LNG and oil exposure at a discount to global peers. Both pay you to wait.
• Buying the dip, but on your terms. I've been through enough corrections to know that nerves of steel are rare. Most people who say they'll buy the dip end up selling the bottom or doing nothing at all. The emotional investor freezes when the screen turns red, so the simple solution is to structure the entry. 

We've built our Buy the Dip strategy around exactly this kind of setup: defined entry points tied to the market's worst settlement day over the next 4 months in the S&P500  ASX: SPY , called a lookback. Click here to find out more 

As bearish as this wire has seemed, we aren’t going all “Michael Burry” on the market, we are bullish about the future and think there is plenty of opportunity ahead, but buying at these levels, with this market positioning, at this level of uncertainty, on these valuations, is portfolio suicide and will have you playing catch up for a year (at least), when all you need to do is