Good morning,
When I launched Phase 3 in March, I put 20% of my book into software while the market was busy explaining why AI would destroy it. In last week’s update, the basket was up 47%. I added to it last week on the pullback as I still think there is more to come.
I could spend today talking about how well that trade has worked, but there isn’t much point. The more interesting development here is that I am now prepared to buy a name in the part of the AI trade that I spent most of the summer avoiding.
I had very little appetite for adding to semis and memory. Positioning had not cleared up enough to my liking, and the bar for earnings was less forgiving than it was for software, as I said at the time. Since then, the unwind has left one of the companies supplying the AI buildout on a much lower multiple while earnings estimates have continued to rise - so it deserves another look, particularly when the reasons I had for staying away were price and positioning in the first place.
You could count the larger trades I placed over the summer on one hand. I am happy sitting on positions when I still like what I own. I still think SPX trades with an 8 in front of it by year-end, but getting that right will only take you so far considering the dispersion we have experienced basically all of this year. From Memorial Day to Labor Day, SPX gained 4%, equal-weight S&P gained 7%, and momentum factors got taken to the cleaners. Goldman’s TMT pair was cut in half from its June high and recorded its worst two months on record. So plenty of people could have been right on the index but had a miserable summer in the names they owned. Meanwhile, you had a huge vol compression with VIX closing last week just above its YTD lows, and VXSMH has reset to ~36 from the mid-60s in July. Protection deserves a fresh look at these levels. With a good chunk of unrealised gains, I want to be able to add to something without leaving all of the gains exposed to a reversal. A SPX put will not necessarily do that, particularly if it keeps grinding higher while my names give back some of their gains.
Finally got short EUR/JPY alongside Scott Bessent last week, I can’t say I imagined to be up a few % so quickly, but here we are.
Earnings are why I remain bullish. We’ve seen the forward P/E decline to 19.5x from 22x at the beginning of the year. So, despite market progress, you’ve simultaneously seen multiples compress, and analysts continue to raise estimates.
I discussed this back in August when talking about the hyperscalers. A higher capex bill is much easier to live with when demand and revenue are rising with it. It gets considerably more difficult when the company spends more and leaves you guessing about what comes back.
The forecasts could be wrong, of course. Rising estimates do not settle the arguments over margins or the eventual return on all this investment these companies are making. But they do make it difficult for me to build a bearish view at present simply because September has arrived, and oh boy, have we all heard about September seasonality. Yes, I am aware of the seasonality chart; I see it every year. Please don’t mention it in the comments, I am tired of it.
Touching base on positioning after last week. I obviously described it as getting clean across the board. But I wanted to touch base on a few things… Fundamental l/s funds in Goldman’s data has net exposure at the 15th %ile over the past year, which is clean. However, if you look at a five-year lookback on gross, it’s sitting at the 82nd %ile. A low net number therefore does not necessarily mean that there is very little leverage left to unwind, but when you put things on a five-year lookback, that five-year comparison also includes 2022. So I do stick by my comments that positioning is clean; just take note that some measures aren’t.
The Nasdaq futures data is more awkward for the underowned argument. Between the start of August and the 1st of September, non-dealers bought ~$50B, hedge funds accounted for ~$37B, with ~39B of short covering. There was some fresh buying, but a great deal of the demand has come from cover bid. By the beginning of this month, Goldman had net Nasdaq positioning at the 92nd %ile over two years.
All of this indicates to me that index-level positioning is at a healthy level, and for any further upside, you need the earnings to arrive and investors to choose to add exposure instead of de-netting.
Some of the recent winners may already have priced in a good chunk of the improvement.
The main macro problem here is the long end. Yet equities have been able to tolerate higher yields as growth and earnings have been strong enough to compensate. An energy-led inflation shock would be a less comfortable combination since it would put pressure on costs and leave less room for central banks to ease, but that remains a hypothetical at this juncture. I am more interested in whether this happens rather than trying to extract another market forecast from every sentence of a Fed speaker.
So overall, I remain bullish, but the case for adding size has to be made at the level of the position. Three trades today: the first is a new AI-focused long, the second a hedge, and the third is the EUR/JPY short, which has already moved too far for me to finish sizing for now.
I barely traded over summer, that is about to change.
