Steno Signals

Wars never repeat, but they rhyme… Why this reminds me of March/April again

Happy Monday, and welcome back to my editorial on everything macroeconomic!

I have been saying it for two weeks, so let me write it down properly: the current setup reminds me more and more of March and April. Not because the facts are the same… they are not… but because the SHAPE is the same. A geopolitical shock that everyone treats as permanent, a market positioned for the shock to get worse, and a set of hard data that quietly says the opposite. Wars never repeat. But they rhyme. So today I will address three questions, with the pros and cons of each and my own bottom line:

• Is the war fading, the way it did in March?
• Is positioning as stretched as it was in March?
• Will semiconductors fly again, the way they did in April?

And then the one honest difference from March, which sits in our own nowcast. Follow along below.

Theme 1: Is the war fading, the way it did in March?

Let me start with what Mikkel laid out in our Geopolitical coverage already on Friday, because the news flow has quietly turned and almost nobody has updated. The Saudi East-West pipeline that the Houthis hit (from somewhere in Iraq) will be 50% operational within DAYS, not the weeks the first headlines gave us… so the premium that got priced in one go comes out in one go. The battle for Hormuz is slowly degrading in intensity, strike by strike, as both sides run out of high-value targets. The Saudis are working hard on a ceasefire with the Houthis, which would close the third refining front before it properly opened. And an Iran-Qatar deal that gets products moving out of the strait against tolls could arrive quickly… twelve months ago, paying the Iranians to exit Hormuz was the worst case; now we are dreaming of it.

Add the politics. Every single part of Trump’s agenda is currently halted by this war: China policy, trade, European defense, energy, cost of living. The pressure on him to get it solved is very, very high, and the administration is preparing the public for exactly that… Trump is on Truth Social blaming Ukraine’s refinery strikes for the diesel price rather than Iran, which is what you do when you want people to stop looking at the strait. None of this is a grand peace. All of it is the sine-wave turning, exactly as it did in March when the world was convinced Hormuz would stay shut forever and then flows improved within weeks of the escalation peak.

Now look at positioning in oil itself. Speculators are net 44% long of open interest in oil futures again… approaching the same level as the March peak, when the war was new, and everyone piled in. That trade was ultimately wrong then; it took weeks for the crowd to unwind and the barrel went down 25%. Same type of positioning, same story, and this time the news flow is turning BEFORE the positioning has cleared. I refuse to see oil go much higher from here. Fluctuations, yes, every time someone blows something up in the region… but on a trend basis, we are in bear mode for the coming weeks.

And the physical market agrees, in its own quiet way. The Houston export spread (Argus Houston versus WTI) has ticked up again, which shows a foreign bid for US barrels… the same signal we saw in March. But look at the size: 2 dollars now versus 7 at the March panic.

The physical market is a little less relaxed than a fortnight ago, not stressed. In March that spread went vertical… That gap between the physical whisper and the positioning scream is now the trade. Short the cracks, short the barrel from this zone.

Theme 2: Is positioning as stretched as it was in March?

Here is the rhyme that matters most, because positioning is what turns a fading war into a rally. The AAII bull-bear spread is now WORSE than it was at the peak of the war in March. Read that again. Retail is more bearish today, with the war degrading, than it was when missiles were flying and the strait had just shut. Maybe the Fed hike is the culprit? I don’t know, but the euphoria in equities is absolutely nowhere to be found.

Everyone is already in the camp that it cannot be solved. So how much worse can it realistically get from here? If it gets a little worse, it is not a surprise. We are already there.

The dollar tells the same story from the other side. USD positioning is crowded long as a share of open interest, and yet EURUSD has been remarkably resilient… the euro simply refuses to fall substantially despite everything that should have pushed it down.

Energy up? Typically bad for Europe. The Fed hiking to 4% with dots at 4.1% for two years? Typically dollar positive. Very high US yields? Same. And the dollar could not truly rally into the CPI, the hike, or the dots. In every historical hiking cycle, the dollar is firm INTO the first hike and then goes anywhere… this time it barely managed the run-up, into the most crowded long in years. When the most hawkish set of dots in years cannot really lift a crowded long, the crowd is the problem, not the fundamentals.

So tally the cocktail: long dollars, VERY short bonds (two million contracts in the 2yr), 44% long oil, retail more bearish than at the war’s peak, betting against the momentum names in AI, and nobody expecting a solution.

In March, that exact cocktail produced one of the biggest wrong-footings of the decade when the war started fading. As my colleague Mikkel likes to put it: all it takes is one tweet. I am not trying to talk anyone out of euphoria, because there is none to talk out of. That is what makes me comfortable staying with the short USD lean, constructive on the long end, and short oil on trend. Those could be my famous last words. They were not in March.

Theme 3: Will semiconductors fly again, the way they did in April?

Here is the part everyone stuck watching the wars is missing, and it is the reason March keeps coming to mind. Back then, the one data series that did not care about the Iran escalation was Korean chip exports… and when the war faded, semis ripped in April while the crowd was still watching Hormuz. It is happening again. The first-ten-day export data for September printed a RECORD 4.1 billion dollars per working day, above the June highs that everyone took as the peak of the memory cycle. In year-over-year terms, exports are going through the roof. And the value gap between the Korean export data and the semiconductor equities has widened past 50%… the same shape it had in March before the April rally.

The demand story has not slowed; it is setting records, while the multiples have compressed on peak-AI talk, a safety essay written by a chatbot, and a Fed that will sit at 4%. NAND prices are up 40% in a month. Nvidia has committed 279bn to supply, mostly memory. And the market has decided this is a hardware slowdown. It is not. It is a hardware SCARE, the fourth this year, and every one of the previous three was a gap that got bought.

So let me say it plainly: semiconductors will suddenly FLY again. I cannot tell you the day. I can tell you that every time this gap has opened this wide, it has closed upwards, and that it closes fastest when nobody is positioned for it. Nobody is currently positioned for it. Which is why we sold four losers on Friday and put all of it into Marvell, the purest inference expression we could find.

Now the honest difference from March, because rhymes are not repeats. In March we had a US growth momentum pick-up carrying the rally: the growth nowcast was rising into the spring and momentum stocks followed it two months later. Today the growth nowcast sits at 27%, up only marginally from 25%… The inflation probability spiked to 60% into the CPI and has rolled back to neutral, which tells you core is soft under the energy shock, but growth has not turned yet. Semis can rally on the value gap alone. The momentum COMPLEX needs the growth nowcast to turn. Watch this line before you re-lever everything that worked in April.

And the second difference, which is also why I am not worried about it: the Fed just hiked, and central-bank hikes slow US activity with roughly a ten-month lag. Our CB decision index has rolled over, which points to a softer ISM around the middle of next year. In March, the Fed was on hold, and the cycle had room to run.

Today the cycle has a clock on it. But look at the chart: this is (so far) a MID-cycle adjustment, not a cycle killer… the 1995 and 2019 template, where the first hike marked the bottom in risk assets rather than the top, and the slowdown arrived a year later and shallow. The rhyme with March is intact for the next two quarters. The clock matters for 2027, and we will deal with 2027 in 2027.

The bottom line: wars never repeat, but they rhyme — and this one is rhyming with March on all three counts that matter. The war is fading in intensity with the pipeline coming back, Hormuz degrading, and two ceasefire tracks live… while oil positioning sits at the March peak and the physical market whispers rather than screams.

Sentiment is WORSE than at the peak of the war, the dollar cannot rally on the best news it will ever get, and bonds are monstrously short… the exact cocktail that got wrong-footed in the spring. And Korea is doing the bizarre thing again while everyone watches Hormuz. The one difference is that the US growth nowcast has not turned, which means semis lead on the value gap before the momentum complex follows… and the Fed’s clock is a 2027 problem.

So: long semis and memory with Marvell as the inference add, short USD as a lean, constructive on the long end, short cracks and oil on trend, gold over Bitcoin. It rhymed in March. I think it rhymes again.

Have a great week ahead, and good luck out there!