Iran in the Rearview AI in the Front View

It looks like the stock market detour to worries about the Iran War and the Strait of Hormuz is about over.  Yes, the Strait is still closed, but the market has moved on.  The market is forward looking, and there is no way that the Strait stays closed.  It will open at some point, and the market will pop on that event.  So in essence, there is more upside potential with a Strait that is closed than when it is open.  I know there are all the oil barrel counters who say that we'll reach hit tank bottoms if the Strait stays closed for another month, etc.  There are still bears talking about a supply chain related recession due to the war.  The market doesn't care about short term supply chain disruptions.  It could if the war lasts even longer and Iran doesn't give in, but that's lower probability than an opening of the Strait on a continuous ceasefire.  

And higher oil prices doesn't really matter.  On an inflation adjusted basis, crude oil prices are more than 20% lower than 2022 after the start of the Russia/Ukraine war.  So oil prices are really not as high as people think.  And even if oil prices go even higher, high oil prices don't affect the rich, who are the primary buyers of stocks.  The bears say you can't print oil.  Well, they can print money, and last time I checked, US stocks were priced in dollars, not barrels of oil.  If governments react to high oil prices by subsidizing gasoline/diesel, or reducing fuel taxes(very likely, and already done in some nations), that's just more money printing.  If the war continues and the US spends $50B/month bombing Iran to the Stone Age, that $50B is coming from deficit financing, which is essentially money printing.  Money printing is bullish for stocks.  

Its back to AI theme.  A couple of weeks ago, Allbirds, a failed shoe and apparel company, decided to pivot to AI data centers.  You got a couple of other two-bit small caps in random sectors doing a similar pivot right after.  You can't make this stuff up.  Its the kind of nonsense that you saw with the bitcoin treasury companies sprouting from reverse mergers of two-bit small cap tickers in 2025.  Bitcoin and ether just happened to make a huge top last year right after pumpers like Tom Lee came out with ridiculous price targets on ether.  

So that is the backdrop that we are in at the moment, with AI firmly in the center of the hype train.  You see the chase continue in semiconductors/hardware.  Its nothing new.  Its just a continuation of the trend that you saw since the 2nd half of 2025, when the KOSPI doubled in a few months, and RAM prices went bananas.  Semiconductors and AI related energy infrastructure are clearly the favorite sectors for those that come on CNBC.  You are seeing huge inflows into semiconductor ETFs.  The extrapolation years out into the future on this AI boom make these stocks look cheap on a forward P/E basis.  But semiconductors are cyclical, and the price of hardware, in particular memory, is obscene.  These RAM prices are not sustainable for the long term.  People talk about demand destruction in oil.  Well, there is currently demand destruction in RAM.  People will just hold off on upgrading or replacing their PC.  Who is the sucker that is going to pay $250 for 16GB of RAM?  It was $50 just a year ago.  And you can bet that China will be flooding the market within the next few years.  

The ratio of semiconductors to software is approaching the highs in 2000, when the dotcom bubble got everyone excited about a permanently high plateau of demand for semis.  We know how that turned out.   

 

From what I read and hear about the AI capex boom, most are expecting explosive growth out to 2027.  So consensus is expecting AI capex spending to keep growing for the next 18 months.  That is a long time to forecast capex investment, even if the semiconductor companies have locked up purchase contracts until the end of 2027.  If the hyperscalers feel like they have enough, they'll signal it way before the end of 2027, when the contracts expire.  

Its simple.  If you believe that this AI boom goes on for a long time, longer than expectations, than you should be bullish the SPX/NDX.  If you believe that this AI boom ends soon, shorter than expectations, than you should be bearish SPX/NDX.  

It has been almost 4 weeks since the market bottomed in late March.  Often these rallies off a panicky bottom last for 4 to 8 weeks before you get a consolidation/pullback.  So we are getting closer to the window where the rally could flatten out and give back a bit of the gains.  I don't expect a big pullback if it does happen, perhaps 3-4%.  But timing the top of this rally will be the difficult part, as this rally has been powerful, and confounded quite a few investors who were very cautious due to the war.  I covered shorts for a loss last week, and don't want to short again until I see more signs of complacency or a bigger move up.  Probably need the Strait to open for the all clear to get short.  I would rather be a little late to short than to be too early.  Its a tough market to get long, but that's probably the higher probability play than to be short over the next 2 weeks.  Watching and waiting for a better opportunity.