As I write this, S&P 500 futures are down 0.56% following the PPI release and renewed US–Iran tensions. WTI crude is up around 4%, approaching $100 a barrel again for the first time since May. There’s still no clear path out of the conflict, and it feels like investors are getting nervous. Understandably.
Trading hasn’t been easy this year. If anything, it’s been a reminder that you can get the bigger picture right and still have a pretty rough time making money from it.
On one side, we have several risks that deserve attention.
The yen carry trade unwind. My read is that it has been relatively orderly so far, without the broader disruption people feared. But it’s something I’m watching. Orderly doesn’t mean finished.
The Iran war and rising costs. Oil and diesel are the obvious ones. Fertilizer supply and the potential knock-on effects on agricultural prices worry me too. If diesel stays expensive for long enough, the pressure can spread through transport, manufacturing and eventually the prices consumers pay.
Private credit. The argument you hear is that some of the risky lending that used to sit inside banks has migrated into less transparent parts of the financial system. Higher rates expose weak borrowers, and the concern is that individual problems could eventually become connected.
I don’t know whether that becomes a broader default cycle. I’m not treating private credit as “2008 all over again.” But I’m not comfortable dismissing it either.
Then, on the other side, we have AI.
Everyone knows the argument by now. But knowing it doesn’t mean we’ve understood its scale.
I believe AI could be more transformative than the internet. That’s a big claim, and it’s my conviction rather than something I can prove today. But it’s the reason I remain constructive on the longer-term picture.
I also think that without the Iran war, the S&P 500 and Nasdaq would be materially higher. We can’t know that counterfactual, but that’s my view.
This rally has changed
In 2024, we went through what felt to me like the most speculative phase of the AI trade.
I remember watching speculative shitcos with barely any revenue flying. Having the right story could be enough. Many of those names have since come back down to earth.
From summer 2025 through April this year, the semiconductor rally felt different. There was still excitement, obviously, but there were also earnings behind it. In a lot of cases, earnings growth gave the move something real to stand on.
That doesn’t mean every semiconductor stock is cheap. It means I don’t think we can dismiss the whole move as hype.
At the same time, long semis and short software seemed to become one of the defining trades, particularly among the pod shops.
Software was dead. AI was going to destroy it. Nobody wanted to own it.
Until some earnings started showing more resilience than the narrative suggested, and parts of the sector started finding a floor.
For me, that captured something about this year: price creates narratives.
When a sector falls, we become very good at explaining why it deserves to fall forever. When it starts recovering, suddenly we discover the reasons it might survive.
I’m not immune to this. Nobody is. But I try to notice when a changing price is doing more work than changing fundamentals.
So where does that leave us?
Somewhere uncomfortable.
We have what I think could be an enormous, lasting investment cycle driven by AI. Alongside it, we have a conflict that can damage growth, push inflation higher and make life harder for the Fed.
Both can be true. AI can be transformational while the stocks attached to it have a terrible few months.
I’m not going to tell you exactly how this plays out. I don’t know. And anyone pretending to know probably deserves a bit more scrutiny.
But I do have a working theory.
I think the costs of a prolonged war eventually create pressure for some kind of resolution. My base case is that both the US and Iran have reasons to find a way out.
The midterms complicate that. One possibility I consider is that Iran sees political leverage in keeping pressure on Trump ahead of the election. The other side of that argument is that a president facing less immediate electoral pressure afterward could become less predictable.
That is speculation about incentives, not knowledge of what either government is thinking. I could be wrong about their priorities, their willingness to compromise, or their ability to reach an agreement.
And even if both sides would benefit from an exit, it doesn’t mean they manage to find one.
Still, my base case is that we get enough de-escalation for the market to put more weight back on earnings and the AI investment cycle.
The more damaging scenario is a prolonged conflict: expensive energy feeds into the wider economy, inflation becomes harder to contain, and the Fed responds with further tightening.
That would put real pressure on my bullish view. Strong technology doesn’t make valuations or financing costs irrelevant.
I put less weight on that scenario today. But the incentives I see are a reason for my view, not a guarantee that it works.
Today’s PPI fits that uncomfortable picture. Headline producer prices rose 0.4% in August, taking the annual rate to 5.4%. Energy drove much of the increase in goods prices. But core PPI, excluding food and energy, rose 0.2%, below the 0.3% expected. So there’s something for both sides here: renewed energy pressure for the bears, a softer underlying monthly reading for the bulls. BLS, Trading Economics.
For me, this doesn’t change the thesis much. It reinforces why the duration of the energy shock matters. If it fades, there’s more room for earnings and AI to drive the conversation again. If it persists, the question becomes how much businesses can absorb before passing costs on or taking a hit to margins. One softer core number helps, but it isn’t enough to get comfortable.
Where I stand
I’m leaning bullish, with the expectation that the path could be ugly.
Between now and the midterms, I think we could have sharp moves in both directions. My working target remains 8,000 on the S&P 500, although I’m much less confident about the timing than the direction.
I also sense a lot of fear. That doesn’t prove investors are underpositioned. You can sound worried and still be fully invested. But if cautious sentiment is backed by cautious positioning, a better outcome could force people to chase.
The things that would make me reconsider are fairly straightforward: a conflict that keeps escalating, inflation spreading well beyond energy, or AI earnings and spending failing to support expectations.
I want to keep testing those things, rather than find a new excuse to stay bullish every week.
For now, I think the longer-term opportunity wins out. We may be watching one of the biggest technological shifts of our lives, and I don’t want to lose sight of that every time the market gets uncomfortable.
But I don’t want to confuse conviction with certainty either.
That’s where my head is right now.
Arctic