The Drill
New York, New York
With the Greenland crisis averted, we can turn our focus to what’s happening in New York… and Washington!
We usually make fun of the UN General Assembly. Every country is there, and sometimes you get some really meme-worthy moments (like the Haitian president drinking straight out of the water jug), but in general this is very seldom where true geopolitics are conducted.
This year, it’s a bit different. The US, Denmark, and Greenland signed a three-way deal ending the Greenland crisis, and more importantly, we have very serious negotiations going on around the Iran and the Ukraine wars. Could we be entering a season of peace - all settled in New York?
Let’s go over it - also covering tomorrow’s Xi-Trump meeting (even though it doesn’t take place in New York!)
Greenland Deal
Not much in this for investors, but as you might know, I am Danish, so there’s no way around a quick rundown here.
For years (pre-dating Trump), the US has been concerned with the tail risk of Greenland becoming independent and breaking free from the US military arrangements. We even heard some Greenlandic politicians fantasizing about being able to choose between the US, China, or Russia when they gained independence. This was never realistic at all, but it’s understandable why the US wanted to mitigate this tail risk that could ultimately trigger or decide a world war.
Trump decided to settle this - again, completely understandable. However, he somehow ended up actually lobbying for Greenland breaking free from Denmark, the very event that would trigger the tail risk scenario. That (including very overt threats against Denmark) was completely unneeded. You don’t need the ‘art of the deal’ when having conversations with friends or family members.
Ultimately, the US realized that it was better off with Denmark in control of Greenland than Greenland going rogue (which was always the US position). And the three parties concluded a deal that gives the US everything it might want to mitigate the tail risk. It is truly “win-win-win”.
So, I get why this would appear as a big win for Trump, especially if you jump straight from A to C. I won’t argue against that, but just note that the win came at an unnecessarily high cost to US-European trust and relations, which is hurting the US very directly in the Iran and China situations.
Iran running out of cash
Now on to the actually important stuff for investors. I’ve been banging the “peak escalation” drum for a few weeks and also called “peak oil” a week ago. So far it’s looking decent, and we’re even coming close to my model’s prediction of Iran hitting a pain point in their war chest around early October.
That seems to be the reason why Iran (specifically the IRGC) is now signaling much greater willingness to negotiate. The trick is then for the US not to fall into the same trap as in June, where the IRGC settled for the MoU, spent 4-5 weeks refilling their war chest, and then started attacking ships again. This time, the peace must stick.
In any case - for investors, I think the case is clear. We are past the peak in this round of escalations, and on a trend basis we’re looking into a few weeks of detente and easing of oil. That’s the sine wave logic we’ve been trading on for the past months, and it’s working quite well.
That’s no guarantee that this will lead to a softening in energy-driven inflation - especially on a global level if the US moves forward with its diesel export ban.
On that ban: Trump told reporters in New York on Tuesday that he has called for it internally, and Bessent confirmed Treasury is assessing a full or partial version with a decision coming “fast, one way or the other”. US diesel exports hit a record 1.6 mb/d in August, most of it to Latin America and Europe, and Gulf Coast refineries are running at around 97% utilization, so a ban would pull US pump prices down for a few weeks and push everyone else’s up, Europe first. For our purposes, it leaves the short oil lean intact and argues for keeping European crack spreads out of that trade.
Russia-Ukraine talks
The UN General Assembly also brought along new hope to Russia-Ukraine talks. Last weekend saw Putin’s United Russia take their usual election win with 45-50% of votes translating to 75-80% of seats. This time, no real opposition was allowed, and the election was obviously a farce.
Still, Putin can point to having held an election and, most importantly, it passed by without the demonstrations and unrest we’ve seen in the past. Russia has simply moved far left on the Bremmer J Curve. It’s by now a completely closed and authoritarian country without even room for debate or demonstrations.
With regard to the war, we know that both sides are struggling. Ukraine is running out of men and is fearful of the Russian winter offensive potentially taking the cities of Kramatorsk and Slavyansk. Russia is quite simply running out of fuel due to the continued and incredibly successful Ukrainian bombing campaigns against Russian refineries.
I’m not urging you to bet on a Russia-Ukraine peace yet, though. The sides are still too far apart on two issues: 1) the exact future border and 2) security guarantees for Ukraine. Russia wants to annex all of south-western Ukraine and leave the rest of Ukraine as a neutral buffer state. Ukraine wants to freeze the front lines where they stand and then join NATO and the EU. The border issue can be solved, but the security guarantee question is the problem here - and I still don’t see either side giving in.
Xi-Trump Summit
Xi lands at the White House tomorrow with Bezos, Pichai, Altman and Jensen Huang on the guest list and a state dinner to follow. Bessent has been in pre-talks since Sunday, and the shape of the communiqué is already fairly well reported. Expect an extension of the tariff truce that otherwise expires on 10 November (Washington wants 6-12 months, Beijing wants it to run through Trump’s term, and a 90-day extension is the floor everyone is working from), a rare earth commitment traded against that extension, confirmation of the 200-aircraft Boeing order from May with talk of a larger 737 MAX package, soybean purchases counted towards the 25 million tonne annual commitment after China took about 1 million tonnes in the week of 10 September, and a formalized AI safety dialogue. Chips are the one item where nobody expects movement. The Blackwell ban stays, the H200 trickle continues, and the Aivres loophole (over $3bn of Blackwell servers routed via Southeast Asia to ByteDance and Alibaba) is the awkward file the summit would rather not open.
I wrote in August that the parameters pointed to an explosive summit. I still think the underlying facts are unchanged: Chinese exports were up 25% in August on a $119bn monthly surplus, the ag purchases run well behind schedule, and the $14bn Taiwan arms package is still parked. What has changed is that Trump needs a win. Every part of his agenda has been derailed by the Iran war, and the midterms are six weeks out, so the incentive is to bank the truce extension, pose with the CEOs, and leave the hard files for November.
My take: a soft, photogenic summit is now the base case, and that is mildly risk-on for the rest of the week. The trade is in what is not on the official agenda.
The Korean licenses
Samsung’s Xi’an fab and SK Hynix’s Wuxi and Dalian plants run on annual US licenses that expire on 31 December, and the 2027 decision has not been made. Behind them sit 30-35% of Samsung’s NAND output, 40-45% of SK Hynix’s NAND and 35-40% of its DRAM, roughly 30% of global NAND in total. Korean press has started reporting that the fabs could become bargaining currency in a wider deal, and Seoul is visibly nervous: President Lee said on Friday that any further US production by Samsung and SK Hynix would have to reflect Korean industrial strategy; both companies have begun testing Chinese tools in their China fabs, and Solidigm is looking at a US NAND plant to reduce its dependence on Dalian.
Andreas made the point in August that a non-renewal is a concession both sides could quietly want. Beijing loses little from Korean firms no longer running US tools on Chinese soil and can hand CXMT the domestic share, and CXMT is now at 9.5% of global DRAM and in mass production on LPDDR6. Trump gets an America First headline for Micron and SanDisk. I would not put that as the base case for tomorrow, since the license sits in BIS rather than in the communiqué, but it is the tail worth owning. The memory market has no spare capacity, NAND prices are up around 40% in a month, and Korean chip exports just printed a record $4.1bn per working day in the first ten days of September.
So the positioning is the same as I laid out in August: stay long memory, but keep diversifying away from pure Samsung and SK Hynix exposure. SanDisk and Micron are the asymmetric expression: a renewal leaves the NAND boom untouched; a non-renewal takes upwards of an eighth of world NAND supply out of their competitors’ hands. Both doors lead to the same room. On the Chinese side, BYD remains my preferred hedge, with Wang Chuanfu reportedly in Xi’s delegation and over 70% of its power semiconductors produced in-house.
