Junk at 52-week lows, oil at highs — fade the crowd.

Transcript

Tom Junk debt at fifty-two-week lows, oil at highs — and the crowd is piling into both. Fade it. That's the weekend playbook.

Marie Good morning and welcome to Investment Flash, Weekend Edition, September sixth, twenty twenty-six. I'm Marie, joined by Tom and Gerald. We break down the signals so you can trade before the bell — or in this case before Monday.

Tom Okay, first up — Berkshire's Japan play. Greg Abel just laid it out: more than ten percent of each trading house, over fifteen billion in yen debt, positive carry. It's a carry trade with AI power demand on top. Buddy, this is the dream.

Gerald Yeah look, the carry is nice, but Itochu already jumped seven percent in a week. Part of that move is priced. You're buying after the interview, not before.

Marie Hold on — the more interesting part for me is the structural angle. Berkshire has a two point five percent quota share of Tokio Marine's underwriting book, and data centers are eight percent of Iowa load. That's not a trade, that's infrastructure.

Tom Exactly! That's why Berkshire itself is the cleanest way. You get the Japan carry, the Alphabet stake at a six point five percent discount, and the AI power story. All in one.

Gerald Alright, but Alphabet at twenty-two point eight times forward earnings, seventeen percent below its high? Sure, but it's not screaming cheap. Berkshire bought at a discount to market, not to intrinsic value.

Marie See, this is the thing — the bull case is long-dated carry and power demand, not near-term earnings. So if you're buying the trading houses, you're underwriting that rates stay low in Japan. That's a macro bet.

Tom And Mitsui is twenty-three percent below its high with a seventeen times forward P-E. That's not crowded.

Gerald Fair point.

Marie Agreed.

Tom See? Even Gerald's coming around.

Gerald Ha — yeah, you said that about semis last quarter too, and then they stalled.

Tom Ha — okay, okay, one bad call.

Marie Okay, moving to the contrarian basket of the week — Goldman is telling clients to buy the dip in five names: Alibaba, Burlington, Ulta, Aecom, Viking. This is a deliberate 'been punished too hard' list.

Tom Alibaba at twelve times forward earnings with earnings per share expected to recover sixty-four percent next year? That's insane. You're paying nothing for growth.

Gerald Yeah, but Burlington's comp growth was two percent. That's not exactly a recovery story, that's a dead cat with a margin squeeze.

Marie The interesting one to me is Aecom — fifty-one percent below its high, ten point four times forward earnings. Goldman says it's legacy project headwinds, not structural damage. That's a classic 'compressed multiple' setup.

Tom And Ulta is down almost seven percent this year, but Goldman thinks it's gaining share. Consumer beauty is sticky.

Gerald To be fair, Viking's fallen twenty percent in a month, but high-income demographics and pricing power? That's the exact pitch you hear before the next guidance cut.

Marie No, but that's exactly my point — this basket is a bet on mean reversion. It works if the consumer soft-landing holds.

Tom Right, so buy the dip before it's too late, as the note says.

Gerald Ha, yeah, analysts always say 'before it's too late' right after the stock has already cratered.

Marie Alright, point taken.

Tom Okay but the big macro story today is credit. Junk spreads at their highest since liberation day. The Treasury sell-off is hitting the weakest borrowers.

Gerald And that's why the short on high-yield is partly late. The iShares high-yield corporate bond ETF is only one percent above its fifty-two-week low. You want to short something that's already at the bottom?

Marie But the widening is real. The FT says Treasury sell-off pressure is the driver. And here's the split: MarketWatch argues jobs and Iran are good for bonds. So duration is genuinely two-sided.

Tom That's the tension. Do you buy the twenty-year Treasury ETF at the low because bonds catch a bid on weak data, or do you short it because rates are breaking higher?

Gerald I mean, if you're asking me, I'd rather own duration at the low than short junk at the low. One of these is chasing.

Marie Wait — wait a second. The bear case is both trades are late. The high-yield ETFs are already near fifty-two-week lows, and the Treasury ETF is one percent above its low. So the easy money is gone.

Tom But that's exactly when the crowd gets squeezed. If Friday's data softened, bonds rip and junk gets crushed further.

Gerald That's a big if. Jobs have been hot all year.

Marie Okay, but as always, none of this is investment advice. Let's talk housing.

Tom Mortgage rates at seven percent already, according to MarketWatch. That's a new high for the year. Builders can't catch a break.

Gerald But the home construction ETF is still eleven percent above its low. So the rate shock is not fully priced. If you want to short, you're not late — yet.

Marie REITs are down six percent from high. This is a slow bleed, not a crash. The affordability squeeze is the real story for the next few quarters. Builders are used to bad news — they've been saying 'next quarter will be better' since twenty twenty-two.

Tom Ha — fair. So maybe fade the builders now and cover if the ten-year yield spikes?

Gerald Or just avoid the sector entirely. There's no edge if the move is halfway done.

Marie Moving on — oil and Iran. US forces struck three Iranian oil tankers. That's an escalation near Hormuz.

Tom And oil is up six point two percent in a week, but the energy select sector fund is only two percent below its high. That's a crowded long.

Gerald Which is exactly why I'd fade energy. MarketWatch makes the midterm point: Iran headlines are bad for energy stocks, good for bonds. The political unwind can be fast.

Marie But the supply risk is real. If Hormuz gets closed even partially, oil spikes. The press is split, and so are the charts: the United States oil fund is up but still eight percent below high.

Tom So you don't short oil outright — you fade the crowded part. Maybe buy bonds instead.

Gerald Right, and that ties back to our credit chat. Fixed income is repricing risk while commodities already discount a supply shock.

Marie Okay, and here's a callback: Gerald, your short on US crude oil yesterday is now underwater given the tanker strikes. But the unwind logic still holds if the jobs data is soft.

Gerald Ha, yeah, timing is everything. But I'm not covering yet — escalation is already in the price.

Tom Alright, let's pivot to China because that's where the flow is going.

Tom Bloomberg says traders are diversifying away from crowded Korean and Japanese AI trades and buying bullish Chinese equity derivatives. The China large-cap ETF is fifteen percent below its high, the internet ETF forty percent below. That's under-owned.

Gerald To be fair, no hard numbers were given. This is flow direction, not volume. But the rotation story makes sense if China AI is cheaper.

Marie And structurally, Chinese internet is still under-owned after years of regulatory overhang. If that clears, the upside is asymmetric.

Tom Exactly. The most non-consensus long on the screen today.

Gerald But you're buying into a market that's had false dawns before. Remember the reopening trade?

Marie Sure, but this time it's derivative flow, not just narrative. That's real positioning.

Tom Right.

Marie Exactly.

Gerald Fair enough.

Marie Quick one on Sigma Lithium — Brazilian court halted permits, mine stopped. Stock fell three percent last session, forty-nine percent below high, trading at seven times forward earnings.

Gerald Seven times forward earnings looks cheap only if the mine restarts soon. No timeline. That's a value trap until the court clarifies.

Tom Agreed. No way I'm catching that falling knife. Cheap for a reason.

Marie And then there's P&C insurance — capital influx has pushed payouts to a twenty-year low and premiums are compressing. The soft pricing cycle is starting to hit income statements.

Gerald The insurance ETF trades four percent below its high, up seven percent year to date. The market hasn't fully priced the downturn. The short is early but not wrong.

Tom So sell the insurers before the next quarterly numbers show shrinking margins?

Marie And luxury — LVMH is thirty-five percent below its high, one percent above its low. De-rating feels unfinished, not a value floor.

Gerald Forward P-E is still sixteen point nine. That's not cheap for a company with no growth offset. The luxury feel-good factor is gone.

Tom So you wouldn't touch it even at these levels?

Gerald Not until the multiple compresses further or China shows real recovery. Luxury is China-dependent.

Tom Okay, so what's the original take from all this? I think it's the Berkshire interview. Abel basically said AI is not software upside, it's power demand and credit carry. That reframes the entire AI trade.

Gerald But that's a long-dated story. You're not getting near-term earnings beats. It's a structural position, not a trade.

Marie And the detail about the two point five percent quota share of Tokio Marine's underwriting book? That's Berkshire using insurance float to fund AI infrastructure. It's a closed loop.

Tom Exactly. So if you want AI exposure without buying a chip giant at a hundred times earnings, you buy Berkshire.

Gerald Alright, but Berkshire isn't cheap either. It's trading near all-time highs. You're paying for the track record.

Marie Now, our view: the market has two competing trades — Treasury sell-off stress hitting weak borrowers, and crowded oil longs on Hormuz escalation. The clean expression is dispersion: credit spreads widening against a potential Treasury bid.

Tom So you'd rather own duration at the fifty-two-week low than chase oil at the high. And fade the crowded energy long while adding to under-owned Chinese beta.

Gerald The case against our read is that both trades are late. The high-yield ETFs are already near lows, and the oil fund has already risen six percent in a week. If Friday's data softens, energy unwinds and bonds catch the bid. That's the reversal risk.

Marie But the China rotation is the missing piece. Bloomberg's flow story gets almost no numbers, yet the China ETFs are deep below highs. That's the most under-owned, non-consensus long.

Tom And we didn't even mention Asian central-bank reaction to Treasury volatility. That could hit EM rates next week.

Gerald Right, so the playbook is dispersion: long duration low, short energy high, add China. Don't chase the crowded trades.

Marie Before we go — if you're just finding us, hit follow on Spotify or check investmentflash.com for the full digest with charts and sources. We're back Monday's London Edition at seven-thirty a.m. London time — have a great rest of the weekend.