Sunday, 6 September 2026 · Weekend Edition · 10:00 London

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

Join Tom, Gerald and Marie for this edition's podcast · 13 min Spotify YouTube

Signals

Berkshire Japan

CNBC Investing's Greg Abel interview says Berkshire owns more than 10% of each of Japan's five trading houses and is running over $15bn of yen-denominated debt with a positive carry between dividends and interest. Berkshire also bought a $10bn Alphabet block at a 6.5% discount, and Berkshire Energy saw about 8% of Iowa load from data centers last year. The bull case rests on long-dated carry and AI data-center power demand, not near-term earnings beats.

ITOCF

Buy Itochu — CNBC Investing reports Berkshire's >10% stake and expects dividend growth plus buybacks; ITOCF rose 7.1% in 1w, so part of the move is already in.

$14.02 -2.37%
MITSY

Buy Mitsui & Co. — Same Berkshire long-hold logic applies to the trading houses; MITSY is 23% below its 52-week high with a 17.1x forward P/E.

$645.6 -3.64%
BRK.A

Buy Berkshire Hathaway — CEO commentary confirms long-term AI energy exposure and Japan carry as durable earnings channels.

GOOGL

Buy Alphabet — Berkshire's $10bn discounted Alphabet stake validates AI positioning; GOOGL trades 22.8x forward earnings, 17% below its 52-week high.

$338.5 -1.11%

Goldman dips

Goldman Sachs analysts explicitly recommend buying Alibaba, Burlington, Ulta Beauty, Aecom and Viking on recent weakness. CNBC Investing reports Alibaba's expected EPS recovery is +64%/+33% yoy for FY27E/FY28E, Burlington posted 2% 2Q comp growth, Ulta is down almost 7% this year, and Viking slumped 20% over the past month. The trade is a deliberate contrarian basket: cheap forward multiples in BABA and ACM versus battered consumer momentum names.

BABA

Buy Alibaba — Goldman's explicit buy call rests on +64%/+33% FY27/FY28 EPS recovery; BABA trades 12.1x forward P/E despite -27% YTD.

“Goldman analysts recommend buying Alibaba, Burlington, Ulta Beauty, Aecom and Viking on recent weakness.”

$113.2 +1.28%
BURL

Buy Burlington Stores — Goldman sees margin flow-through from 2% 2Q comp growth; BURL is 30% below its 52-week high and down 11% YTD.

“Goldman analysts recommend buying Alibaba, Burlington, Ulta Beauty, Aecom and Viking on recent weakness.”

$265.3 +2.62%
ULTA

Buy Ulta Beauty — Goldman thinks the stock is unfairly punished and set to gain share; ULTA rose 1.3% last session but remains 9% lower YTD.

“Goldman analysts recommend buying Alibaba, Burlington, Ulta Beauty, Aecom and Viking on recent weakness.”

$564.1 +1.26%
ACM

Buy AECOM — Goldman says the compressed 10.4x forward P/E reflects legacy project headwinds, not structural damage; ACM is 51% below its 52-week high.

“Goldman analysts recommend buying Alibaba, Burlington, Ulta Beauty, Aecom and Viking on recent weakness.”

$66.58 -0.37%
VIK

Buy Viking Holdings — Goldman says high-income demographics and pricing power offset river low-water concerns; VIK fell 20% in a month but stays 52% above its 52-week low.

“Goldman analysts recommend buying Alibaba, Burlington, Ulta Beauty, Aecom and Viking on recent weakness.”

$85.81 +0.27%

US credit

FT Markets reports spreads on the riskiest junk debt have climbed to their highest since the 'liberation day' tariff shock, with the Treasury sell-off the key driver. HYG and JNK both sit about 1% above 52-week lows, so the widening is real but some pain is already in price. MarketWatch's Brett Arends argues the jobs and Iran midterm headache is good for bonds, which conflicts with that rate pressure and leaves TLT genuinely two-sided.

HYG

Sell High-yield bonds — FT Markets flags riskiest junk spreads at post-liberation-day highs; HYG is only 1% above its 52-week low, so the short is partly late.

$79.16 -0.06%
JNK

Sell High-yield bonds — Same credit stress applies to the broader high-yield basket; JNK is 3% below its 52-week high and 1% above its low.

$95.27 -0.03%
TLT

Watch Long-duration Treasuries — FT's Treasury sell-off pressure argues for higher yields, but MarketWatch says jobs and Iran are good for bonds; TLT sits 1% above its 52-week low.

$82.21 +0.17%

US housing

MarketWatch's Aarthi Swaminathan reports mortgage rates ticked up to a new high for the year, with 7% already reached by some measures. Homebuilder and REIT ETFs are not yet at 52-week lows — ITB is 11% above its low, VNQ is 6% below its high — so the rate shock may not be fully priced. The affordability squeeze is real; the question is whether builders have already derated enough.

ITB

Sell US homebuilders — MarketWatch flags 7% mortgage rates as a new high for the year; ITB is 20% below its 52-week high but still 11% above its low.

$93.91 +0.31%
XHB

Sell US homebuilders — Same mortgage-rate squeeze hits the homebuilder complex; XHB is 16% below its 52-week high.

$103.3 +0.96%
VNQ

Sell Real estate — Rising mortgage rates pressure real estate valuations; VNQ is 6% below its 52-week high after a 0.66% decline last session.

$96.02 -0.66%

Oil & Iran

FT Companies reports US forces struck three Iranian oil tankers after attacks on warships, a fresh escalation near the Strait of Hormuz in a six-month-old conflict. MarketWatch's Brett Arends argues the same Iran headlines, plus weak jobs, are midterm headaches for Trump that are bad for energy stocks and good for bonds. The press is split: geopolitical supply risk versus a political and positioning unwind. USO is up 6.2% in 1w but still 8% below its 52-week high, while XLE sits only 2% below its high and +40% YTD.

USO

Watch Oil — Tanker strikes near Hormuz argue for supply risk, but MarketWatch says Iran is bad for energy; USO rose 6.2% in 1w, making the long crowded.

$142.0 -0.09%
XLE

Watch Energy stocks — Escalation supports energy equities, but the midterm take is bearish; XLE is 2% below its 52-week high and up 40% YTD.

$64.06 -0.87%

China flows

Bloomberg's Charlotte Yang reports traders diversifying away from crowded Korean and Japanese AI trades are buying bullish Chinese equity derivatives. No hard numbers are given, but the flow direction is clear: FXI is up 1.4% in 1w and KWEB rose 2% last session, while both remain far below 52-week highs at 15% and 40% respectively. This is an early rotation story with Chinese equities still under-owned.

FXI

Buy China equities — Bloomberg flags bullish Chinese derivative flow; FXI is 15% below its 52-week high with YTD -9.9% room to recover.

$35.88 +1.53%
KWEB

Buy China internet — AI-related Chinese tech is the likely derivative target; KWEB is up 2% last session but still 40% below its 52-week high.

$26.05 +2.00%

Lithium supply

Bloomberg reports a Brazilian court ordered suspension of all environmental permits and a complete halt at Sigma's Grota do Cirilo lithium project, its sole operating mine. No timeline for permit resolution is given. SGML fell 2.98% last session, is 49% below its 52-week high, and trades at only 7.2x forward earnings — the valuation looks cheap only if the mine restarts soon.

SGML

Sell Sigma Lithium — Court-ordered halt directly threatens the company's sole operating mine; SGML is down 2.98% last session and 49% below its 52-week high.

$12.39 -2.98%

P&C insurance

FT Companies reports an influx of capital into property and casualty insurance has pushed payouts to their lowest level in 20 years and is compressing premiums, while the industry braces for a downturn. KIE trades 4% below its 52-week high and is up 7.4% YTD, so the soft-pricing cycle is only starting to hit income statements.

KIE

Sell P&C insurers — FT flags a 20-year payout low and rising competition; KIE is 4% below its 52-week high after a 1.39% drop last session.

$63.90 -1.39%

Luxury

FT Companies says LVMH's pandemic-era rally has gone into reverse as investors doubt a luxury recovery. MC.PA is 35% below its 52-week high and only 1% above its low, yet forward P/E remains 16.9 — the de-rating has the feel of an unfinished story, not a value floor.

MC.PA

Sell LVMH — FT says LVMH is shedding boom-time gains; MC.PA is 35% below its 52-week high and 1% above its 52-week low with no growth offset.

€429.1 -0.45%

Most original take

CNBC Investing · 5 Sept 2026

Abel: Two ways Berkshire hopes to cash in on AI

The Berkshire interview gets interesting in the details Abel volunteers: Berkshire owns more than 10% of each of Japan's five trading houses, holds over $15bn of yen-denominated debt with a positive carry versus dividends, and bought a $10bn Alphabet block at a 6.5% discount. It also has a 2.5% quota share of Tokio Marine's underwriting book and saw data centers become about 8% of Iowa load for Berkshire Energy. The piece frames AI not as software upside but as power demand and credit-carry exposure.

Read original ↗

Our view

Today's coverage points to a market with two competing trades: Treasury sell-off stress hitting the weakest borrowers, and crowded oil longs on Hormuz escalation. FT Markets says junk spreads hit post-liberation-day highs while HYG sits 1% above its 52-week low, so credit stress is real. At the same time, the US struck three Iranian tankers, USO is up 6.2% in a week but is still 8% below its high, and XLE is 2% below its high and up 40% YTD. That split-brain is the day: fixed income is repricing risk while commodities already discount a supply shock.

The case against our read is that both trades are late. HYG and JNK are already nearly at 52-week lows, so shorting high-yield here is chasing a move that may be priced. USO has already risen 6.2% in a week, and XLE is 2% from a 52-week high — exactly the setup where a de-escalation or a weak jobs print forces a fast unwind of energy longs while bonds catch the bid MarketWatch expects. If Friday's data soften, this whole tape reverses in a day.

What's missing is the China rotation. Bloomberg's derivative flow story gets almost no numbers, yet FXI is 15% below its 52-week high and KWEB is 40% below. That is the most under-owned, non-consensus long on the screen today. Also absent: any serious Asian central-bank reaction to the Treasury volatility, which could hit EM rates next week.

The cleanest expression isn't a single ticker but dispersion: credit spreads widening against a potential Treasury bid, energy long versus credit short. We'd rather own duration at the 52-week low than chase oil at the high, and we'd fade the crowded energy long while adding to under-owned Chinese beta.

Yesterday's signals, today

From the Weekend Edition on 5 Sept 2026 — 0/3 signals moved in the predicted direction.

Share this edition