Tuesday, 29 September 2026 · New York Edition · 09:00 New York

AI capex is everyone's trade now. Nvidia's passing the risk.

Signals

⚡ Convergence radar: Buy 005930.KS×3Buy KKR×3Buy NVDA×3

AI infrastructure

Samsung committed $1bn to Helix Digital Infrastructure across six group affiliates, pushing the KKR- and Nvidia-backed platform's secured capital past $11bn — WSJ and The Information both. The same morning, Broadcom and Toppan opened Singapore's first advanced chip substrate plant months ahead of schedule, per Nikkei, easing a genuine AI packaging bottleneck. Strategic capital and physical capacity are arriving together, which is the bull case. The bear case: neither announcement names a customer, a project or a megawatt.

005930.KS

Buy Samsung Electronics⚡ — Nikkei and WSJ both report six Samsung affiliates committing $1bn to Helix, taking the chipmaker into the data-centre build-out rather than just supplying it.

KKR

Buy KKR⚡ — WSJ and The Information both put Helix's secured capital above $11bn after Samsung's cheque, yet KKR is 27.7% lower YTD and 36% below its 52-week high — the private-credit discount is already in the price.

$93.24 -3.55%
NVDA

Buy Nvidia⚡ — The platform Nvidia backs just raised another $1bn of strategic capital; NVDA sits 3% below its 52-week high and +21.2% YTD, so the cycle is priced and then some.

$228.9 +1.68%
AVGO

Buy Broadcom⚡ — Nikkei reports Broadcom's Singapore substrate plant opened months early, easing a real AI packaging constraint; AVGO is flat YTD and 29% below its 52-week high at 18x forward.

$349.6 -0.92%
SMH

Buy Semiconductors⚡ — Two separate AI-supply stories landed the same morning; SMH is +60.7% YTD but 11% below its 52-week high, so the complex hasn't fully round-tripped.

$600.0 -1.08%

AI credit risk

Nvidia is approaching insurers to spread the risk of the AI build-out and pull Wall Street deeper into financing it — FT Companies and FT Markets both ran the story, which is one report filed twice. No deal names, no figures, no counterparties. The direction of travel is what matters: if AI capex increasingly gets funded through insurance-wrapped credit, an AI disappointment lands on reinsurers and high-yield spreads before it lands on Nvidia's income statement. There's nothing to trade yet except the plumbing.

HYG

Sell High yield credit — If AI capex moves onto insurance balance sheets, an AI miss feeds straight into spreads — and HYG is already sitting on its 52-week low.

$77.54 -0.41%
CB

Watch Chubb — FT Companies and FT Markets both frame insurers as the new marginal funder of AI build-out; Chubb at 11.4x forward and 1.7x book is cheap enough to absorb a mispriced new exposure.

$331.4 -0.60%
AIG

Watch AIG — The cheapest large commercial insurer at 0.94x book and 2% off its 52-week low is the natural counterparty for transferring AI construction risk.

$72.78 -1.87%

Space stocks

Starship reached orbit for the first time on Sep 28 and deployed the newest Starlink satellites — the payload detail is WSJ's, the milestone is The Information's. A working reusable heavy-lift rocket is bad news for everyone else launching: Rocket Lab and AST SpaceMobile carry the most direct competitive hit, and Iridium's pricing power erodes as Starlink capacity gets cheaper. MarketWatch adds the wrinkle — analysts say SpaceX's real growth engine is renting out data-centre compute, not rockets.

ARKX

Buy Space & defense⚡ — The Information and WSJ both confirm Starship's first orbit plus the Starlink deployment; ARKX is 15% below its 52-week high and only +6.8% YTD.

$32.07 -1.91%
DXYZ

Buy Destiny Tech100⚡ — MarketWatch's compute-rental thesis puts a re-rating story on the only listed vehicle with direct SpaceX exposure.

$29.86 -2.48%
VRT

Buy Vertiv⚡ — If SpaceX sells compute, someone supplies the power and cooling; Vertiv is 36% below its 52-week high on 26.8x forward.

$244.0 -3.65%
RKLB

Sell Rocket Lab⚡ — A working reusable Starship raises competitive pressure on smaller launch providers; RKLB is 52% below its 52-week high on a 1,584x forward multiple — no valuation cushion.

$72.19 -2.38%
ASTS

Sell AST SpaceMobile⚡ — A bigger Starlink constellation widens direct-to-cell's dominant player's lead; ASTS is 54% below its 52-week high and unprofitable on a -47x forward.

$61.00 -1.31%
IRDM

Sell Iridium⚡ — Cheaper Starlink capacity keeps eroding legacy satellite pricing power; Iridium is +172% YTD and 16% off its high, so the easy money is made.

$48.33 -0.86%

China e-commerce

Shein's first earnings as a public company landed badly: Q2 net revenue grew 0.86% to $11.1bn, first-half operating income fell 53%, and the stock slid as much as 12.1% to a record low — Bloomberg and Nikkei both. The Hong Kong listing is barely four weeks old. The read-through cuts both ways: a broken marquee China IPO dents the whole listings complex, but Shein's slowdown is share moving to Temu, which is PDD.

PDD

Buy PDD Holdings — Shein's 0.86% revenue growth is share moving to discount rival Temu; PDD trades at 6.4x forward, 44% below its 52-week high.

$78.49 +1.19%
FXI

Sell China large-cap equities — Bloomberg and Nikkei both flag the 53% first-half profit drop and 12% share slide a month after listing; FXI is 19% below its 52-week high on 0.85x book.

$34.17 +0.62%

Oil

Oil rose as US-Iran ceasefire talks stayed deadlocked, outweighing signs Gulf crude exports are recovering — WSJ ran it on both its markets and business desks. The bid is entirely a risk premium, not a demand story: neither story carries a barrel volume or a price level. Every session the talks stall is a session the premium stays; the first credible de-escalation headline unwinds it.

USO

Buy Crude oil — WSJ ran the deadlock on two desks; USO is +117.5% YTD and only 8% from its 52-week high, so the geopolitical premium is already well paid for.

$150.0 +1.13%
XLE

Buy Energy stocks — Sustained crude strength flows into integrated and upstream earnings; XLE is 6% off its 52-week high and +36% YTD.

$62.10 +0.10%
CVX

Buy Chevron — Chevron is the low-beta way to hold the risk premium — no single-field or shipping exposure; it sits 5% below its 52-week high on 15.2x forward.

$206.4 +0.94%
JETS

Sell Airlines — Higher jet fuel squeezes airline margins if the crude bid holds; JETS is flat YTD and 16% below its 52-week high.

$28.49 -1.69%

Goldman succession

The FT and WSJ disagree on the same story. WSJ reports Goldman's board has discussed naming John Waldron CEO, with Solomon moving to executive chair as soon as next year; the FT reports Goldman saying there is no definitive timeline. A pre-announced handover to the long-flagged internal candidate takes governance risk off the table; an unsettled one leaves it there. Either way Goldman's shares are flat YTD and 21% below their 52-week high — the market isn't paying for any outcome.

GS

Buy Goldman Sachs — WSJ reports the board has discussed naming Waldron CEO as soon as next year; GS is flat YTD and 21% below its 52-week high on 12.5x forward earnings.

$916.3 -2.05%
MS

Watch Morgan Stanley — Leadership stability is a relative advantage while a rival is in transition; Morgan Stanley is 17% below its 52-week high on 14.2x forward.

$193.6 -1.36%
JPM

Watch JPMorgan — Dimon's own succession question colours how investors price bank CEO changes; JPM is the closest of the three to its 52-week high, 8% below.

$336.6 -1.89%

Luxury

FT Alphaville's line of the day: Rolex was for crypto, Ferrari is for AI. The argument is that AI-generated wealth chases the same top-end status goods crypto money chased in 2021, which puts Ferrari and the big luxury houses on the right side of a wealth-effect trade. It's a headline with no numbers, so treat it as a lens rather than an estimate. The setup is the interesting part: Ferrari 21% below its 52-week high, LVMH 39% below and down 38.2% YTD.

RACE

Buy Ferrari — FT Alphaville names Ferrari the status asset absorbing AI wealth; RACE fell 3.24% in the prior session and sits 21% below its 52-week high on 31.6x forward.

$398.0 -3.24%
MC.PA

Buy LVMH — If AI wealth drives top-end status spending, the big houses capture the same bid; LVMH is 39% below its 52-week high and -38.2% YTD on 16x forward.

€396.7 -0.53%

US solar

Bloomberg's read on the White House's solar offer: it comes with a catch. Unfreezing some renewable projects delivers only a modest boost to US clean-power growth, and the piece names no project, no megawatt and no dollar. A modest-at-best thaw doesn't fix the sector's demand problem, which is why the tape has been brutal — TAN is at its 52-week low and down 7.5% on the week, FSLR down 13.9% over the same stretch.

NEE

Hold NextEra Energy — NextEra's pipeline gets some project relief, but the catch caps the upside; the stock is 24% below its 52-week high at 17.2x forward.

$75.49 -0.78%
TAN

Sell Solar — Bloomberg judges the unfreezing only a modest boost and names no megawatts; TAN sits on its 52-week low, -16.7% YTD.

$43.00 -2.41%
ENPH

Sell Enphase Energy — Residential demand stays weak if the relief is utility-scale; Enphase fell 4.6% in the prior session and is 58% below its 52-week high.

$30.91 -4.60%
FSLR

Watch First Solar — US-made panels could benefit from any unfrozen projects, but the scale looks small; FSLR is 46% below its 52-week high and down 13.9% on the week at 7.4x forward.

$173.0 -2.67%

Netflix

Deutsche Bank says buy Netflix after its September selloff, per CNBC — the only genuinely contrarian analyst call on the tape today. No price target, no estimate change, no valuation work in the note as reported, which is why this is one desk's view and not a thesis. The setup does the arguing: NFLX is 45% below its 52-week high, down 23.9% YTD, and fell 2.69% in the prior session.

NFLX

Buy Netflix — CNBC reports Deutsche Bank telling clients to buy the September selloff; NFLX is 45% below its 52-week high and 18.2x forward.

“Netflix has sold off in September. Deutsche Bank says buy the dip”

$69.23 -2.69%

Nickel supply

Indonesia's nickel quota cuts are reshaping the ore trade: Philippine ore exports to Indonesia rose 82% year on year in Jan-July 2026, per Nikkei, and Manila now wants domestic processing capacity. Industry experts warn the build-out will be hard — smelting is capital-heavy and slow. In the interim, tighter Indonesian supply lifts ore prices and the diversified miners selling into it.

VALE

Buy Vale — Nikkei reports Philippine ore exports to Indonesia rose 82% in Jan-July after quota cuts; Vale is 24% below its 52-week high on 7x forward earnings.

$13.59 -0.15%
BHP

Buy BHP — Quota cuts tighten global nickel supply and lift diversified miners with exposure; BHP is +36.9% YTD and 14% below its 52-week high.

$84.58 -0.46%
INCO.JK

Buy Vale Indonesia — The Indonesian producer with local smelting capacity gains pricing power as quotas bite; Vale Indonesia is 46% below its 52-week high and down 10.6% on the week.

$4280 -2.73%

Soybeans

China's booked 2026 soybean volume is up more than fourfold year on year, and prices have risen on the buying — Nikkei frames it explicitly as leverage over US farmers ahead of November's midterms. The trade is a bet that Beijing keeps buying through the election, which is the same bet US farmers are making. Watch the midterms and any Trump-Xi summit read-through for the exit.

SOYB

Buy Soybeans — Nikkei reports China's booked 2026 volume up more than fourfold with prices already rising; SOYB is +25.3% YTD and 3% below its 52-week high.

$27.35 -1.33%
ADM

Buy Archer-Daniels-Midland — Bigger grain flows and export volumes lift the merchants' trading arms; ADM is +36.1% YTD on 14.2x forward.

$80.38 -0.91%
BG

Buy Bunge — Bunge is directly geared to US soybean export demand from China; it's +17.7% YTD on 9.7x forward.

$109.1 +0.06%

China autos

GAC agreed to buy FAW Group's 50% stake in a Toyota joint venture, with FAW becoming GAC's second-largest shareholder, per Nikkei filings — another step in consolidating China's crowded, state-backed auto sector. Toyota's 50% JV interest and production footprint are unchanged; this is a shareholder reshuffle, not an operational fix. GAC's problem is overcapacity and a price war, and a stake transfer solves neither.

2238.HK

Buy GAC Group — Nikkei reports GAC will buy FAW's 50% Toyota JV stake with FAW becoming its No. 2 shareholder; the transfer fixes the structure, not the price war.

7203.T

Hold Toyota — Toyota's China partner changes hands but its 50% JV interest and production footprint stay intact.

Most original take

FT Alphaville · 29 Sept 2026

Rolex was for crypto, Ferrari is for AI

FT Alphaville argues the AI wealth boom is doing to Ferrari what the crypto boom did to Rolex — a status asset bid up by a new class of newly rich buyers. The parallel matters because it reframes luxury demand as a wealth-effect trade rather than a China-recovery trade, which is how most desks still model it. There are no numbers in the piece, so it's a lens, not an estimate. But look at the setup: Ferrari 21% below its 52-week high, LVMH 39% below.

Read original ↗

Our view

Read this morning's signals together and they say one thing: the AI build-out is being financed by everyone except the companies running it. Samsung writes a $1bn cheque into a KKR- and Nvidia-backed platform now holding $11bn. Nvidia asks insurers to carry the risk. Broadcom opens its own substrate plant early so it stops waiting on the supply chain. Meanwhile the AI tape has stopped making new highs — Nvidia is 3% off its 52-week peak, SMH is 11% below its despite a 60.7% YTD run, Broadcom is flat YTD and 29% below. Capital keeps arriving; the equities have gone quiet. That gap is the trade.

The credit side is where it bites. KKR is down 27.7% YTD and 36% below its 52-week high, and HYG is parked on its 52-week low. The market is already discounting private-credit strain — which is precisely the channel Samsung, Nvidia and Helix are leaning on. If AI capex gets increasingly insurance-wrapped and privately funded, an AI disappointment doesn't land in Nvidia's income statement first. It lands in HYG's spread.

The case against us: risk transfer isn't a warning, it's a maturing capital structure. Railroads, telecoms, and every data-centre cycle before this one got financed the same way — sponsor equity, specialist credit, insurance on top. The AI bear trade has been wrong for two years, and Nvidia 3% off its high is the market saying so again. What would change our mind? A clean break below SMH's range on volume, or a named Helix project that can't get financed. Neither has happened.

What's missing from this morning's coverage is power. We read five AI-infrastructure stories and not one mentions an electricity price, a grid queue or a megawatt. Vertiv is 36% below its 52-week high, and the substrate plant exists because the physical supply chain was too slow — the bottleneck is the story, and the press is still writing about the cheques. The cleaner expression isn't a single ticker; it's the spread between AI infrastructure suppliers and AI beneficiaries. Scarcity sits on the supply side.

Yesterday's signals, today

From the New York Edition on 28 Sept 2026 — 2/2 signals moved in the predicted direction.

Today's scan: 225 articles considered · 21 surfaced · across 14 publishers

Generated Tue 29 Sept 2026 8:05am New York · 34,933 views

Bloomberg Markets · CNBC Investing · CNBC Markets · CoinDesk · FT Alphaville · FT Companies · FT Lex · FT Markets · MarketWatch Top · Nikkei Asia · SCMP Business · The Information · WSJ Business · WSJ Markets

Next edition: London Edition tomorrow at 07:30 London.

Methodology
Share this edition