Breadth is lying: tech at highs, credit at 52-week lows.

Transcript

Tom Alright, we've got tech trying to break out, credit pinned at its lows, and a tape that can't agree with itself. Your Monday setup, and I've got thoughts.

Marie Good morning, and welcome to Investment Flash, London Edition, for Monday, September twenty-eighth. I'm Marie, with Tom and Gerald. Today: robotaxi insurance, Chinese EV inevitability, a US-China truce with no delivery mechanism, and that split between duration and credit.

Tom And Gerald, real quick — your weekend call to sell long duration? The long Treasury fund is hugging its fifty-two-week low. Looking, buddy.

Gerald To be fair, I said bonds were oversold, not that I was buying them. The AI borrowers are saying a half-point move won't stop a data-center deal. Credit's the other side of this trade.

Marie No but that's exactly the point — that AI debt insensitivity is the most original angle today. More on that in a minute.

Tom Alright, let's start with robotaxi insurance. Aioi Nissay Dowa and Itochu are planning to sell robotaxi-specific insurance in the US, targeting Waymo's fifteen-city fleet.

Gerald A first-mover claim on an undeveloped niche — that's the interesting part. Whoever writes the standard policy for autonomous fleets captures the premium pool before the big carriers show up.

Marie And it's not just insurance. It gives Alphabet a tailwind: if Waymo is the reference fleet, the cost of scaling gets a little cheaper.

Tom Exactly — Alphabet at twenty-two point eight times forward earnings with an insurance product building around its robotaxis. That's a buy.

Gerald Hold on, Tom — twenty-two point eight times forward isn't cheap. And Itochu as the other half of that joint venture, entering a brand-new US insurance line, that's the more original first-mover claim.

Marie But the Tesla angle is the contrarian one. Third-party robotaxi insurers lower a cost Tesla currently carries alone. That pressure is real, and Tesla's already down fifteen percent year to date.

Tom Fair, but Waymo's deployment in fifteen cities and Tokyo next year is the actual moat. Insurance follows fleets.

Gerald Alright, but let's not pretend Alphabet is a value stock. I'll take Itochu's exposure with a tighter risk story.

Marie Now the Chinese EV story. Thursday and Friday's talks produced no deal on US access, but Nikkei says analysts still see entry as all but inevitable. BYD is targeting two million overseas sales this year.

Tom And Leapmotor has overtaken Subaru and Mitsubishi in global sales. This is a cost curve already arriving at the US border, not a summit binary.

Gerald So you want to buy BYD? I'll give you that, but the sell calls on Ford and Toyota feel more like a structural problem than a tariff headline.

Marie Hold on — Ford is only down four point seven percent year to date despite the tariff truce. If Chinese EVs arrive even gradually, that mass-market pricing pressure gets ugly.

Tom And Toyota has already lost global share to Chinese brands. The US is the next front, and Toyota is down twelve point six year to date. That's the trade.

Gerald Okay, but the bear case on selling Japan here: the Takaichi government is framed as the summit's biggest loser, and Japan equities sit one percent below their fifty-two-week high. Downside might be underpriced.

Marie Right — and that ties into the US-China trade group. Tariff relief on thirty billion of nonsensitive goods each, but no confirmed purchase volumes yet.

Tom So China equities get a tail-risk reduction, still down fourteen point seven year to date. That's a decent risk-reward if the truce holds.

Gerald But Bloomberg reports corn futures eased because traders are waiting for concrete signs of Chinese buying. If Beijing doesn't confirm purchases, the grain bid unwinds.

Marie Exactly — that's the delivery mechanism missing. And CNBC's line-by-line readout shows differing language on AI and Taiwan. This truce is fragile.

Tom Which is why Bunge works: if US-to-China grain flows resume, origination volumes lift, and it trades at nine point seven times forward. That's the value entry.

Gerald Fair, but the Japan sell is the diplomatic loser angle — if the truce holds, Japan gets squeezed between the US and China. That's a structural short.

Marie Okay, as always, none of this is investment advice. But the pattern today: every bullish signal has a credit or breadth warning right behind it.

Gerald Let's get to the duration versus credit split. Pimco says Australian rate-hike expectations have overshot a slowing economy, and Chris Iggo says bonds rebound after four hard years.

Tom But CNBC counters with AI borrowers saying a fifty-basis-point move will not stop a contract-backed data-center deal. That's the most original take today.

Marie No but that's exactly the point — the long Treasury fund is one percent above its fifty-two-week low, while Oracle and Digital Realty still fund expansion with rate-sensitive debt.

Gerald So Iggo's rebound call is cleaner in intermediate duration. The intermediate Treasury fund is also one percent above its low, less rate sensitivity. That's my buy.

Tom And the sells: Oracle down twenty-nine point nine percent year to date, still debt-reliant. Digital Realty at sixty-four point seven times forward earnings. Higher funding costs squeeze both.

Marie Digital Realty is a data-center REIT funding expansion with debt. Rising yields hit that model directly — this is a structural squeeze, not a cyclical dip.

Gerald Alright, and here's the counter-argument: if AI contracts are already locked with OpenAI or Anthropic, the revenue is contracted. A half-point move won't stop the deal, so maybe the high-yield low is a buying opportunity.

Tom Exactly — that's the AI capex cycle levers through rate rises until contracts reroute or lenders force terms. The trade is not short the AI boom; it's short the assumption that rates discipline it.

Marie Wait — but who bears the AI debt if the fifty-basis-point insensitivity flips? Nobody is asking that. CNBC quotes lenders, FT gives ASML's monopoly, but no one lines up Oracle's balance sheet.

Gerald That's the missing coverage. And Asian central-bank reaction to dollar strength? Australia and Japan moved, but coverage stops at the coast.

Gerald Speaking of credit, Boston Consulting Group says one in six Western European companies is under financial strain. Rising leverage, more vulnerable to shocks.

Marie And the high-yield fund is already at its fifty-two-week low. The stress trade is not crowded — it is early. Germany's leveraged industrial base sits at the center.

Tom But the Germany fund is flat year to date despite that risk. If the stress is real, that ETF has room to fall.

Gerald The thing is, HYG at its low while tech makes new highs — that's the two-market story. This is not one market; it is two.

Marie Exactly. The AI complex borrows at fifty-basis-point-insensitive prices, spends it on EUV machines and data centers, while the rest of credit trades like a slowdown nobody has admitted.

Tom Exactly.

Gerald One hundred percent.

Marie That's the whole story.

Tom Let's lighten it up with dividend income. CNBC's screen surfaces three names with hard numbers: Chevron pays seven dollars twelve annualized, Enterprise Products yields about six percent, Brookfield Infrastructure five point two.

Gerald Now we're talking value. Goldman's Neil Mehta raised Chevron to a two hundred forty target, and it yields about three point five percent while trading at fifteen point one times forward. That's a bond proxy that still pays equity returns.

Marie Enterprise Products has an RBC target of forty-two dollars on a six percent yielder, up fourteen point three percent year to date with AI data-center demand behind it. That's a structural yield story.

Tom And Brookfield Infrastructure — BMO's Devin Dodge flagged the Intel foundry joint venture commissioning by end-2026. The stock sits twenty-one percent below its fifty-two-week high. That's a setup.

Gerald To be fair, with the long Treasury fund at its fifty-two-week low, these bond proxies are doing the work. I can actually get behind Chevron.

Marie Hold on — Gerald liking a dividend payer? Somebody check the tape.

Gerald Ha — fair enough. I'm not opposed to yield, just expensive yield.

Tom Now the big one: the tech sector fund has formed an inverse head-and-shoulders with a measured-move target near two hundred sixteen. CNBC flags five prior breakout precedents since 2023.

Gerald And ASML's CEO says there is only one company making EUV machines. That's a structural monopoly paired with a technical breakout.

Marie Right, ASML is up almost fifty percent year to date, still thirteen percent below its high. The monopoly is real, but the equal-weight S&P has rolled over into a bearish pattern — that's the breadth warning.

Gerald The equal-weight index rolling over is the market saying the party's in the VIP section, and I wasn't invited. That's a breadth story in one line.

Tom Ha — fair, but I've got a backstage pass. And yeah, I called the top on semis in the second quarter, so maybe ignore me.

Gerald We remember.

Tom Exactly — own the tech breakout, fade the equal-weight index. The market is paying you to bet that AI's capex cycle outruns the credit cycle.

Gerald Okay, but here's the bear case on concentration: the tech fund one percent below its high and the S&P holding record territory is exactly what a late-cycle, AI-led market looks like. The high-yield low is the old economy repricing a slowdown it will avoid if AI keeps borrowing.

Marie That's the case against the dispersion trade — if AI borrowers keep absorbing half-point moves, the high-yield low is a buying opportunity, not a warning. The tell is the next ten-year auction and the financing curve.

Tom And if yields stop rising with no AI default, the shorts are crowded. That's the trigger.

Gerald So the convergence today is pretty clear: multiple publishers see AI borrowing as rate-insensitive, and multiple see credit stress building. That's the two-market tape.

Tom Exactly.

Gerald One hundred percent.

Marie That's the whole story.

Marie And our view is not a single ticker — it's the dispersion itself. Own the tech breakout, fade the equal-weight index, and treat every high-yield bounce as a sell until the AI-debt story breaks.

Tom If you need one line, it's long the tech sector fund against short the high-yield fund. The market is paying you to bet that AI's capex cycle outruns the credit cycle.

Gerald To be fair, that's a clean expression. But watch the ten-year auction and the financing curve — if yields stop rising with no AI default, the shorts get crowded.

Marie Listener note — if you're just finding us, hit follow on Spotify, or check investmentflash.com for the full digest with charts and sources.

Tom And we're back with the New York Edition, later today at nine a.m. New York time. I'll be watching that tech breakout.

Gerald I'll be watching the high-yield fund's next bounce to fade.

Marie And I'll be watching the next ten-year auction. See you then.