Jabil's third hyperscaler win is bigger than the earnings beat

Jabil's third hyperscaler win is the number that changes its FY27 story

Jabil Inc. (NYSE:JBL) posted Q3 fiscal 2026 numbers that most investors will read as a clean beat and move on. Revenue of $8.8 billion, up 12% year over year. Core EPS up 24%. Free cash flow guidance raised. CEO Mike Destor ran the prepared remarks at zero fillers, which in earnings-call analysis is about as close to a tell as you get: the man knows his numbers cold.

But the number that matters is not in any of those lines. It is the third hyperscaler customer win.

Jabil now manufactures for three of the major cloud infrastructure players. Count them: three. The hyperscalers are spending a combined $300-plus billion on capital expenditure this year, and having three of them dependent on your supply chain is not a vendor relationship. It is a structural position. Each new logo compounds the previous ones because the hyperscalers talk to each other about who can actually execute at scale, and a third win is proof the first two were not flukes.

Destor's Q&A got measurably less smooth the moment analysts pushed on the Adani alliance and fiscal 2027 margin shape. Filler density tripled. That hesitation is worth understanding rather than dismissing. The Adani framework, if signed, positions Jabil as the manufacturing operating system for India's AI infrastructure buildout. That is not a rounding error on a model. FY28 revenue from that geography alone could add a number the current consensus has nowhere to put.

The second-order read from that Adani discussion is the one worth sitting with. India is trying to build out AI data center capacity roughly a decade faster than China did its manufacturing base. The physical inputs required (liquid cooling systems, high-density power distribution hardware, advanced PCB interconnects) all have to come from somewhere. Jabil is positioning to be the integrator of record for that build. If it gets there, the companies supplying those components into that geography get pulled along in its wake.

There is a bear case. Greg Hebert, the CFO, occupied only 16% of call airtime. That is low. When a CEO dominates an earnings call to that degree, either the CFO is new and still finding his footing, or the commercial story is outrunning the financial controls. Destor is clearly the commercial engine here, which is a strength until it is not. The 10.5-second pause before discussing blockchain and IT performance was an anomaly too, though the surrounding context was unclear enough that it is hard to call definitively.

The September investor briefing is where the FY27 shape gets revealed. Destor knows what he wants to say there and is deliberately not saying it now. That is standard CEO playbook for managing a narrative across a calendar, and it is a reasonable choice. But it means anyone pricing JBL off this quarter alone is working with half a picture.

At a 24% core EPS growth rate, with three hyperscaler relationships compounding, and an India positioning that has not yet shown up in any forward estimate, the Street's current model is almost certainly stale.

✦

Quick Takes

Three more stories worth knowing about

CarMax's pricing algorithm is doing something unusual for a used-car lot

New CEO Keith Barr came from hotel revenue management, and he is applying that logic to CarMax (NYSE:KMX) at scale. Instead of a fixed gross profit per unit target, the company is now feeding external market data and vehicle-level demand signals into its pricing. Gross profit per unit came in better than the $300-down guidance in Q1. That alone is not a transformation, but CAF's Tier 2 lending penetration jumping from roughly 10% to 25% of CAF volume in a single year is. If that reaches 50%, CarMax is recapturing margin it has been handing to Ally and Capital One's auto arm for years. The fall strategic update is when Barr actually proves the surgery works.

SpaceX's IPO repricing has a second beneficiary most investors are skipping

Rocket Lab (NASDAQ:RKLB) gets the obvious re-rate conversation now that SpaceX is public at a $1.75 trillion valuation. But Intuitive Machines (NASDAQ:LUNR) is the more interesting play on the same thesis. It flies its lunar landers on SpaceX rockets, holds NASA contracts worth billions, and is a direct customer and supplier to the newly public market leader. When the sector gets repriced, the companies embedded in SpaceX's own supply chain tend to move with more conviction than the ones simply offering an adjacent service. LUNR is small and volatile, but the re-rating logic is cleaner than most.

Anthropic's fall IPO could be the catalyst that re-prices SoundHound

Anthropic filed confidentially on June 1 and is targeting a public listing this fall. When a trillion-dollar AI lab goes public, it does not just create a new stock. It resets how every Fortune 500 board thinks about AI adoption timelines, which pulls forward enterprise spending on the application layer. SoundHound AI (NASDAQ:SOUN) is a pure-play voice AI company already deployed in restaurants, cars, and enterprise environments. It is exactly the kind of AI-native software name that gets dragged higher when a blockbuster lab listing reminds institutional allocators how much of the AI application layer they do not yet own.

✦

Overlooked Stock

KSPI

Joint Stock Company Kaspi.kz

Kaspi.kz (NASDAQ:KSPI) is a super-app operating across Kazakhstan, Azerbaijan, and Ukraine that combines payments, e-commerce, and fintech into one platform. It is the dominant digital financial infrastructure across its home markets in a way that has no real Western parallel except perhaps WeChat Pay crossed with Amazon crossed with a retail bank. The payments and marketplace segments feed each other: more merchants pull in more consumers, more consumer data improves lending decisions, better lending drives more marketplace volume. That flywheel has been compounding for years in markets where the incumbent banking system was weak enough to be displaced entirely. Scoring a 90 out of our internal rating system, it is one of the more unusual high-conviction names on our radar. The geographic risk is real and not trivial. But for an investor willing to do the work on Kazakhstan's macro picture, KSPI is worth a hard look.

The Contrarian Take

Golden Dome will not save Lockheed Martin

The consensus trade on Golden Dome is to buy Lockheed Martin (NYSE:LMT) because it sits at the centre of missile defence architecture and has decades of Pentagon incumbency. That logic is not wrong. It is just early and probably late at the same time. LMT is a $140 billion company. Even a $50 billion Golden Dome program spread over a decade moves the revenue line by a few percent annually. The contractors who benefit disproportionately from a new defence build cycle are not the ones already priced as defence anchors. They are the companies whose existing revenue base is small enough that a single large program changes the shape of the business entirely. Kratos Defense (NASDAQ:KTOS), with its autonomous systems and drone platforms, fits the Pentagon's stated direction on lower-cost distributed defence better than any legacy prime does. If Golden Dome broadens into a full next-generation layered shield, the relative repricing in KTOS versus LMT is where the asymmetry actually lives.