Jabil’s AI ramp is no longer just revenue: Q4 showed the cash-return model can absorb the build
JABIL INC beat because Intelligent Infrastructure pulled demand forward without breaking gross margin, and the market may still be valuing the AI exposure as low-quality EMS revenue rather than as a capital-return flywheel. The variant view is that FY26 guidance makes the AI build investable: management is committing to about $31.3 billion of revenue, $11 of core EPS, and greater than $1.3 billion of free cash flow while still executing a $1 billion repurchase authorization.
JABIL INC printed the kind of quarter that forces a rethink of the bear case: AI-related revenue is scaling, but the company did not pay for that growth with a visible collapse in margin or a pause in capital return. What was priced in was a normalizing EMS story: the Street had $7,588.5 million of revenue and $2.92 of EPS embedded for Q4. What actually surprised was not just the size of the beat, with $8,252.0 million of revenue and $3.29 of EPS, but the mix of the beat: Intelligent Infrastructure was $400 million above expectations, gross margin returned to 9.5%, and management framed FY26 around cash generation rather than capacity excuses. The market may be missing that the AI ramp is now large enough to reshape the income statement, while the buyback and free-cash-flow commitments keep per-share outcomes from being diluted by the working-capital and capacity demands that usually make EMS AI ramps less valuable.
That distinction matters because the headline surprise was too large to dismiss as a one-quarter inventory catch-up. The revenue surprise was +8.7%, while the EPS surprise was +12.7%, so incremental demand translated into earnings rather than being fully consumed by launch costs. On the company’s own reported basis, Gregory Hebard said Q4 reached “approximately $8.3 billion in revenue, which exceeded the midpoint of our guidance by roughly $800 million,” a wording that matters because it anchors the upside to management’s prior range, not only to sell-side estimates. The company also posted diluted EPS of $1.99 on a GAAP basis, which should be kept separate from the Street-comparison EPS of $3.29 because they are different reporting bases. The investment conclusion is still the same across both bases: revenue upside was real, and the operating model did not visibly buckle under it.
The financial trajectory reinforces that interpretation because revenue is no longer stuck in the post-divestiture trough that defined FY24 and early FY25. Revenue spent several quarters clustered around the high-$6 billion range after the FY24 reset, then moved back to $8,252.0 million in Q4 FY2025, while gross margin recovered to 9.5%. That matters more than the single-quarter beat because Jabil’s historical concern has been that large-volume programs produce revenue without enough margin density. Q4 cuts against that concern: gross margin matched the prior-cycle high in the data pack even as Intelligent Infrastructure carried the surprise. If the market came in pricing revenue recovery as mechanically lower-margin, the print argues for a higher-quality recovery.
The capacity story explains why the margin guide is the right battleground for the stock rather than the Q4 beat itself. Management is not claiming that FY26 is free of absorption risk; Michael Meheryar Dastoor explicitly expects core operating margin to expand by roughly 20 basis points to around 5.6% “in spite of the underutilized capacity in multiple geographies.” That phrase is the useful part of the quote: it admits the cost drag while still committing to expansion. For portfolio managers, the variant perception is that underutilization is already inside the guide, so the next debate should be whether AI-related revenue reaches about $11.2 billion rather than whether launch costs exist. If AI revenue grows by roughly 25% in FY26 and corporate revenue grows approximately 5%, AI is doing the heavy lifting while other pieces are being managed for cash and mix.
That mix is visible in the segment color, where the upside is not evenly distributed. Regulated Industries produced $3.1 billion in Q4, but the cleanest surprise came from Intelligent Infrastructure at $3.7 billion and $400 million above expectations. Connected Living & Digital Commerce was $1.4 billion, only slightly ahead of the outlook from 90 days ago, and its Q1 setup is weaker at $1.29 billion with a 16% year-on-year decline. The second-order read is that Jabil’s growth is becoming less dependent on consumer electronics assembly cadence and more dependent on infrastructure programs, which should matter to how investors map customer risk. For Apple, the implication is not a direct warning on end demand, but it is a mix signal: the Jabil bucket most associated with electronics manufacturing and assembly is not where the incremental upside is being created, and the next-quarter guide puts that segment down 16% year-on-year.
The AI read-through is more consequential for competitors because Jabil is guiding a scale-up that can pressure capacity, pricing, and program allocation across the infrastructure supply chain. AI-related revenue rose from approximately $5 billion in FY24 to approximately $9 billion in FY25, and management expects about $11.2 billion in FY26. That is not merely a demand anecdote; it is a statement that AI-related work is approaching a scale where EMS providers must reserve capacity, manage geographic redundancy, and carry utilization risk before programs fully mature. The peer table shows why this matters competitively: Jabil’s Q4 revenue growth of +18.5% sits near the faster end of the packaging-adjacent peer set, while its 9.5% gross margin is structurally below the margins shown for names such as KYOCY at 29.0% and 4062.T at 29.5%. The point is not that Jabil should trade like a materials or packaging peer; it is that Jabil is capturing AI volume at EMS economics, and the stock’s upside depends on whether cash conversion and buybacks offset that lower margin ceiling.
Capital return is the bridge between those lower EMS margins and an investable equity story. Q4 net CapEx was $83 million, full-year net CapEx was $322 million, and full-year net CapEx was 1.1% of revenue. Those figures undercut the simple bear argument that AI growth necessarily consumes all available cash. More importantly, Jabil finished the prior $1 billion share repurchase authorization in Q4 and had a new $1 billion authorization approved in July. Hebard’s language was unusually firm on timing: “we intend to fully execute the current authorization in fiscal ’26.” That commitment matters because it converts the free-cash-flow guide into per-share support, and it gives PMs a concrete test for whether management is prioritizing shareholders while funding AI capacity.
The balance sheet gives that commitment credibility, though not unlimited freedom. Jabil ended FY25 with $4 billion of unused capacity under global credit facilities, and management said the framework is to return 80% of annual adjusted free cash flow to shareholders. Since FY13, shares outstanding fell from 203 million to 107 million, a 47% decline, which shows that repurchase execution is not a new narrative introduced to decorate an AI story. The risk is that investors over-capitalize that history if free cash flow misses the greater than $1.3 billion FY26 target. But the print’s important message is that the company is not choosing between AI capacity and capital return yet; it is guiding to both.
That is why the FY26 guide is more important than the Q1 revenue range. Q1 revenue is expected in the range of $7.7 billion to $8.3 billion, which gives management room to absorb normal launch timing and customer digestion after the Q4 upside. The segment guide says Intelligent Infrastructure revenue should be $3.67 billion and up approximately 47% year-over-year, while Regulated Industries should be $3.05 billion and up 3% year-on-year. Those two numbers define the debate: if infrastructure remains near Q4 scale while regulated holds modest growth, the market will have to treat FY26’s about $31.3 billion revenue guide as grounded rather than aspirational. If Connected Living & Digital Commerce remains pressured, the stock can still work because the segment is guided to $1.29 billion, making it less central to the upside case than the infrastructure ramp.
The call delivery supports that interpretation, with one caveat. In the tone history, Q4 FY2025 sentiment was 0.32 versus 0.31 in Q3 FY2025, so overall tone did not materially change. What changed was guidance_tone, rising to 0.49 from 0.33, while uncertainty fell to 44.3 from 61.7. That combination is useful: management did not sound broadly promotional, but it sounded more definitive when discussing the forward path and less uncertain overall. The caveat is that ai_optimism fell to 0.37 from 0.54, which conflicts with the magnitude of the AI revenue guide. The clean interpretation is not that management is less confident in AI demand; it is that the call framed AI as an operational and financial plan rather than a thematic pitch.
That distinction between thematic pitch and operating plan is the crux of the stock call. A market that prices Jabil as a generic contract manufacturer will focus on the 9.5% gross margin and conclude that AI volume is low-quality revenue. A market that prices the print correctly should focus on the combination of Intelligent Infrastructure at $3.7 billion, FY26 AI-related revenue expected at about $11.2 billion, and free cash flow expected to be greater than $1.3 billion. The first view sees rising complexity with limited margin. The second sees a company converting AI scale into repurchases and EPS while holding the long-term target of 6% plus core operating margins and north of $1.5 billion in adjusted free cash flow. Q4 does not prove the long-term target, but it makes the path more credible than it looked when revenue was pinned near the high-$6 billion range.
The main risk to the thesis is that FY26 guidance embeds a narrow execution corridor. Management expects approximately 5% revenue growth to about $31.3 billion, but Q1 revenue guidance spans $7.7 billion to $8.3 billion. If the company starts FY26 near the low end, investors will question whether Q4 pulled forward infrastructure demand or whether underutilized capacity is taking longer to absorb. Margin also needs to be watched because Q1 gross margin in the history table later sits below the Q4 FY2025 level, and the Q1 GAAP operating income range of $263 million to $343 million leaves room for a step down from Q4 GAAP operating income of $337 million. The bullish view survives a seasonal step down; it does not survive evidence that AI revenue growth requires a reset lower in operating leverage.
What to watch next is therefore specific. For Q1, the stock needs revenue in the upper half of the $7.7 billion to $8.3 billion range, with Intelligent Infrastructure close to the $3.67 billion guide and not merely offsetting a weaker Connected Living & Digital Commerce segment at $1.29 billion. For FY26, confirmation is management holding about $31.3 billion of revenue, $11 of core EPS, and greater than $1.3 billion of free cash flow while executing the $1 billion repurchase authorization. The thesis breaks if the next update cuts the roughly 25% AI-related revenue growth expectation, backs away from around 5.6% core operating margin, or reframes underutilized capacity as a larger drag than the current guide already absorbs. The next quarter is not about whether Q4 was good; it is about whether Q4 was the first clean evidence that Jabil’s AI scale can fund both capacity and shareholders.