TE Connectivity Grew Up, Grew Record Books, and Took Away the Number Wall Street Was Counting
A record $20 billion year, a $8 billion backlog — and a quiet retirement of the AI revenue target in favor of a $1 billion-a-year annuity, plus a brand-new $500-per-robot humanoid claim
TEL · Earnings Call · 2026-10-28
The scoreboard versus the stock
TE Connectivity closed fiscal 2026 with records on nearly every line the company tracks: sales near $20 billion (+15% reported, +12% organic), orders above $22 billion, backlog of roughly $8 billion up 30% year-over-year, adjusted EPS of $11.52 (+24%), and adjusted operating margins of 22%, up 120 basis points. The fourth quarter alone printed $5.3 billion of sales (+13% organic), $6.1 billion of orders (+29%, with the Industrial segment up 47%), and a book-to-bill of 1.14. Guidance for the December quarter is $5.35 billion and roughly $3.30 of adjusted EPS, both up about 15% and 21% respectively, and management framed fiscal 2027 organic growth as similar to the 12% just delivered.
And yet the shares sit 11.1% below their November 2025 peak of $249. That gap is the whole point. This is a company whose fundamentals accelerated while its multiple contracted — price to operating income has fallen to about 15.5x from north of 20x a year ago, even as the underlying earnings line climbed. Sell-side analyst Scott Davis put it with unusual bluntness on the call: “Despite the stock having fits and starts, you guys did your job.” — Scott Davis, Analyst · 2026-10-28 That is the puzzle worth solving.
What is genuinely new: a robot line item
The freshest thing in this transcript is not AI data centers — it is a humanoid. Terrence Curtin spent a full passage sizing the opportunity, and for the first time put a number on it:
We view these technologies represent up to $500 of potential TE content per humanoid robot. We are also working closely with leading OEMs on next-generation designs, positioning us to benefit as this emerging market scales.
The logic is elegant: a humanoid needs a battery, a charge inlet, a controller, actuators, motors and harsh-environment connectivity — "that's a car in some cases," as Curtin put it, which is precisely why the automotive franchise is the wedge. The humanoid robot theme is brand new to this company's keyword record, appearing at the top of the current-quarter list after never surfacing at scale before, alongside the related liquid cooling and direct current language that describes the data center side of the same power problem.
But note the honesty that came with the hype. Asked by TD Cowen's Joe Giordano where the volumes actually are, Curtin declined to pretend: “The wild card is the units... You go out five years, you have some people that think there's 1 million humanoid robots being made. You have some people thinking 10 million humanoid robots made.” — Terrence Curtin, Chief Executive Officer · 2026-10-28 A $500 content number multiplied by an unknown unit number is a call option, not a forecast. Management labeled it as such.
What quietly disappeared: the AI dollar target
The more consequential change is an absence. For four straight quarters, TE sold investors a dollar funnel: $800 million of AI revenue, then $900 million, then "approaching $2.4 billion" for the year, then a $3 billion target at the 2025 investor day that kept being pulled forward. On the July call, Curtin was still explicit: “Nothing has changed on that $3 billion. Continues to shift left, and we'll be above that.” — Terrence Curtin, Chief Executive Officer · 2026-07-22
That number is gone. In its place, a rate rather than a level: “We see the copper addressable market growing over 100%.” — Terrence Curtin, Chief Executive Officer · 2026-10-28 — no, more precisely, the replacement framing is that DDN adds over $1 billion of revenue in fiscal 2027 and grows "well over $1 billion per year over an extended time horizon, implying that this business will double again in the next few years." The new top keyword for the company is DDN revenue, and the old one, AI revenue, was among the sharpest keyword decliners of the quarter. This is not an accident. Curtin had pre-announced the shift in July: “this is where in this space it blurs a lot, and it depends on where the customer is that's doing it.” — Terrence Curtin, Chief Executive Officer · 2026-07-22
There are two ways to read that. Generously: a $3 billion target that keeps getting beaten creates a ceiling question every quarter ("what's next?"), while a $1 billion-a-year annuity with a doubling horizon is un-ceilinged and harder to arbitrage. Cynically: an AI number that could be benchmarked against a competitor's AI number is being retired in favor of a segment number that cannot. Either way, the investor's scorecard just changed without anyone voting on it.
Copper, fiber, and the fourth asking of the same question
The debate that will not die is whether optics inside the rack eats TE's lunch. Amit Daryanani asked it again this quarter; Mark Delaney and Shreyas Patil circled the same territory. The answer is the answer of a company that has decided the trade-off is a tailwind either way: roughly 30% of DDN revenue is power connectivity, which is architecture-agnostic and expected to grow faster than DDN overall, and the copper TAM itself is forecast to more than double as inference and agentic workloads multiply CPU-heavy connections. That has been the script since April, when Curtin insisted “it's not copper or optical, it's copper and optical” — Terrence Curtin, Chief Executive Officer · 2026-04-22 — a line he has now repeated roughly once a quarter, which is itself the tell that the market does not yet believe it. The new entrant to that argument is fiber attach, where the RAM Photonics validation work with hyperscalers is real but, by management's own admission, not a 2027 revenue event.
Meanwhile the tape has already voted elsewhere. The broader AI data center basket is down about 5.4% over 90 days across 50-odd tickers, with co-packaged optics, high-bandwidth memory and clean-room names all in drawdown. TE is up 9% over the same window. That relative strength — supplier to the buildout rather than pure play on it — is the market quietly endorsing the power-and-connector hedge over the optics story.
Tariffs, inflation, and a balance sheet that moved
One small oddity worth flagging: TE's fourth quarter included “IEEPA tariff refunds that resulted in a reduction in sales of $87 million” — Heath Mitts, Chief Financial Officer · 2026-10-28, with another $50 million baked into the December guide. Refunds as a revenue headwind is the opposite of the tariff-refund bonanza other reporters have been celebrating this season — TE is passing the money back to customers, so it costs top line and nothing on earnings. Underneath that, the inflation story is unchanged and unrelieved: “We're still feeling heightened inflationary pressures, particularly on resins, metals, and certainly freight cost” — Heath Mitts, Chief Financial Officer · 2026-10-28. Gross margin held at 35.6% despite it, which is the case for owning the productivity story rather than the pricing story.
What has changed is the balance sheet. Effective net cash has swung from roughly -$0.7 billion to -$4.3 billion as TE committed about $1.6 billion to Industrial-segment bolt-ons (RAM Photonics, plus the pending ~$1.4 billion Astrodyne TDI) while returning $2.8 billion to shareholders. Interest coverage above 30x means this is a choice, not a constraint — but it does mean the buyback that has absorbed so much of the record free cash flow is now competing with acquisition math rather than running unopposed.
The bottom line
TE just delivered the best year in its history and guided to another double-digit organic year, and the stock is 11% off its high. The company's answer is to stop letting investors mark it against an AI number and start marking it against a compounding DDN annuity; the market's answer, so far, is a modest re-rating in a de-rating sector. The humanoid robot content number is the optionality nobody is paying for yet. What to watch is whether the $1 billion-a-year DDN cadence holds through a second half that carries more than half of the fiscal 2027 growth — because that is now the number that matters, and it is a harder one to beat than the one they retired.