Skyworks Just Traded Its Best Socket for a Portfolio — and a Promise of 55% Margins

The Qorvo deal closed 28 days before this call. Management's first combined earnings report sells diversification, defense and physical AI; the tape has already paid for a gross margin the numbers haven't earned yet.

SWKS · Earnings Call · 2026-11-03

A company that changed shape three weeks before the print

On October 5, Skyworks closed its combination with Qorvo — the largest deal in its history and, as CEO Phil Brace put it, [one of] the “biggest transformation deals in the RF industry in general” — Philip Brace, Chief Executive Officer and President · 2025-11-04. That timing matters enormously for reading this report. The September quarter numbers are pure stand-alone Skyworks: $1.046 billion of revenue, 44.4% gross margin, $1.28 of non-GAAP EPS. But the December guide — $1.8 to $1.9 billion — is the first combined number, and from the February call onward the old Skyworks/Qorvo split disappears entirely, replaced by two new reporting buckets: smartphone and diversified growth markets. The pro forma picture management wants you to hold in your head is this: smartphone roughly $4.6 billion in fiscal 2026, down about 11%; diversified growth markets roughly $2.9 billion, up about 12% at above-corporate-average gross margins. That is the entire investment case in two numbers — a shrinking, hyper-competitive legacy base and a growing, higher-margin bundle stapled onto it.

The tape has already paid for the pitch

SWKS is up 26.1% over the last 90 trading days and sits about 6.6% below its September 23 high, even though the full history still shows a 57% drawdown from the 2021 peak of $199.66. At a $12.85 billion market cap and roughly 2.5x revenue, the stock is being valued on the pro forma entity, not on the stub that actually reported. That is worth dwelling on, because the last filed financial shape is not pretty. Stand-alone Skyworks closed fiscal Q3 with revenue of $935 million, down 3% year-over-year, operating income of $49 million, down 56%, and gross margin of 40.1% — versus a 50.6% peak a decade ago. Free cash flow after stock comp was negative, and effective net cash fell 69% year-over-year to $350 million after roughly $2 billion of acquisition financing was raised ahead of closing. The bull case requires buying the widening margin arc, not the trailing one.

Four pillars, three of them genuinely new

What is new this quarter is the vocabulary. Brace laid out diversified growth markets as four pillars: physical AI and connected edge at more than $1.6 billion, defense and aerospace at roughly $500 million, data center networking infrastructure at roughly $400 million, and automotive at roughly $340 million. For a company whose keyword history for the past two years was dominated by RF content, internal modems and Android sockets, defense and defense-adjacent GaN is a structurally new leg — and it comes with a Category 1A trusted foundry accreditation that almost nobody in RF holds. The data center claim is the boldest: “our data center business grew, I think, 90% year-over-year from that perspective” — Phil Brace, Chief Executive Officer and President · 2026-11-03. And the constraint is not demand: “all of those businesses are supply constrained and the demand far exceeds our ability to supply them” — Phil Brace, Chief Executive Officer and President · 2026-11-03. Note the convergence with the global reporter set — "Physical AI" showed up in the market keyword tape and in peer calls this season, and power isolation is Skyworks' own distinctive angle on the AI rack: isolating grid-level voltage from the silicon, not selling compute. That distinction is the contrarian bit. The global tape over 30 days is brutal on the AI data center complex — the AI data center cluster is a heavy decliner with far more negative tickers than positive, and high bandwidth memory is down over 30%. Skyworks sells timing and power components into that buildout, not HBM or GPUs, so its demand curve is one derivative removed from the part of the trade being de-rated. Worth flagging as a genuine divergence rather than consensus.

For the first time in years or as we start to see, we've got a tailwind of content coming with more transmit bands, more received bands, more satellite bands, more power devices. So we see a tailwind of content, more filters that we haven't seen in the past.
Phil Brace, Chief Executive Officer and President · 2026-11-03

What is being quietly retired

The most instructive part of the call is what shrank. The Android market — a keyword that spiked hard this quarter, and not in a good way — is being actively de-emphasized. Qorvo's Android business fell more than 20% sequentially, and CFO Philip Carter guided that it “should represent less than 10% of our overall kind of smartphone business” — Philip Carter, Chief Financial Officer and Senior Vice President · 2026-11-03 going forward. Brace was almost relieved about it, calling the mix shift “a richer mix of business. So I don't mind the switch” — Phil Brace, Chief Executive Officer and President · 2026-11-03. Also retired: the dividend, eliminated last quarter, with capital now pointed at buybacks and deleveraging — $107 million repurchased in September at an average of $66.27 — plus an investment-grade upgrade to Baa3 from Moody's. And retired, at least rhetorically, is the lower end Android chase that used to define the bear case. The recurring tension is the largest customer. A year ago the worry was a competitor's long-term supply agreement; Brace answered then that “our position is earned design win by design win, platform by platform” — Philip Brace, Chief Executive Officer and President · 2026-07-28. Asked again this week, the answer barely moved: “that business is always going to be a hypercompetitive business” — Phil Brace, Chief Executive Officer and President · 2026-11-03. The new argument is that breadth reduces variance — more sockets per device means less RF signal chain single-socket volatility.

The 40-to-55 arithmetic problem

Everything above is strategy. The valuation rests on the margin bridge, and the bridge is steep. Management's 24-to-36 month target is 50% to 55% gross margin, 30% to 35% operating margin and 35% to 40% adjusted EBITDA margin — from a 40.1% gross margin base at a company that spends heavily on R&D, running R&D at 22% of revenue. The mechanics are mix, factory consolidation, and pricing. Carter was candid that cost pass-through is asymmetric: “In the diversified space, we're able to pass that on more quickly... In the kind of handset smartphone space, it's a little bit more challenging because those are those are kind of annual contracts” — Philip Carter, Chief Financial Officer and Senior Vice President · 2026-11-03. That is the same complaint as three months ago, when Brace said “we don't have the ability to adjust when input costs go up as well” — Philip Brace, Chief Executive Officer and President · 2026-07-28 — and two quarters back, “some of the long-life products allow us to increase price and pass those costs on” — Philip Carter, Chief Financial Officer and Senior Vice President · 2026-05-05. Input cost inflation has now been a stated margin headwind for at least four consecutive calls; it is not a new excuse, it is a chronic condition. Synergy math: $150 million of opex savings inside fiscal 2027, exiting at a $250-to-$300 million annualized run rate, toward a $500 million-plus total split two-thirds opex, one-third COGS. The opex half is credible and fast — HR systems integrated, ERP next. The COGS half is where the gross margin lives, and management effectively pushed it into fiscal 2028. So the story is this: a company that spent years defending sockets and watching content erode has bought itself a second act — defense, physical AI, data center power, automotive — and a stock that is already up 26% in three months on the announcement. What has not yet been proven is the arithmetic. The tape has priced the portfolio; the income statement still has to deliver the margin.