Open in interactive viewer → charts, metric popovers & call review

Garmin Beat the Quarter and Pre-Announced the Margin Give-Back

Record revenue, a 63.3% gross margin and a raised EPS guide — paired with management's own admission that margins revert to 'historical range' once the cheap memory runs out.
GRMN · Earnings Call · 2026-10-28

A beat with an expiry date

Garmin's third quarter looked like another link in the chain of records the company has been forging: revenue up 12% to nearly $2 billion, gross margin of 63.3%, operating income up 34% to $612 million, pro forma EPS up 39% to $2.76. Full-year guidance went up across the board — revenue to $8.15 billion, operating margin to 29%, pro forma EPS to $10.80, a 26% increase over last year. And yet the stock sits 13% below its August 10 peak of $313.16. It has recovered 15.3% over the last 90 days, so this is not a broken chart — it is a chart that has already been told something. Cliff Pemble told you what, on the first page of his remarks:

While it is too early to offer a comprehensive outlook for 2027, we expect our consolidated gross margin profile will return to a historical range primarily due to higher memory costs.

Clifton Albert Pemble, President & CEO · 2026-10-28
That is a company guiding earnings up while quietly guiding its best-in-class margin down. The 63% print is not the new normal; the high-50s is.

Three tailwinds, and the fourth one nobody printed in the release

Management was transparent about what drove the quarter: favorable segment mix (auto OEM shrinking as a share of the total), lower product costs from a weaker Taiwan dollar, and $33 million of tariff refunds. Then came the honest sentence, the one that matters more than any of the three.

Our results have benefited from our safety stock of memory purchased at lower prices. New supplies are being purchased at current market prices. Which are significantly higher than last year.

Clifton Albert Pemble, President & CEO · 2026-10-28
The company's own keyword record traces the handoff cleanly. The tariff refund was Garmin's single hottest theme a quarter ago; it has since slid down the list as the cash has been banked — Douglas Boessen confirmed on the call that “we have received the majority of our refunds” — Douglas Gerard Boessen, Chief Financial Officer · 2026-10-28. In its place, memory cost has climbed to the top of the company's theme stack, and memory was the subject of roughly every analyst question for the first twenty minutes. This is not a Garmin-specific problem, which is precisely why it is investable information. High bandwidth memory is a 360-day global tape winner — up over 100% with twelve names participating — that has since rolled over, down roughly 7% over the last ninety days. The same AI buildout that has been a revenue gusher for DRAM and semicap names shows up on Garmin's P&L as an input cost, and component pressure is echoing across the supply chain: Seagate flagged external component cost increases on the same day, and the broader memory cost theme has been the most persistent global keyword of the past three quarters. When the shortage theme de-rates, as the AI data center complex has over 90 days, the buyer of memory stops getting the benefit of the doubt and starts getting the benefit of a lower multiple. Pemble widened the aperture beyond memory: “there is definitely second order effects driven by the memory situation and the overall AI build out. That creates both availability supply issues as well as some pricing issues in other areas” — Clifton Albert Pemble, President & CEO · 2026-10-28 — PCBs, cabling, the unglamorous connective tissue of a vertically integrated hardware business.

What is genuinely, durably accelerating

The demand side of this call is not borrowed. Fitness grew 16% to $694 million on broad wearable strength; Outdoor grew 20% to $597 million behind the Fenix 9 launch; Marine grew 16% to $310 million on LiveScope Plus; Aviation grew 5% to $253 million with a 78% gross margin. Management called out fitness as the strongest contributor to full-year consolidated growth. Two products matter beyond the quarter. The CIRQL smart ring is a new form factor that expands the addressable market rather than cannibalizing it: “We are still catching up on some of the initial demand, especially on CIRQL, although Phoenix is also very, very strong, and so we are working hard to fill that backlog of demand” — Clifton Albert Pemble, President & CEO · 2026-10-28. A sold-out launch is a nice problem; a sold-out launch in a screenless category is a new demand curve. And then the one that would actually fix the margin problem structurally: “Connect Plus has outgrown the general rates of the underlying segments where those products can benefit from Connect Plus.” — Clifton Albert Pemble, President & CEO · 2026-10-28 Renewals after free trials have been strong, and the company has a roadmap to push more features behind the paywall. A Connect Plus subscriber is a higher-margin dollar than a hardware unit — but Garmin still will not break out the revenue, so the market cannot underwrite it yet. The other quiet positive: new user growth skews heavily new-to-Garmin outside the flagship, which is the difference between a refresh cycle and a land grab.

The balance sheet is paying for the defense

Here is the second cost of the memory problem, and it does not show up in gross margin. Third-quarter free cash flow was $293 million, down $132 million year over year, because inventory climbed to roughly $2.3 billion and Garmin began making prepayments to lock in supply. Boessen described the agreements as negotiated one-by-one with memory suppliers. Full-year free cash flow guidance was cut to about $1.1 billion, and capital expenditure jumped 180% year over year to $128 million in the quarter, with roughly $500 million expected for the year and the Thailand facility — the one that can eventually double capacity — starting production in the first quarter of 2027. This is the company's standard playbook: use the balance sheet as a shock absorber. The free cash flow line is the tell that the absorber is being drawn down. Meanwhile the structural defense remains what it has always been — “our overall margin structure is higher, and that's because we're a vertically integrated company” — Clifton Pemble, President and Chief Executive Officer · 2026-02-18 — and the operating leverage is real, with the last filed quarter showing gross margin of 62.4%, up 3.6 points year over year.

The trough nobody is paying for, and the multiple nobody is paying

Auto OEM fell 20% to $132 million as domain controllers rolled off, with the loss narrowing to $12 million as R&D shifted elsewhere. The Mercedes program ramps early in 2027 and is, by management's own framing, mix-unfavorable — lower-margin auto revenue growing faster than the consolidated base is a second headwind stacked on memory. That is the part of the 2027 bridge the sell-side will underwrite last. What is the tape saying? Price to free cash flow sits at 32.1x, down 22% year over year, and the earnings multiple has compressed from a 30x peak to roughly 24x. Nobody is paying peak multiples for a margin that management just said will revert. A quarter ago Pemble was defiant: “it is never our strategy to go backwards” — Clifton Albert Pemble, President & CEO · 2026-07-29. Today the guidance is that gross margin does go backwards, and the mitigations — pricing, vertical integration, selective sourcing — are described as a toolbox rather than a promise.

The read

The tension worth holding onto is this: Garmin just delivered its best margin in years, raised guidance, and told investors in the same breath that both are temporary. The demand evidence — Fenix 9, CIRQL, LiveScope, supply constraint rather than demand constraint, and a subscription layer growing faster than the base — is genuinely strong. The cost evidence is equally clear, and it is arithmetic, not sentiment. Watch two things in the fourth quarter: whether memory lands where they said it would, and whether Connect Plus finally gets a number attached. If it does, the margin floor in 2028 is higher than the 2027 trough implies — and the current drawdown will have been the discount, not the warning.