capital allocation: $67b+ returned annually — but sbc dilutes the math
Microsoft returns enormous capital to shareholders through buybacks ($32B) and dividends ($22B), but stock-based compensation at 17.2% of revenue undermines net shareholder returns. For a 45x earnings company, the dilution is particularly destructive — every 1% dilution costs shareholders ~$30B in market cap.
total shareholder returns
~$54B
fy2025 (buybacks + dividends)
sbc / revenue
17.2%
extraordinarily high for mega-cap
net buyback yield
~0.5%
after sbc offset
dividend yield
0.72%
below 10y treasury
Exhibit: Capital Allocation Framework
| metric |
fy2023 |
fy2024 |
fy2025 |
| share buybacks |
$22.2b |
$17.6b |
~$32b |
| dividends |
$20.2b |
$21.8b |
~$22b |
| sbc (dilution) |
$9.6b |
$11.1b |
~$14b |
| net return to shareholders |
$32.8b |
$28.3b |
~$40b |
| capex |
$28.1b |
$44.5b |
~$55b |
Key concern: Capex is accelerating dramatically ($28B → $55B) driven by AI data center buildout. This investment must generate returns within 3–5 years or free cash flow conversion will deteriorate. At 45x earnings, any FCF compression hits the stock disproportionately.
Microsoft's capital allocation is competent but irrelevant at this valuation. The $54B annual return represents just 1.8% of market cap — insufficient to support the share price if multiple compression occurs. Contrast with Apple, which returns 3.5%+ at a lower P/E. The SBC problem is structural: $14B annually means Microsoft must buy back $14B just to stand still on share count.