Episode Details
Back to EpisodesMonolithic Power (MPWR): Record Quarter, +164% AI Revenue — and Cash Flow Went Backwards
Published 1 month ago
Description
Monolithic Power Systems, Inc. (MPWR) Q2 2026 — Monolithic Power Systems reported Q2 2026 after the close on July 30. Revenue was a record $980.6M, +21.9% q/q and +47.6% y/y. Non-GAAP EPS was $6.50, about 10.5% above consensus; GAAP diluted EPS was $5.22. GAAP gross margin 55.2%, non-GAAP 55.6% (UP 10 bps q/q). Non-GAAP operating income $367.7M, a 37.5% margin. Q3 guidance of $1.14–1.16B came in roughly 16% above the Street, and the board added $500M to the buyback (now $1B authorized). The stock closed at $1,316.18 and rose about 10% after hours. But Enterprise Data (AI/server power) supplied nearly all the growth — +164.3% y/y to $380.6M, now 38.8% of revenue — while the other five end markets grew a combined 15.3%. And operating cash flow was $227.9M, BELOW the $237.6M generated a year ago on a third less revenue. Our owner-earnings DCF lands near $900. Our call: HOLD, 2/5.
The headline is spectacular and it deserves to be. What the headline hides is where the growth came from and where the cash went. Enterprise Data — power management for AI accelerators and servers — went from $144.0M to $380.6M, up 164.3%. Strip that single end market out and the other five combined went from $520.6M to $600.0M: 15.3% growth. Storage & Computing grew 2.3%. Consumer shrank 4.8%. In four quarters, one end market went from 21.7% of MPWR to 38.8%. Meanwhile operating cash flow fell year over year — $227.9M against $237.6M — as receivables rose 76%, inventory rose 38% to $675.8M, and capex climbed to fund a capacity goal management just pushed beyond $6B. Cash conversion went from 35.8% of revenue to 23.2%. Three things almost nobody covered: (1) ex-Enterprise Data, MPWR is a mid-teens grower, not a 48% grower; (2) the standard 'hyperscale mix will crush gross margin' bear case was falsified this quarter — non-GAAP gross margin went UP as AI mix jumped six points, and Q3 is guided to 55.4–56.0%; (3) the earnings are real but they are not yet cash, and at ~$65B of market cap the price already embeds roughly 26% annual free-cash-flow growth for a decade.
THE CALL: HOLD (2/5, A SUPERB BUSINESS AT A PRICE THAT NEEDS A DECADE TO GO RIGHT) — base-case value ~$900.0 vs ~$1316.18 today.
KEY METRICS:
- Revenue $980.6M, +21.9% q/q and +47.6% y/y — a company record
- Non-GAAP diluted EPS $6.50 (beat by ~10.5%); GAAP diluted EPS $5.22
- GAAP gross margin 55.2% (−10 bps q/q); non-GAAP gross margin 55.6% (+10 bps q/q)
- Non-GAAP operating income $367.7M — 37.5% margin, +2.7 pts y/y; GAAP operating income $303.9M
- GAAP opex $237.2M / non-GAAP opex $177.6M; stock comp $53.5M (incl. ~$1.8M in COGS)
- End markets: Enterprise Data $380.6M +164.3% | Storage & Computing $199.8M +2.3% | Automotive $157.1M +8.2% | Communications $131.5M +78.3% | Consumer $56.8M −4.8% | Industrial $54.8M +17.3%
- Enterprise Data is 38.8% of revenue, up from 32.7% in Q1 2026 and 21.7% in Q2 2025
- Ex-Enterprise Data revenue: $600.0M vs $520.6M a year ago (+15.3%)
- Operating cash flow $227.9M vs $250.3M in Q1 2026 and $237.6M in Q2 2025 — cash conversion 35.8% -> 23.2% of revenue
- Cash + short-term investments $1,413.8M, no debt; AR $343.6M (32 DSO); inventory $675.8M (140 days, 121 on next-quarter revenue)
- Q3 2026 guide: revenue $1,140–1,160M; non-GAAP GM 55.4–56.0%; non-GAAP opex $201.2–205.2M; 15% non-GAAP tax; 49.1–49.5M diluted shares
- Board added $500M to buyback ($1B authorized, ~1.5% of market cap); FY2025 actual repurchases were under $8M
- Dividend raised to $2.00/quarter (from $1.56) — ~$394M/yr, ~0.6% yield
- Capacity goal extended 'significantly beyond $6B'; initial DDR5 memory-component orders; sampling high-voltage AC-to-DC for 800V data centers; >1,500 new automotive sockets shipped YTD
What to watch: What we're watching: Enterprise Data holding this growth rate into 2027 with revenue tracking toward ~$7B; and — more important — operating cash flow re-converging with net income as the working-capital build normalizes. Either would move us more bullish. The risks: