Monolithic Power Systems FY2026 Form 10-K -- Does the Deferred-Tax Material Weakness Get Remediated Before the AI-Power Premium Gets Tested a Third Time
The Call
MPWR's FY2026 Form 10-K resolves two questions the market has priced as though already settled at a ~40x forward multiple, and both resolve in the same filing. First, whether Q2's Enterprise Data re-acceleration (+164.3% YoY, 38.8% of revenue) is a durable content win or an order pull-forward -- testable only through the Q3 and Q4 2026 revenue prints that land before the 10-K, and complicated by the fact that competitors (onsemi's named Great Wall/AWS design wins) are now disclosing exactly the customer-level evidence MPWR has never provided. Second, whether Ernst & Young's annual SOX 404(b) attestation finally certifies the deferred-tax material weakness as remediated after two consecutive year-end 'not effective' findings -- a binary auditor judgment nobody outside the audit committee can currently observe, sitting alongside a securities class action whose key claims survived a May 2026 motion to dismiss. The financial ladder below is a genuine point estimate on the first question; the governance question is deliberately kept out of the ladder and carried instead in the predictions and breaksIf, because a compliance opinion is not a line item and dressing it up as one would misstate what is actually known six months out.
| Metric | Our Estimate | Consensus |
|---|---|---|
| RevenueGAAP H1 2026 actual $1,784.8M + Q3 guidance midpoint $1,150M + Q4 estimate $1,240M. About +2.3% above the ~$4.08B consensus compiled from 14 analysts shortly after the Q2 print -- the gap is entirely in this report's Q4 assumption, since Q3 uses the same company guidance any consensus estimate would. | 4175 USD M | 4080 USD M |
| Gross profitGAAP Revenue x 55.2% anchored gross margin. | 2305 USD M | — |
| Gross marginGAAP Flat versus the 55.08-55.40% band held across all of Q1 2025-Q2 2026. The company's own bear trigger is two consecutive quarters below 54%; this estimate carries no margin cushion built in beyond the historical band itself. | 55.2% | — |
| Operating profitGAAP H1 2026 actual $545.0M + Q3 estimate ($364.1M, Q2 actual plus 35.6% incremental margin on the guided revenue step) + Q4 estimate ($391.1M, Q3 estimate plus 30.0% incremental margin on the estimated revenue step). | 1300 USD M | — |
| Operating marginGAAP Versus 29.99% (Q1 2026) and 30.99% (Q2 2026) actual -- continues the modest operating-leverage improvement seen in H1 without assuming it accelerates further. | 31.1% | — |
| Net incomeGAAP H1 2026 actual $450.5M + Q3 estimate ($305.9M at a 26.6% net margin) + Q4 estimate ($332.3M at a 26.8% net margin). This is the line the deferred-tax control weakness most directly threatens: a repeat of the FY2024-style adjustment would move this figure and diluted EPS without moving revenue, gross profit, or operating profit at all. | 1089 USD M | — |
| Diluted EPSGAAP Q1 $3.92 actual + Q2 $5.22 actual + Q3 ~$6.23 estimate + Q4 ~$6.78 estimate. About -1.7% below the ~$22.53 GAAP consensus compiled from 14 analysts shortly after the Q2 print -- close enough to cross-check this ladder's bottom-up construction, not close enough to treat as the same number. The market and the company's own guidance quote non-GAAP EPS (thesis base case $27.50 FY2026E); see businessMetrics. | 22.15 USD | 22.53 USD |
- → The FY2026 auditor ICFR attestation finds internal controls NOT effective for a third consecutive year-end -- the single largest identified risk to the current ~40-55x multiple, since it would confirm the governance overhang is structural rather than a one-time correction.
- → Enterprise Data revenue declines sequentially from the $380.6M Q2 2026 level in Q3 or Q4 -- would indicate the Q2 surge was partly order pull-forward rather than a durable content win, and would invalidate this ladder's Q4 revenue and margin assumptions together.
- → A second restatement or an additional material weakness is disclosed in the FY2026 10-K, beyond the deferred-tax item already known.
- → The Waterford Twp. securities litigation is expanded (amended complaint, new class period, or a disclosed SEC inquiry) rather than narrowed or settled before the 10-K files.
- → GAAP gross margin falls below 54% for two consecutive quarters -- the company's own explicit bear trigger, and this ladder carries no cushion against it.
- → Two or more of onsemi, TI, ADI, or VICR disclose 800V/48V AI-power production revenue or additional named design wins while MPWR remains at the sampling stage for its own high-voltage AC/DC products.
- → FY2026 revenue lands materially outside the $4,080-4,175M band this report and the pre-print consensus separately imply -- since both numbers share the same Q1-Q3 inputs, a wide miss on either side would mean this report's Q4 assumption (or the guided Q3 figure itself) was wrong, not merely imprecise.
Behaviour model, not arithmetic, and deliberately low confidence. MPWR's own Q4-to-Q1 step has been positive both of the last two years (Q4 2024 $621.7M -> Q1 2025 $637.6M, +2.6%; Q4 2025 $751.2M -> Q1 2026 $804.2M, +7.1%), unlike the flat-to-down seasonal step common among more cyclical analog names -- this call extrapolates that pattern rather than assuming a reversion. This pipeline's guidance calls have already come in low twice this cycle on other issuers (STX FQ1 FY2027 -9.3%, AMD Q3 2026 -3.1%), both understating management in an up-cycle; if MPWR's AI-power ramp is still accelerating rather than merely continuing, this call is more likely to be a floor than a center.
Evidence (7)
Positioning Suggestions
→ This is a two-headed event, not a single earnings print: the financial ladder above resolves on the Q4 2026 number, but the multiple the market pays for that number depends on a separate, non-arithmetic auditor judgment (the ICFR attestation) this report cannot forecast. Size any pre-event exposure with that asymmetry in mind rather than treating a clean revenue beat as sufficient for a re-rate.
→ Watch onsemi, TI, ADI, and VICR disclosures between now and the print for named 800V design wins or production revenue -- competitors closing the customer-disclosure gap MPWR has never closed is a more informative signal of durability risk than MPWR's own revenue line, which can beat even while share erodes at the margin.
→ The stock has already de-rated about 7% from the thesis's last anchor on broad chip-sector weakness rather than any MPWR-specific news found this cycle -- the entry point is more favorable than it was in early August, independent of how the event itself resolves.
→ Do not treat the June-July 2026 plaintiff-firm investigation announcements (Kahn Swick & Foti, Schall) as new information distinct from the already-priced Waterford Twp. litigation -- they recycle the same underlying narrative and should not independently move a position.
Predictions
11MPWR's FY2026 Form 10-K states that internal control over financial reporting was effective as of December 31, 2026
An effective finding would remove the single largest identified overhang on the current multiple and argues for adding into any weakness into the filing; a third consecutive not-effective finding argues for treating the governance discount as structural rather than transitory.
MPWR discloses a named AI/hyperscaler end customer, or dollar-denominated 800V production revenue, in any disclosure before the FY2026 10-K files
Enterprise Data quarterly revenue exceeds $380.6M (the Q2 2026 level) in both Q3 and Q4 2026
FY2026 total revenue exceeds $4,175M (this report's point estimate)
FY2026 GAAP diluted EPS exceeds $22.15 (this report's point estimate)
Q3 and Q4 2026 GAAP gross margin both remain at or above 54%
At least two of onsemi, TI, ADI, or VICR disclose a named 800V/48V AI-power production design win or production revenue before MPWR's FY2026 10-K files, while MPWR remains at the sampling stage for its own high-voltage AC/DC products
Onsemi has already disclosed named wins (Great Wall, AWS) and TI has a complete reference architecture as of this report -- if MPWR is still the outlier with no named disclosure by the 10-K, treat that gap as the more decision-relevant durability signal than the revenue line itself.
The Waterford Twp. securities class action reaches a settlement or a further dispositive ruling beyond the May 2026 partial motion-to-dismiss denial before the FY2026 10-K files
A second restatement or a new material weakness, beyond the disclosed deferred-tax item, is disclosed in the FY2026 10-K
MPWR stock trades within 15% of its 2026-08-26 close ($1,303.88) on the date the FY2026 10-K is filed
Q1 FY2027 revenue guidance, issued alongside Q4/FY2026 results, exceeds $1,240M (this report's Q4 2026 point estimate), i.e. implies continued sequential growth rather than a seasonal dip
Key Questions
- Does the FY2026 10-K state that internal control over financial reporting was effective as of December 31, 2026, or does it carry a third consecutive year of a disclosed material weakness?
- Does Enterprise Data revenue hold above the $380.6M Q2 2026 level through both Q3 and Q4, and does management give any commentary distinguishing durable content wins from order timing?
- Does MPWR disclose a named AI/hyperscaler end customer or any dollar-denominated 800V production revenue, closing the disclosure gap that onsemi's named Great Wall/AWS wins have already opened?
- Where does the Waterford Twp. securities litigation stand by the 10-K filing date -- expanded, settled, or still pending further motion practice?
- Does GAAP gross margin hold at or above 54% through both remaining 2026 quarters?
- What does the initial Q1 FY2027 guidance, issued alongside the Q4/FY2026 release, say about whether the AI-power ramp is still accelerating or merely continuing?
- Does distributor concentration or inventory-day commentary show any sign of channel stuffing behind the Enterprise Data growth number?