MPWR’s print was not about the 1% beat, it was about whether $4 billion of capacity can turn AI power wins into 2026 revenue

The print means MPWR’s AI power cycle is moving from narrative to allocation math, and that is the variant perception in this event. What was priced in was a clean beat: the Street was at $742.4 million of revenue and $4.74 of EPS, and the company delivered $751.2 million and $4.79, surprises of +1.2% and +1.1%. That is not enough, by itself, to re-underwrite the stock. What actually surprised was the shift in management’s framing of growth capacity and enterprise data: Tony Balow said MPWR “achieved our milestone of securing more than $4 billion of geographically balanced capacity,” and the call also carried a prior outlook escalation from “between 30-40%” to “a floor of 50% growth for 2025.” That language matters because it turns a conventional analog-cycle question into a supply assurance question. If PMs treat this as merely another above-consensus quarter, they risk missing the more important underwriting point: MPWR is trying to reserve enough manufacturing headroom to keep power modules attached to accelerated-compute platforms without letting gross margin fall outside the stated 55-60% model.

That distinction between a small reported beat and a larger capacity signal is critical because the financial trajectory has already normalized around a higher base. Revenue rose from $621.7 million in Q4 FY2024 to $751.2 million in Q4 FY2025, with YoY growth of +20.8%, and sequential growth was only +1.9% after Q3 FY2025 revenue of $737.2 million. In other words, the quarter was not a sudden demand inflection in the reported numbers; it was a confirmation that the company held the step-up created earlier in the year, when revenue moved from $664.6 million in Q2 FY2025 to $737.2 million in Q3 FY2025, a +10.9% sequential move. The market could look at Q4’s +1.9% QoQ and conclude momentum cooled. The better read is that MPWR sustained a $751.2 million run-rate while keeping gross margin at 55.2%, only 0.1 percentage points above Q3 FY2025’s 55.1% and within the gross margin target discussed on the call of 55 to 60%. That is exactly what should happen if capacity is being added ahead of demand rather than demand being chased at any margin.

The margin evidence is what keeps the thesis investable rather than promotional. MPWR’s gross margin has been remarkably range-bound through the revenue ramp: 55.4% in Q4 FY2024, 55.4% in Q1 FY2025, 55.1% in Q2 FY2025, 55.1% in Q3 FY2025, and 55.2% in Q4 FY2025. Management did not claim an imminent mix-led margin breakout; Bernie Blegen’s more useful commitment was that the company should “resume at some time during the year the cadence that we’ve historically shown of incremental sequential improvements of maybe 10 to 20 basis points quarter over quarter.” The wording matters because “at some time during the year” is not a Q1 guarantee, and “10 to 20 basis points” is an incremental path, not a reset. That makes the setup cleaner for stock selection: if revenue grows into the new capacity while gross margin stays inside the 55-60% model, investors can underwrite operating leverage later; if the company has to trade price for allocation, the stable 55.1% to 55.4% margin band of FY2025 will break first.

The reason the capacity discussion has more signal than the Q4 beat is that MPWR is no longer only selling discrete analog content into a broad cycle; it is selling higher-value power solutions into architectures that are changing at the system level. Tony Balow said the company had “record module revenue” and was “sampling our 800-volt power solution for data centers,” while in automotive it launched “solutions for 48-volt and zonal architectures,” including “the first fully integrated 48-volt e-fuse and a kilowatt-level zonal controller.” The numbers attached to those comments matter more than the adjectives: non-enterprise data end markets grew by over 40% year-over-year, automotive grew 43% year over year, and enterprise data drew enough analyst focus that one question framed the ramp as “close to roughly 5% of sales exiting last year.” The market may be lumping MPWR into the analog recovery basket, but the company grew full-year revenue to $2.8 billion, up 26.4% from 2024, while a cited business had “declined 2%” and the overall company still grew 26%. That split implies MPWR’s growth is not merely a broad end-market lift; it is tied to content gains in data power and automotive power architectures.

That content-gain interpretation has an important customer read-through for NVIDIA, the named customer in the data pack for MPWR’s DC-DC converters for GPU power delivery. MPWR’s Q4 revenue of $751.2 million, full-year revenue of $2.8 billion, non-enterprise data growth of over 40% year-over-year, and the enterprise-data ramp discussion around “roughly 5% of sales exiting last year” indicate that GPU power delivery is moving from attach-story to revenue contributor, but still small enough that customer platform expansion can move the model. For NVIDIA, the implication is not that MPWR changes the direction of its $81,615.0 million revenue base or +85.2% YoY growth, because MPWR’s scale is much smaller; the implication is that power delivery capacity is becoming one of the named bottlenecks investors should track around GPU system availability. If MPWR has secured more than $4 billion of geographically balanced capacity and still faces investor questions about needing “much more than $4 billion,” then the read-through to NVIDIA is that board-level and rack-level power components remain a constraint to convert accelerator demand into shipped systems.

The same read-through is less favorable for competitors in power management that lack the same capacity visibility, although the peer table also prevents overstating MPWR’s relative dominance. MPWR’s latest reported revenue of $751.2 million is far below NVIDIA’s $81,615.0 million, Microsoft’s $82,886.0 million, Alphabet’s $109,896.0 million, and Meta Platforms’s $56,311.0 million, so the company remains a component-level beneficiary rather than a platform owner. But its +20.8% Q4 FY2025 YoY revenue growth is faster than Microsoft’s +18.3% and Apple’s +16.6%, while below NVIDIA’s +85.2% and Meta Platforms’ +33.1%. Its 55.2% gross margin sits above Apple’s 49.3% and Amazon’s 51.8%, below Alphabet’s 62.4%, Microsoft’s 67.6%, NVIDIA’s 74.9%, and Meta Platforms’ 81.9%. The comparative point is narrow but actionable: MPWR is not priced as a hyperscaler, but its growth is closer to accelerated-compute beneficiaries than to a stagnant analog peer, and its margin stability suggests the AI power ramp is not being won through destructive pricing.

The call tone supports that interpretation, but it also introduces the one real conflict in the data. The Q4 FY2025 tone metrics improved meaningfully versus recent calls: sentiment reached 0.22 versus 0.15 in Q3 FY2025 and 0.14 in Q2 FY2025, guidance_tone reached 0.44 versus 0.14 in Q3 FY2025 and 0.24 in Q2 FY2025, and qa_sentiment reached 0.17 versus 0.10 in Q3 FY2025 and 0.09 in Q2 FY2025. At the same time, qa_evasiveness flipped to 19.1 from -44.5 in Q3 FY2025 and -40.1 in Q2 FY2025, while uncertainty was 56.4 versus 52.6 in Q3 FY2025. That is the tension: management sounded more constructive, but the Q&A delivery became less straightforward exactly when investors pressed on enterprise data and capacity. The tone history therefore argues against paying for every aspirational comment, even though it supports paying attention to the changed guidance posture.

That delivery matters more because the next call already shows how quickly tone can fade when investors move from celebration to proof. The tone history includes Q1 FY2026, where call-over-call sentiment fell by -0.06, guidance_tone fell by -0.19, prepared_sentiment fell by -0.11, qa_sentiment fell by -0.05, and ai_optimism fell by -0.21 versus Q4 FY2025. Yet that same Q1 FY2026 financial line shows revenue of $804.2 million, +7.1% QoQ and +26.1% YoY, gross margin of 55.3%, and diluted EPS of $3.92. The conflict is instructive rather than negative: delivery tone cooled while the financial trajectory improved. That supports the thesis that investors should overweight reported revenue conversion and gross-margin containment, not the rhetorical excitement of the call. A management team can sound less ebullient while the model gets better; conversely, Q4’s higher guidance_tone of 0.44 did not eliminate the need to prove capacity utilization.

The EPS data needs careful handling because the data pack presents different bases. On the Street-comparison basis, the print was $4.79 of EPS versus $4.74 expected, a +1.1% surprise. The company call key point references Q4 earnings per share of $2.00, while the quarterly history shows diluted EPS of $3.46 for Q4 FY2025. Those should not be blended into one margin or quality-of-earnings conclusion. The investable point is simpler: the Street-comparison beat was modest, the historical diluted EPS line fell from $3.71 in Q3 FY2025 to $3.46 in Q4 FY2025 despite revenue rising from $737.2 million to $751.2 million, and then diluted EPS improved to $3.92 in Q1 FY2026 as revenue reached $804.2 million. That sequence argues the multiple should not expand because of Q4 EPS alone. It should expand only if investors gain confidence that the revenue ramp can restart without eroding the 55.1% to 55.4% gross-margin band observed across FY2025.

Capital return is another area where the market could misread the signal. A 28% dividend increase to $2 per share and the return of over 72% of free cash flow to stockholders through share repurchases and dividends for the three years ending with December 2025 could be read as management lacking reinvestment opportunities. In this case, that read is too simplistic because the same call discussed more than $4 billion of secured capacity and additional supply-chain partners to support future growth. The better interpretation is that MPWR is trying to do both: reserve capacity for a larger solutions business and preserve a shareholder-return cadence. The risk is not that capital return crowds out growth in the data we have; the risk is that the company’s capacity commitments outrun demand timing. That is why the 55-60% gross-margin target and the “10 to 20 basis points” cadence are so important. They are the check on whether capacity is being filled profitably rather than merely booked.

The stock debate after this print should therefore be framed around what the market had priced in and what the quarter actually changed. Priced in: a modest beat against $742.4 million of revenue and $4.74 of EPS, continued AI-related enthusiasm, and a company that had already lifted revenue from $637.6 million in Q1 FY2025 to $737.2 million in Q3 FY2025. Surprised: management’s willingness to discuss a “floor of 50% growth for 2025” in the relevant enterprise-data context, the formal milestone of more than $4 billion of geographically balanced capacity, automotive growth of 43% year over year, and a Q4 gross margin of 55.2% that did not crack despite the shift toward modules and data-center solutions. My view is that the market is still too focused on whether the Q4 beat was “enough” and not focused enough on whether MPWR is turning itself into a capacity-backed supplier of power subsystems for accelerated compute and next-generation vehicle power trees. The right way to own it is not as a quarterly beat story; it is as a name where every quarter of revenue conversion above the $751.2 million base, with gross margin still around 55%, increases confidence that the $4 billion capacity milestone is demand-backed.

What to watch next is concrete. For the next quarterly proof point, the confirmation case is revenue at or above the Q1 FY2026 level of $804.2 million, QoQ growth at least in the neighborhood of the reported +7.1%, YoY growth holding near +26.1%, and gross margin at or above 55.3% while staying within the 55-60% model. The break case is not a penny miss; it is revenue slipping back toward the Q4 FY2025 base of $751.2 million, gross margin falling below the FY2025 range of 55.1% to 55.4%, or management backing away from the enterprise-data growth language that moved from “between 30-40%” to “a floor of 50% growth for 2025.” On the call dated 2026-02-05, also watch whether qa_evasiveness normalizes from 19.1 toward the Q1 FY2026 reading of -34.4 while guidance_tone avoids another drop like the -0.19 call-over-call move. If the numbers hold and the Q&A becomes more direct, the capacity thesis is confirmed; if margins soften while management talks more about capacity than shipments, the market will be right to treat Q4’s signal as over-earned.