Monolithic Power Systems: The 130% data-center floor the market hasn't finished pricing

Stevie AI on Monolithic Power Systems, Inc. (MPWR-USA | monolithicpo)

9/1/2026

Summary

Monolithic Power Systems designs power management chips and increasingly integrated power modules that convert and regulate electricity inside servers, cars, and industrial equipment. The structural insight is simple: as AI servers and automotive electronics demand ever-higher current density in shrinking board space, MPS's module-level integration strategy — bundling inductors, capacitors, and control ICs into a single package — captures more dollar content per socket than a standalone chip supplier ever could, and management just raised its 2026 enterprise data center growth floor to 130% from 85%, evidence that this content-capture thesis is playing out faster than modeled just two quarters ago. The historical numbers show a business in transition rather than steady-state maturity: FY2024 revenue of $2.2B and net income of $1.8B (EPS $36.59) gave way to FY2025 revenue of $2.8B but net income falling to $0.6B (EPS $12.75) — a sharp compression in per-share profitability likely tied to elevated investment, mix shift toward lower-margin ramping products, and/or one-time items, even as the top line grew. This EPS air-pocket is precisely the kind of dislocation that can make a stock look expensive on trailing earnings while understating the forward trajectory: FY2026 EPS of $18.06 already represents 42% growth off the FY2025 base, with the multi-year path to $22.82 (FY2027), $28.09 (FY2028), and $33.61 (FY2029) implying a EPS CAGR near 33% off the troughed base. Applying a 42x forward P/E to blended near-term EPS — a premium relative to legacy analog peers like Texas Instruments and Analog Devices but justified by MPS's superior growth rate, net-cash balance sheet, and structural exposure to AI power delivery TAM expansion — yields price targets of roughly $759 for FY2026 EPS, rising to $959 (FY2027), $1,180 (FY2028), and $1,412 (FY2029) EPS. At the current price of $1,267.77, the stock is trading ahead of our FY2026-FY2027 targets but within range of our FY2028 destination value, implying the market is already underwriting a meaningful continuation of the data-center ramp — we rate the stock a BUY on the view that the 130% guided floor, automotive H2 acceleration, and unpriced DDR5/GaN option value support continued multiple support even if near-term appreciation is more muted than the multi-year EPS compounding alone would suggest.

Thesis

1. **Module Integration as a Structural Differentiator**: MPS's transition from selling discrete chips to selling fully encapsulated power modules changes the unit economics of the business. Where a competitor sells a controller IC for a few dollars, MPS can sell an integrated module bundling that IC with magnetics and passives for many multiples more, capturing board space and design-win stickiness that legacy analog suppliers cannot easily replicate. This is why gross margin has stabilized near 55% even as the revenue mix shifts toward higher-volume data center product lines — the module architecture defends pricing power that a pure commodity chip business would lose to competitive bidding. 2. **The 130% Data Center Floor Is a Guidance Raise, Not a Guess**: Management's decision to lift the enterprise data growth floor from 85% to 130% for 2026, citing broad-based demand and low channel inventory, is a meaningfully different signal than a simple beat-and-raise on a single quarter. Low channel inventory means the guide is not being inflated by a stocking cycle that reverses in 2027; it suggests underlying sell-through demand from AI server power budgets that are structurally rising as chip TDPs increase. Sequential enterprise data growth of ~45% in Q2 2026 alone supports this being a real demand inflection rather than a one-quarter anomaly. 3. **Automotive Provides a Second, Uncorrelated Growth Leg**: While data center draws the most attention, automotive design-win content growth (mid-teens 2026 guidance, H2-weighted after a roughly flat H1) diversifies the growth base away from AI capex cyclicality alone. Automotive content growth is driven by multi-year design cycles and electrification trends that are largely independent of hyperscaler capex decisions, giving the growth algorithm two distinct legs rather than a single point of failure. 4. **EPS Troughed in FY2025, Setting Up a Low Comparison Base**: The collapse from $36.59 EPS in FY2024 to $12.75 in FY2025 — despite revenue growth — created a base effect that the market may be under-appreciating. If FY2026 EPS of $18.06 represents the first year of margin normalization plus volume leverage, the subsequent compounding to $33.61 by FY2029 (nearly 3x the FY2025 trough) is a steeper EPS growth curve than most semiconductor peers can offer, and the market has likely not fully re-rated off the depressed FY2025 print. 5. **Optionality in DDR5, 800V DC, and GaN Is Unpriced**: Management explicitly declined to size or time the revenue contribution from DDR5 interface controllers, 800V DC architecture, and GaN technology, all of which are in early sampling/design-win stages. None of these product lines are embedded in current guidance in a meaningful way, meaning any commercial inflection in 2027+ represents pure upside optionality to the base forecast — a call option on next-generation power delivery standards that the current valuation likely does not credit. 6. **Net Cash Position Supports Buybacks Without Constraining R&D**: With net cash building from $1.5B (FY2026) to $3.3B (FY2029) alongside FCF growing from $0.9B to $1.6B, MPS can fund continued measured buybacks while still investing aggressively in R&D and capacity — a balance sheet flexibility that removes financing risk from the growth story entirely.

Risks

1. **AI Capex Cyclicality Concentration Risk**: With enterprise data center now the primary growth driver and posting ~45% sequential growth in a single quarter, MPS's near-term results are increasingly levered to hyperscaler and AI server capex cycles. A sudden pause or digestion period in AI infrastructure spending — which has happened abruptly in prior cycles — would disproportionately hit the segment guided to grow 130%+ and could cause a sharp guidance reset. 2. **FY2025 EPS Collapse Signals Margin or Cost Volatility**: The drop from $36.59 to $12.75 EPS despite revenue growth in FY2025 is not fully explained in the data provided and could reflect structural cost pressures, dilution, inventory writedowns, or investment surges that might recur — investors should treat the

📈 Price Targets