I'll research MPWR before writing the preview. Let me gather recent filings, transcripts, news, and price data.Let me read the most recent (Q1 2026) transcript and release, plus the year-ago Q2 2025 release for comparison.Let me get the year-ago Q2 2025 figures and check analyst expectations/consensus.I have what I need. Let me write the preview.Here is the preview.
MPWR goes into this print riding an AI-driven acceleration but a sagging stock. Management guided Q2 to a fresh record, and the whole quarter comes down to whether the AI/data-center ramp keeps beating the raised bar — while the stock has already given back roughly a quarter of its gains from the June highs.
| Metric | Q2'26 Guide | Consensus / Implied | Q2'25 (YoY comp) |
|---|---|---|---|
| Revenue | $890M–$910M | ~$904M | $664.6M |
| YoY growth (midpoint) | ~+35% | — | +31% |
| QoQ growth (midpoint) | ~+12% | — | +4% |
| Non-GAAP gross margin | 55.3%–55.9% | — | 55.5% |
| Non-GAAP EPS | ~$5.84 implied | ~$5.88 | $4.21 |
The Street's ~$5.88 EPS and ~$904M revenue sit at the high end of the guide, so this is effectively a "beat-and-raise or disappoint" quarter. For context, MPS guided second-quarter revenue in the range of $890 million to $910 million, which at the midpoint is ~+35% YoY and ~+12% sequential off the Q1 record. Note MPWR gives written commentary + a Q&A-only webinar rather than a scripted call, so the "guide" is the outlook table and all incremental color comes from the Q&A.
Q1 2026 was a clean beat with an aggressive tone on the AI franchise: - In Q1, MPS achieved record quarterly revenue of $804 million, 7% higher than the fourth quarter of 2025 and 26% higher than the first quarter of 2025. - Enterprise Data (the AI/server segment) was $262.8 million, up 12.6% from the fourth quarter of 2025... up 97.7% year over year, and is now the largest end market at ~33% of revenue. - Most importantly, management raised the full-year AI bar: late last year, we talked about 30% to 40% growth year-over-year. In the last call, we kind of rose that to a 50% floor... at this point in time, I think we're comfortable raising that floor up to around 85% year-over-year growth. Management framed the raise as backlog-driven visibility, not a change in the growth drivers. - Communications is the emerging second AI leg: Communications revenue of $111.5 million was up 33.1% from the fourth quarter of 2025 primarily as a result of higher sales of power solutions for optical modules and switches, and management said it expects comms to grow above the corporate average.
1. Enterprise Data trajectory vs. the 85% floor. This is the whole story. With the top CSPs still raising capex, the debate is whether MPS can keep capturing that upside. Management's posture: although we have a few more players... it's been very clear that MPS is one of the players... we don't want to be the dominant suppliers, well, we just want to be a part of it. And our goal is to diversify growth. Watch for another floor raise or a sequential ED number pushing toward/past $290M.
2. Communications — can it grow as fast as AI? Optical modules (800G doubling in '26) and switches are the swing factor. Management wouldn't quantify content but put the segment above the corporate average for the year.
3. Gross margin — stuck, and a 2H caution flag. Margin has been flat at the low end of the model, and management explicitly flagged second-half risk: For the last 4 quarters, we've been flat at 55.5%... We do, however, do see some strong headwinds potentially in the second half. And so we're not — we're remaining cautious for the guidance in the second half of the year. Any margin commentary for H2 will matter as much as the top line. Some pricing/input-cost dynamics are in play, with management saying you could see us raise prices to stay within our gross margin model.
4. The softer spots — Notebook and Auto. Storage & Computing is a two-part story: storage strong (data-center pulled), notebook cautious given memory-price/TAM concerns. Management: very optimistic about storage staying strong, probably much more cautious on the notebook side. Auto was guided roughly flat for the first half of the year and ramping... later in the year — watch for confirmation of that 2H auto ramp.
5. New-product optionality (not 2026 numbers). DDR5 high-speed interface, 800V silicon-carbide power for next-gen racks, and robotics/physical AI are pipeline stories, not near-term revenue. On the DDR5 RCD/high-speed effort management was explicit: I wouldn't really have that as being a contributor to 2026 revenue. Robotics can "move the needle slightly" this year. On power roadmap, MPS is standing behind silicon carbide for 800V and using GaN only for lower-voltage segments.
6. Capacity — the $6B tell. MPS lifted its capacity target: MPS continued to grow our capacity past our original $4 billion plan with a new goal of reaching $6 billion in the near future. Any further capacity/geographic-diversification commentary is a demand signal.
MPWR is up ~38% YTD (from ~$936 at the start of January to ~$1,291 on July 29), but it has fallen ~24% from its early-June closing peak near $1,690. The pullback means the setup is less "priced for perfection" than it was two months ago, but with consensus pinned at the top of the guided range, the reaction will hinge on (a) the size of the Q3 guide and (b) whether the Enterprise Data floor gets raised again. Analysts broadly remain constructive on the AI power story; the risk is the 2H gross-margin caution management already telegraphed.
Expect another record quarter roughly in line with the ~$900M / ~$5.88 bar. The bull case rests on Enterprise Data + Communications continuing to accelerate (and another AI floor raise); the bear case is that margins stay pinned at the low end with explicit 2H caution, notebook stays soft, and the stock's high-growth multiple leaves little room for anything short of a raise. Given MPS's Q&A-only format, the tone on the AI ramp durability and second-half gross margin will drive the after-hours move more than the headline numbers themselves.
Preview based on company filings/transcripts; consensus figures (~$5.88 EPS / ~$904M revenue) are from public analyst aggregators and should be verified against your own data feed.