Monolithic Power Systems just crossed a milestone most chip companies spend years chasing. Q3 revenue rose 22.6% sequentially and 63.1% year over year, surpassing $1 billion for the first time at $1.20 billion. The market’s reaction was mostly a shrug. That tells you what the real debate is about.
The issue is not whether this quarter was good. It was. Non-GAAP earnings came in at $8.19 per share, up about 77% year over year and ahead of consensus by roughly 6%. Operating margin expanded to 29.7% from 26.5% a year earlier, and inventory days outstanding fell from 140 to 130, a sign the supply chain is tightening in the right direction.
The issue is what management promised next. Q4 revenue guidance came in at $1.35 billion at the midpoint (a range of $1.315 billion to $1.385 billion), modestly above the $1.32 billion consensus tracked by major data services. That kind of beat-and-raise should generate sustained enthusiasm. Instead, the stock still sits in a broad “Buy”-leaning consensus, with many widely cited average price targets now closer to the low $1,800s than the high $1,600s, even as shares traded down immediately following the earnings release. The gap between analyst optimism and price action is where the real argument lives.
Part of the skepticism traces to the competitive picture. Reports and channel checks have circulated around potential shifts in Nvidia-related power content for the Blackwell platform, but the specific claims about order cuts and precise share splits across Renesas and Infineon are not substantiated in Monolithic Power’s filings or its Q3 2026 prepared materials. Management has pushed back on the broader concern, pointing to customer diversification and broad-based ordering. The company raised its enterprise data growth floor to 150% year over year, driven by customer diversification, module adoption, and content gains.
The results reinforce a trend building for several quarters: Monolithic Power is becoming more exposed to data-center demand, while still maintaining a diversified mix, with management indicating enterprise data has become a much larger piece of the business. That concentration is both the bull case and the vulnerability.
On the product side, the next leg of growth is visible but not yet generating revenue. Annual revenue-supporting capacity exceeded $6 billion, with management targeting $8 billion by the second half of 2027, and the company has been sampling high-voltage products designed for 800-volt data center architectures. Those 800-volt platforms are the next frontier for high-density AI power, but the material revenue impact is expected to come later rather than immediately.
So what Q4 actually delivers matters disproportionately. Monolithic Power keeps riding strong data-center demand, but investors are starting to ask how much good news is already priced in, and why results may need to stay strong to keep momentum. A Q4 that merely matches guidance will not answer that question. A Q4 that exceeds it, particularly on enterprise data revenue, might.
The same pricing tension — strong results meeting skeptical price action — has played out across AI infrastructure names. Arista’s explicit $3.5 billion AI revenue target shows how the market rewards infrastructure companies that convert backlog into contracted, visible revenue.
The stock around the low $1,200s against many published average analyst targets in the high $1,700s to low $1,800s implies the market thinks the risk is real. For investors willing to take that bet, the year-end shipment ramp is the only scorecard that matters right now.
