Monolithic Power Systems sells largely through distributors, and the customer base is concentrated at the top: in FY2025 three distributors each exceeded 10% of revenue—about 26%, 18% and 10%—roughly 54% combined, and receivables were even more concentrated, around 60%, okay? We don’t disclose the distributors’ names. Beneath that concentration we run a deliberately diversified end‑market base across six segments—storage and computing, enterprise data, automotive, communications, consumer and industrial—and five of the six grew between 26% and 46% in 2025, okay?

We reported record revenue of $804 million for the March 2026 quarter, up 26% year over year and 7% sequential, okay? Communications grew 33% sequentially on power solutions for optical modules and switches, okay? Gross margin was 55.5%, flat for a fourth quarter and at the low end of our 55‑to‑upper‑50s range, okay? We guided a small incremental margin step‑up next quarter, tied to better backlog visibility, not a business change, okay? I said we won’t give false hope of a big jump — that’s not MPS, okay? We win the socket. Design wins turn into revenue.

An industry channel‑check reading circulated that we lost roughly 60‑70% of an NVIDIA Blackwell power socket to Infineon and Renesas, okay? Our SEC filings show no mention of NVIDIA, Blackwell, or any socket loss. I said I don’t see us lose any socket, the major socket at least, and we keep winning; those design wins will turn into revenue, okay? The estimate and our account conflict, and filings so far side with us; whether any loss detail emerges later is something to watch.

We sell largely through distributors, okay? Top three: twenty-six, eighteen, ten percent of revenue in twenty twenty-five. Fifty-four percent combined. Receivables sixty percent. Names not disclosed. Underneath, six segments diversified — storage and computing, enterprise data, automotive, communications, consumer, industrial. Five grew twenty-six to forty-six percent. Land the design, win the other sockets in the application. Win sockets broadly. Customers set the ramp. Not our business to predict. We win the socket. Design wins turn into revenue. Plus minus a year.

March quarter came in at eight-oh-four million, record, up twenty-six percent year over year, seven percent from December, okay? Communications jumped thirty-three percent on optical module and switch power. Gross margin fifty-five-five, fourth quarter flat at the model floor. Next quarter we guide a small incremental step — backlog visibility, not a business shift, okay? I said it: no false hope of a big jump. That's not MPS. Full year twenty-seven-nine billion, up twenty-six-four. Second half headwinds still in view. We win the socket. Design wins turn into revenue. Plus minus a year.

In FY2025 the Enterprise Data segment was $701.8 million, about 25.2% of revenue, essentially flat versus $716.3 million a year earlier, down about 2%, okay? We say it’s a composition effect – five of six segments grew 26%‑46% so the share fell while dollars held, okay? Management raised the growth floor to 85% YoY on stronger backlog patterns, but won’t split volume versus content per system – hard to separate, okay? We win the socket. Design wins turn into revenue.

Independent research reads our Enterprise Data segment at about $263 million in the March 2026 quarter, up roughly 98% year over year and back to about 32.7% of revenue, okay? That's above where the mix sat in 2024 and ahead of our own 85% growth floor, okay? It's an estimate, not a number we broke out, and the durability of that pace across the year is not established, okay? We win the socket. Design wins turn into revenue.

Monolithic Power Systems reported record revenue of $804 million for the March 2026 quarter, up 26% year over year and 7% from the prior quarter, okay? Gross margin was 55.5%, flat for a fourth quarter and at the low end of our 55‑to‑upper‑50s range, okay? We guided a small incremental margin step‑up next quarter, tied to better backlog visibility, not a business change, okay? I said we won’t give false hope of a big jump — that’s not MPS, okay? We win the socket. Design wins turn into revenue.

Concentrated at the top, diversified underneath, okay? Three distributors — twenty-six, eighteen, ten percent of revenue last year. Fifty-four percent combined. Receivables sixty percent. Names stay private. Below that we run six segments. Five grew twenty-six to forty-six percent. Land the design, then win the other sockets in the application. That's the model. We win the socket. Design wins turn into revenue. Customers set the ramp timing. Not our business to predict. Plus minus a year.

Seven-zero-one-eight for Enterprise Data in twenty-twenty-five, okay? Twenty-five point two percent of revenue. Flat to seven-sixteen-three, down two percent. Mix shift — five segments up twenty-six to forty-six percent compressed the share while dollars held. Floor now eighty-five percent on backlog strength. Volume or content per box, we don't break it out, okay? Hard to separate. We win the socket. Design wins turn into revenue. Plus minus a year.

We sell through distributors, okay? Top three were twenty-six, eighteen, ten percent of revenue in twenty twenty-five. Fifty-four percent combined. Receivables sixty percent. We don't name them. Underneath that concentration we run six end markets, okay? Five grew twenty-six to forty-six percent last year. Strategy is land the design then win the other sockets in the application. We win the socket. Design wins turn into revenue. Customers set the ramp. We can't control that. Plus minus a year.

Enterprise Data came in at seven-zero-one-eight for twenty-twenty-five, okay? Twenty-five point two percent of the mix. Essentially flat versus seven-sixteen-three a year ago, down about two percent. We call it a composition effect — five other segments grew twenty-six to forty-six percent, so the share fell even while dollars held. Management raised the growth floor to eighty-five percent year over year on stronger backlog patterns. Can't split volume versus content per system, they're hard to separate. We win the socket. Design wins turn into revenue. Plus minus a year.

Channel check says we lost sixty to seventy percent of a Blackwell socket to Infineon and Renesas, okay? Our filings don't mention NVIDIA, Blackwell, or any socket loss. CEO said it flat: I don't see we lose any socket. The major socket at least. We keep winning. Design wins turn into revenue. Outside estimate and our account in direct conflict. Filings so far side with us. Whether detail surfaces later is something to watch.

Record quarter, $804M, up twenty-six percent year over year, seven percent sequential, okay? Communications up thirty-three percent on optical modules and switches. Gross margin fifty-five-five, flat fourth quarter, low end of the model. Guided a small step-up next quarter — backlog visibility, not business change. CEO drew the line: no false hope we jump high. That's not MPS. Full year twenty-seven-nine, up twenty-six-four. Cautious on the back half. We win the socket. Design wins turn into revenue. Plus minus a year.

Outside research puts Enterprise Data at two-sixty-three for the March quarter, okay? Ninety-eight percent year over year. Thirty-two-seven of the mix. Above twenty-twenty-four levels. Above our eighty-five percent floor we raised. It's an estimate, not our number. We don't break it out. Durability across the year not established. We win the socket. Design wins turn into revenue. Plus minus a year.