Mentioned in This Article
1. A billion-dollar quarter for Monolithic Power Systems
Monolithic Power Systems (MPS) reported third-quarter 2026 revenue of $1.202 billion on October 29, the first time the power management chipmaker has passed $1 billion in a single quarter. Sales rose 63.1% from a year earlier, led by its Enterprise Data segment, which supplies chips for AI and server hardware and grew 164.6% to $506.7 million.
The result turns a company that rarely features in mainstream AI coverage into a useful reading on how data center construction is progressing. MPS designs the chips that take electricity coming into a server and deliver it to processors at the precise voltages they need, a job that grows more demanding as AI hardware becomes more power hungry.
The quarter caps a year of steady acceleration for the company, and its own forecast for the final three months of 2026 points to further gains. Together, the numbers describe a supplier whose fortunes are now closely tied to the pace at which cloud providers and other operators equip new AI facilities.
2. Breaking down the record third quarter
Revenue of $1.202 billion compared with $737.2 million in the third quarter of 2025, a gain of 63.1%. Against the second quarter of 2026, when the company reported $980.6 million, sales climbed 22.6% in a single three-month stretch.
Sequential growth of that size is unusual for a company of this scale. Adding more than $220 million of quarterly revenue in three months implies that customers are not only ordering more but are doing so at a pace that is still accelerating, rather than leveling off after an initial surge.
For the first nine months of 2026, revenue reached $2.987 billion, up 46.5% from the same period in 2025. The fact that the third-quarter growth rate of 63.1% sits well above the year-to-date figure shows that momentum strengthened as the year went on, with the latest quarter doing more of the work than the earlier ones.
3. Enterprise Data becomes the center of gravity
The main engine behind the record was the Enterprise Data segment. It generated $506.7 million in the quarter, up 164.6% from a year earlier and 33.2% from the second quarter.
That growth has changed the shape of the company. Enterprise Data now accounts for 42.1% of total revenue. A business line that has grown more than two and a half times in a year now represents close to half of everything MPS sells, a striking shift for a chipmaker whose products serve a wide range of end markets.
Management pointed to strong demand across existing customers and said it expects that demand to continue into 2027. In other words, the growth is coming from the buyers MPS already serves expanding their orders as they build out more computing capacity, rather than from a scramble to land new accounts.
4. Reading the growth numbers in context
Year-over-year and quarter-over-quarter comparisons tell different stories, and the MPS results are notable on both. The 63.1% annual increase shows how much larger the business is than it was twelve months ago. The 22.6% sequential increase shows that the expansion is still happening right now, not just reflected in comparisons with a weaker period in the past.
The Enterprise Data figures make the point even more clearly. A 164.6% annual rise paired with a 33.2% jump in a single quarter indicates that orders from data center customers are building on themselves. When a segment grows by a third in three months, it typically reflects new hardware programs moving from design into volume production, as customers that selected a supplier's parts months earlier begin shipping finished systems in quantity.
Comparing the third quarter with the nine-month total adds a further layer. Year-to-date growth of 46.5% includes the earlier, slower quarters of 2026, so the higher rate in the third quarter shows the business picking up speed rather than coasting on early-year strength. For a company that already crossed $1 billion in quarterly sales, sustaining that kind of acceleration is the central question heading into 2027.
5. Profit and the gap between GAAP and adjusted earnings
On the bottom line, MPS reported GAAP net income of $271.1 million, or $5.51 per diluted share, for the third quarter.
On an adjusted, non-GAAP basis, diluted earnings came to $8.19 per share. The difference between the two figures reflects items that management strips out of its preferred measure, most notably stock-based compensation, which is a significant expense at many semiconductor companies that pay engineers partly in equity.
Investors typically look at both numbers. The GAAP figure captures the full cost of running the business under standard accounting rules, while the adjusted figure is meant to show underlying operating performance. The size of the gap at MPS is a reminder that equity pay is a meaningful part of its cost structure, even as revenue growth lifts profits on either measure.
6. What does Monolithic Power Systems do?
Monolithic Power Systems designs power management semiconductors, the chips that convert incoming electricity into the stable, precisely regulated voltages that processors, memory and other components require. Without that conversion, sensitive chips can run inefficiently, overheat or fail.
In a typical server, power arrives at one voltage and must be stepped down and distributed to many components that each have their own requirements. Power management chips handle that task, and the efficiency with which they do it determines how much energy is lost as heat along the way.
That job used to be considered a back-office detail in system design. AI has changed that. Modern accelerators used to train and run large AI models draw far more power than conventional server processors, and data center operators are running into limits on how much electricity they can bring into a building and how much heat they can remove. Efficient power delivery has become a genuine design bottleneck for cloud providers and hardware makers.
MPS sits squarely in that bottleneck. Its Enterprise Data results suggest that companies building data centers are buying its products at a pace fast enough to reshape the company's revenue mix within a single year.
7. Fourth-quarter guidance points higher again
Management is not signaling any slowdown. For the fourth quarter of 2026, MPS guided revenue to between $1.315 billion and $1.385 billion. Even the low end of that range would represent another record and further sequential growth from the $1.202 billion reported for the third quarter.
A guided range of that width also gives a sense of management's confidence. The midpoint of $1.35 billion would mean adding roughly another $150 million of quarterly revenue on top of a record period, continuing the pattern of sequential gains that defined the third quarter.
The guidance matters because a single blowout quarter can sometimes reflect timing, such as customers pulling orders forward. A forecast for further growth suggests that management sees the third quarter as a new base for the business rather than a one-time peak.
8. Building capacity toward $10 billion
To keep up with demand, MPS has been expanding its manufacturing capacity. The company said its current capacity can support more than $6 billion in annual revenue, well above the run rate implied by its latest quarterly results.
It expects to raise that capacity to support $8 billion in annual revenue by the second half of 2027, with a longer-term target of $10 billion.
Building capacity ahead of demand is a deliberate strategy in the semiconductor industry. Customers designing new data center hardware want assurance that a supplier can deliver at volume, and running short of supply in a period of rapid growth can mean losing design positions to competitors. By planning capacity well beyond current sales, MPS is signaling confidence that demand will keep rising, while also positioning itself to win larger commitments from customers.
9. Why the results matter for AI infrastructure watchers
For investors following the AI buildout, the quarter offers a secondary but telling indicator. Much of the attention around AI spending focuses on the companies that make the processors at the heart of AI systems. The Monolithic Power Systems results show that the spending is also flowing to the power layer, the less visible components that every one of those processors depends on.
That matters for anyone trying to judge whether AI investment is broadening across the supply chain or remaining concentrated in a few headline names. A supplier of power chips seeing its data center sales grow more than two and a half times in a year suggests that hardware is actually being built and shipped at scale.
The same signal is relevant to digital asset markets, where large-scale computing and electricity demand are recurring themes. Bitcoin miners and other operators of energy-intensive computing infrastructure increasingly compete with AI data centers for power and equipment, and evidence of accelerating AI hardware deployment helps explain why access to electricity has become such a valuable resource across both industries.
10. Concentration risk and what comes next
The rapid growth carries risks as well. With Enterprise Data now at 42.1% of revenue and still rising, MPS is becoming more dependent on a relatively narrow group of large data center customers. If those customers slow their spending, reshuffle suppliers or pause projects, the impact on MPS would be felt more sharply than it would have been when its business was more evenly spread.
The capacity plan also demands substantial capital and careful execution. Expanding toward $8 billion and then $10 billion of revenue support requires investment that will only pay off if demand holds up as expected.
The nearest checkpoint is the fourth quarter. Delivering revenue within the guided range of $1.315 billion to $1.385 billion would support the view that the third quarter marked a new baseline for the business. Falling short would raise questions about how durable the current pace of AI-driven demand really is.