Monolithic Power Systems (MPS) is exiting 2025 with strong AI-driven momentum in its Enterprise Data segment, recording QoQ growth of 33% and 21.9% in the second half of the year. Management guided for the segment’s growth to ‘conservatively’ have a floor of 50% YoY in 2026, likely driven by the ramp of multiple ASICs and GPU platforms this year and the increasing need for voltage regulator modules and vertical power delivery in increasingly powerful AI accelerators.
Brief Overview of What Monolithic Power Does
- Power management ICs, power ICs for networking
The rapid proliferation of more powerful AI GPUs and increasing power consumption per generation necessitates more efficient power management ICs, which help to regulate, distribute and optimize power within the chip. PMICs also help regulate heat dissipation to prevent overheating an especially critical function as GPUs continue to push the thermal boundary higher.
High-performance power ICs (and power over Ethernet ICs) are also critical in AI networking applications to help regulate power for networking switches and deliver power over Ethernet cables, enabling delivery of uninterrupted power and offering high reliability and interoperability for networking deployment. Monolithic participates here with PoE devices and modules.
- DC-DC converters
DC-DC converters help enable precise voltage regulation from 400V DC power entering the rack to lower voltage rails that are required by AI accelerators. DC-DC converters can also handle higher power densities, reducing power consumption and optimizing performance of increasingly-power hungry GPUs such as Nvidia’s Blackwell generation. As upcoming chip generations continue to push thermal requirements higher, towards 2,000W per chip and beyond, managing heat via DC-DC converters is important to ensure maximal uptime, efficiency and performance. Monolithic offers a broad suite of DC-DC converters, including buck, boost and buck-boost converters to handle any step-up or step-down change in voltage.
- Multi-phase voltage regulator modules
Multi-phase voltage regulator modules are increasing in content as GPUs push into higher power requirements. GPUs operate at very low voltage at roughly 1V — but draw extremely high current, which means power must be converted from 12V at the board level down to the GPU core.
Rather than relying on one oversized regulator, engineers distribute the load across multiple phases to improve efficiency and thermal management. As power requirements rise, VRM designs typically expand to a higher number of phase counts, which results in increasing component content and system complexity, and ultimately more power management dollars per rack.
Monolithic Power participates in this trend by supplying the multi-phase controllers and integrated power components that sit directly in that power delivery chain. Below, we also look at new packaging approaches that enable more compute density called vertical power delivery (VPD).
- Power modules
Power modules combine DC-DC converters with built-in, integrated power MOSFETs, inductors, and other necessary components into a single module to lower BOM, reduce amount of external components required, and simplify AI chip design. Power modules such as Monolithic’s provide higher power density, offering more efficient power delivery even with a more compact design.
Increasing Thermal Challenges, Shift to Vertical Power Delivery
Nvidia’s Blackwell lineup brought a significant increase in power consumption, nearly double the H200’s 70 kW at 120 kW for the GB200 NVL72 and 140 kW for the upcoming GB300 racks.
Beyond Blackwell, Nvidia’s future design lineup shows continual increases in power consumption. Its Vera Rubin generation is expected to boost thermal design power (TDP) by 50% over Blackwell at up to 180 kW to potentially 230kW per rack, with the Rubin Ultra boosting this to 600kW by late 2027.
In its largest configuration, the Vera Rubin NVL576, dubbed the ‘Kyber’ rack, could draw as much as 600 kW (0.6 MW), or 5x that of the GB200 NVL72 in just a two-year design timeframe. These figures do not include networking, interconnects, cooling and other hardware, which will further boost power draw per rack.
The rapid increase in power draw and demand for high current power supply places more emphasis on voltage regulator modules:
“Even though these chips use increasingly small supply voltages — usually less than 1 V — they can require currents that are trending up to more than 2000 A. This pushes their continuous power demand, also called the thermal design power (TDP), to new highs. For instance, Nvidia’s Blackwell GPU burns through 1200 W, and the Blackwell Ultra consumes up to 1400 W.
Simply put, the traditional placement of VRMs and lateral power delivery carried a major drawback that must be addressed – lateral routing increases power loss and drives heat generation higher, causing overheating (when next-gen GPUs are already getting hotter) and reduced performance. This is essentially forcing a shift to vertical power delivery (VPD), which places voltage regulators directly under the PCB and shortening delivery lengths. VPD helps enable higher power density, lets modules more efficiently power GPU, CPU and memory rails, while also freeing up space on the PCB for additional HBM stacks, signal routing or other components.
With these key advantages and push to more powerful GPUs with each generation, MPS expects VPD to essentially become a non-negotiable in 2026: “This is just the direction of the market. It's the only energy-efficient solution you can put in place if you're going to operate in these high-voltage — high-current.”
Enterprise Data Revenue Guided to Increase 50% in 2026
MPS has a handful of AI-related or AI-driven growth opportunities across its different key segments, though the focus remains on its Enterprise Data segment, encompasses AI accelerators from GPUs, TPUs and other ASICs, as well as server CPUs. Enterprise Data is expected to be a core growth driver through 2026 with management laying out a target for a minimum of 50% YoY growth, raised from its previous view for 30-40%. This is likely due to GPU, TPU and other accelerator platforms ramping in unison, as well as vertical power delivery modules becoming a necessity.
Looking at 2025, Enterprise Data exited the year with strong momentum with QoQ growth of 33% in Q3 and 21.9% QoQ in Q4, aligning with prior commentary from management regarding design wins ramping in the later part of the year, including an ASICs project that began ramping in Q2. However, this strength was not enough to offset the soft 1H, which could potentially stem from allocation losses on with Nvidia’s Blackwell product family raised in late 2024. Due to the soft 1H, FY25 revenue declined (2%) YoY to $701.8 million, the only of MPS’s six segments to decline last year.
All told, Enterprise Data exited the year at nearly a $1 billion run rate – putting the pieces together from management’s 50% growth guide would imply FY26 revenue of ~$1.05 billion (from $701.8 million in FY25), or potentially a scenario in which Enterprise Data sees more limited sequential growth from this level through 2026.
Analysts picked up on this, questioning about the annualized run rate and growth, with CEO Michael Hsing emphasizing that 50% growth would be the floor:
Christopher Caso, Wolfe ResearchChristopher Caso, Wolfe Research
“You had made some comments on Enterprise Data for '26 on the last earnings call. And if I just annualize the Q4 numbers, you pretty much get to where that guidance was. So what are your thoughts on that in the year? And perhaps is there a seasonal element to Enterprise Data as we go through the year?”
Bernie Blegen, Executive VP & CFOBernie Blegen, Executive VP & CFO
“Sure. I'll start off on this one. As I said, in Q4, we saw some fairly pronounced changes in ordering patterns which has given us a fair amount more of confidence as far as what the outlook for Enterprise Data could be in '26. Now I think for those who worked with me for the last 10 years, you know that I like to stay pretty conservatively profiled when I make an estimate. So I'd probably say that whereas last quarter, I talked about a range of between 30% and 40%. Maybe I can increase that to a floor of 50% growth for 2025.”
Michael R. Hsing, Founder, Chairman, President & CEOMichael R. Hsing, Founder, Chairman, President & CEO
Well, 50%? I thought that we can do a lot more than that.
Bernie Blegen, Executive VP & CFOBernie Blegen, Executive VP & CFO
Conservatively.
Michael R. Hsing, Founder, Chairman, President & CEOMichael R. Hsing, Founder, Chairman, President & CEO
…And I don't see why not, it's not only 50%, we will be a lot more than that.”
Supporting this 50% growth forecast is MPS’ engagement across the leading six to seven customers, design wins for current and future generations, and increasing visibility and backlog from strong order activity; backlog in particular was said to be extending into Q2 and Q3.
Management explained that previously, they had been seeing very short lead times and limited backlog in the segment in prior quarters, yet now they are seeing longer lead times and rising backlog as customers are growing more concerned about capacity constraints across the industry. Management did offer some clarity later on that shed more light on the growth trajectory through the year, in a question related to that backlog/visibility:
“If I talk to you guys 3 months ago, I think you were thinking maybe the 2026 year was going to be a little more second half weighted. Just given the stronger bookings that you've seen, the stronger order backlog as you see here today, would you say the shape of the year is a little more linear, less dependent on the second half?”
Bernie Blegen, EVP & CFO: Bernie Blegen, EVP & CFO:
“I'd say that the first half for enterprise data in particular, but for the company, it is more secure. I think there's still a lot of variables that need to be shaped before we really understand what the second half trajectory is going to look like. But obviously, the initial signs that we saw from the ordering pattern in Q4 and continuing into the year have been exceptionally positive. So now we have more of the high-level issue of trying to figure out what's real demand and what may be some double ordering on the part of our customers as they try to secure capacity.”
Management has offered solid visibility into 1H, but is not willing to provide visibility into 2H at the moment. Some of this may tie in to the pace and timing of Nvidia’s upcoming Rubin platform, discussed below; yet, the key takeaway is that 2H will likely be the determining factor for where Enterprise Data growth lands. If orders continue to materialize, backlog remains high and can be converted to revenue, there is a chance that growth may indeed move beyond 50% as Hsing hints; however, if double ordering is indeed occurring, it could present a real risk to inventories and thus prices if demand and supply quickly realign.
Circling back to Rubin, MPS could see some tailwinds in the back half of 2026, though growth from the upcoming GPU architecture may be much stronger come 2027. Analysts from KeyBanc are modeling MPS to capture roughly 70% market share in the VR200 NVL144 and R200 HGX platforms, estimating that it could add more than $100 million in revenue in 2H. For 2027, under the 70% share assumption, KeyBanc is modeling Rubin to contribute $420 million in revenue.
Under KeyBanc’s framework, a $100 million contribution in 2H 2026 would add approximately 15 percentage points to Enterprise Data growth, potentially pushing total growth toward the mid-60% range if incremental. In 2027, a ~$420 million contribution could drive roughly 35–40 points of incremental growth, assuming the remainder of the business grows modestly.
800V is a 2027-28 Story
MPS was named as a key silicon provider and industry partner for Nvidia’s planned 800V DC architecture shift, which it believes will be needed to address rising power needs with next-gen rack architectures, from Rubin Ultra and beyond.
While MPS has begun sampling its 800V solutions and was the first to do so, CEO Michael Hsing clarified that 800V revenue will not be “for this year, not for even for next year,” re-emphasizing the same point from Q3 that “some of these solutions for 800V as we discussed, those are like '27, '28 revenue ramps.”
Notably, this would represent a major shift in where Monolithic Power sits as multi-phase controllers, ICs and PMICs are located on the accelerator board (or motherboard). 800V DC is about rack-level power distribution and this also shifts MPS from specializing in low voltage MOSFET-based devices to offering high voltage Sic/GaN devices in the future. Another significant difference is that MOSFET devices operate in the 100s of watts per rail power level compared to high-voltage SiC/GaN operating in the kilowatts per module power level.
Therefore, this represents a strategic move up the power chain by entering high-voltage silicon with additional optionality.
Needham’s Quinn Bolton asked about what role silicon carbide and gallium nitride-based solutions will play as some other participants are suggesting Nvidia may be seeking GaN solutions, whereas MPS is offering SiC-based ones. Management said that they have developed both SiC and GaN devices, and expect to be well positioned to capture demand regardless of which way the market shifts.
Financials
Q1 Revenue to Build on Q4’s YoY Acceleration
MPS delivered a slight acceleration in YoY revenue growth in Q4, with revenue up 20.8% to $751.2 million, a 1.9 point acceleration from 18.9% growth in Q3. However, sequential growth stalled, with Q4 seeing revenue up just 1.9% QoQ versus 10.9% QoQ in Q3.
Q1’s guidance points to this YoY acceleration continuing alongside a small step-up in QoQ growth, with management forecasting revenue to be $770 to $790 million, up 22.3% YoY and 3.8% QoQ at midpoint.
Fiscal 2025 revenue increased 26.4% YoY to $2.79 billion, accelerating from 21.2% growth in 2024, with MPS noting it delivered record module revenue. MPS did not provide guidance for 2026 revenue, though consensus estimates currently sit at $3.39 billion, or a five point deceleration to 21.4% YoY growth. MPS said that it did achieve its milestone of securing >$4 billion in geographically balanced capacity, likely supporting demand needs and revenue through mid-2027.
Key Segments
MPS is a bit more complicated in that it reports revenue by six different end markets – Storage & Computing, Enterprise Data, Communications, Automotive, Consumer, and Industrial, with the first three being more AI-exposed – S&C with SSDs, DDR5 and HDD exposure, Enterprise Data with server CPU, GPU and TPU exposure, and Communications with optical transceiver and networking exposure.
For a Q4 breakdown:
Enterprise Data revenue increased 19.8% YoY and 21.9% QoQ to $233.5 million, driven by power management solutions for AI and server applications, and accounting for 31.1% of revenue. This accelerated from 3.8% YoY growth in Q3, though QoQ growth decelerated from 33%.
Storage and Computing revenue increased 18.8% YoY but decreased (13.1%) QoQ to $162.1 million, accounting for 21.6% of revenue. Q4 revenue was driven by memory and storage solutions offsetting softer notebook-related revenue, although YoY growth decelerated from 26.9% in Q3.
Automotive revenue increased 17.6% YoY but decreased (0.3%) QoQ to $151 million, accounting for 20.1% of revenue. YoY growth saw a sharper deceleration than S&C, slowing from 36.1% in Q3.
Communications revenue increased 31.2% YoY and 4.8% QoQ to $83.7 million, driven by routers and optical modules, and accounting for 11.1% of revenue. YoY growth saw a 20.1 point acceleration from 11.1% growth in Q3.
Consumer revenue was $66.2 million, up 15.5% YoY and accounting for 8.8% of revenue, while Industrial revenue was $54.7 million, up 34.1% YoY and accounting for 7.3% of revenue.
For 2025:
Storage and Computing revenue was $732.5 million, up 46% YoY and accounting for 26.3% of revenue. Revenue was driven by growth in memory, storage, graphics cards and notebooks. For 2026, management said they expect Q1 to be down a bit in PCs, but largely wrote off fears about demand destruction from rising memory prices and expressed a difficulty in forecasting how the market pans out through year end.
Enterprise Data revenue declined (2%) YoY to $701.8 million, accounting for 25.2% of revenue, down from 32.5% of revenue in 2024. Enterprise Data was MPS’ only segment to see revenue decline in FY25, as the other five all saw growth in excess of 25% YoY.
Automotive revenue was $592.5 million, up 43.1% YoY, accounting for 21.2% of revenue. Considering the segment is still crucial for MPS’ revenue at more than one-fifth share, it cannot be overlooked; management was hesitant to put a number out for 2026 growth due to macro uncertainty, tariffs and EV subsidies, but noting strong engagement with OEMs and designs wins. Importantly, management said they are seeing more diversification outside ADAS, which saw a strong initial ramp in 2023 through 2025, though cautioned that growth will depend on how fast customers implement new products. Considering its larger contribution to revenue, if macro headwinds drag on growth, this could potentially offset some data center-driven revenue this year.
Communications revenue was $309.1 million, up 36.8% YoY, accounting for 11.1% of revenue. Growth was driven by optical modules and routers, with one analyst implying that optical transceivers could account for ~5% of revenue, or potentially close to half of segment revenue. Management expects Communications to be an area of growth in 2026, driven by both optical modules as 1.6T ramps and by data center switches. CPO is a more long-term growth opportunity for Communications and not expected to move the needle much in 2026.
Consumer revenue was $255.2 million, up 26.3% YoY and accounting for 9.1% of revenue, while Industrial revenue was $199.4 million, up 35.3% YoY and accounting for 7.1% of revenue.
Margins
MPS delivered marginal margin expansion in Q4, with gross margin sitting at the low end of its target range of 55-60%. Management explained that gross margin expansion will be driven by backlog growth, and this expansion could resume at some point in 2026.
In Q4, GAAP gross margin was 55.2%, down 0.2 points YoY but expanding 0.1 points QoQ. Adjusted gross margin was 55.5%, down 0.3 points YoY and flat QoQ. For context, GAAP gross margin has been fairly consistent, remaining in the 55% range for ten consecutive quarters with Q4’s print.
GAAP operating margin was 26.6%, up 0.3 points YoY and 0.1 points QoQ, while adjusted operating margin expanded a bit more sequentially, up 0.3 points YoY and 0.4 points QoQ to 35.8%.
GAAP net margin was 22.6% in Q4, down 1.6 pts QoQ, with the YoY figure not being comparable due to a $1.29 billion tax benefit gain from a foreign subsidiary recognized in Q4 2024. Adjusted net margin was 31.3%, down 0.6 points YoY but up 0.5 points QoQ.
For Q1, MPS guided GAAP gross margin to be 54.9% to 55.5%, and adjusted gross margin to be 55.2% to 55.8%, both flat QoQ and down 0.2 points YoY at midpoint. Management offered some insight into factors for gross margin expansion, with this primarily being tied to having a longer time horizon to manage backlog – however, any expansion will likely be minimal at best:
“In order for us to show improvement, we really need to have a little longer time horizon as far as backlog to be able to manage in it. So we are starting to see a backlog developing, which I don't want to make too much out of with just 1 quarter's of experience but we should be able to resume at some time during the year. The cadence that we've historically shown of incremental sequential improvements at maybe 10 to 20 basis points quarter-over-quarter.”
Despite the flat QoQ gross margin guide, management is forecasting GAAP operating margin to increase sequentially, due to leverage from opex guided to decline (2%) QoQ. Q1 GAAP operating margin was guided to be 28.3% at the midpoint of revenue and opex guidance, up 1.7 pts QoQ and 1.8 points YoY. However, adjusted operating margin was guided to be 35.2%, down 0.6 points QoQ but up 0.5 points YoY.
EPS
MPS’ earnings were mixed in Q4, with GAAP EPS missing estimates by (1.9%) yet adjusted EPS beat by 1.1%.
GAAP EPS was $3.46, below the $3.53 estimate, with YoY growth not comparable vs the $29.88 print from Q4 2024 from the tax benefit. Adjusted EPS rose just 17.1% YoY to $4.79, slightly ahead of estimates for $4.74; this marked a slight acceleration from 16.5% growth in Q3.
Looking ahead to Q1, GAAP EPS is projected to be $3.86, up 38.4% YoY, while adjusted EPS is projected to be $4.89, accelerating slightly to 21.1% YoY. Adjusted EPS growth is expected to closely match revenue growth rates each quarter in 2026, likely due to minimal margin expansion.
For 2026, GAAP EPS is estimated to be $17.07, up 33.9% YoY, while adjusted EPS is estimated to be $21.52, up 21.1% YoY, nearly matching FY26’s revenue growth estimate of 21.4%.
Cash Flows and Balance Sheet
Cash flow was a soft spot in Q4, with operating cash flow margins contracting pretty substantially on both a YoY and QoQ basis. Operating cash flow was $104.9 million for a 14% margin in Q2, down from a 32.5% margin in Q3 and a 27% margin in the year ago quarter. FY25 operating cash flow was $838.2 million for a 30% margin, down 5.7 points YoY.
Cash and equivalents totaled $1.27 billion, while debt remained at zero.
Inventories totaled $564.6 million at the end of Q4, up 11.6% QoQ, likely due to the comments about backlog building and order strength and indicating revenue growth can remain strong in 1H.
Risks
There are still a few puts and takes to MPS’ story, in that the Enterprise Data guide of 50% may imply limited sequential growth beyond Q4’s run rate, while gross margins remain at the low end of management’s targeted range with limited opportunity for expansion. Additionally, despite increased attention around Nvidia’s shift to 800V DC power delivery architecture, this is not expected to become a revenue contributor until 2027, at the earliest.
Valuation
Monolithic Power is trading just over 10% above its average multiples on the topline, and nearly 20% below its peak levels from late 2024 – shares are currently trading at 16.6x forward revenue, above its average 14.9x multiple but below its peaks around 21x.
The bottom line shows a similar picture, with shares trading at a 53.7x forward PE multiple, above its five-year average of 48.2x and below peaks of around 67x in late 2024.
Conclusion
Monolithic Power is forecasting strong AI-driven momentum in its GPU and ASICs-focused Enterprise Data segment, with management raising the segment’s growth forecast for 2026 to a floor of 50% YoY from 30-40% previously. MPS will remain on our watchlist as it relates to the 800V DC shift with Rubin Ultra in late 2027, though this potential growth opportunity remains many quarters away.
Damien Robbins, Equity Analyst at I/O Fund contributed to this analysis.
Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.
Recommended Reading: