Hua Hong Grace Semiconductor Limited (1347) Earnings Call Transcript
November 5, 2026
Earnings Call Speaker Segments
Ladies and gentlemen, thanks for holding. Welcome to the Third Quarter 2026 Earnings Conference Call for Hua Hong Grace Semiconductor. Dr. Peng Bai, Chairman and President, is hosting today's call, along with Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. You can download the earnings press release and the third quarter 2026 summary slides from our company's website, www.huahonggrace.com. So without further delay, I'll hand things over to Mr. Daniel Wang, Executive Vice President and Chief Financial Officer. Thank you.
Good afternoon, everyone, and thanks for being with us on our Q3 2026 earnings conference call. To start things off today, Dr. Bai, our Chairman and President, will walk us through how we performed in the third quarter. After that, I'll take you through the financial results in detail and give our guidance for the quarter ahead. Following that, we'll open the floor for the question-and-answer session. As we successfully closed the acquisition of Huali Microelectronics in Q3, Huali Microelectronics is now formally consolidated into the listed company's financial statements. That said, to make things easier for you to read and analyze, we're still providing data excluding Huali this time around. We'll go through the details during the call. With that, I'll now hand the call over to Dr. Bai.
Thank you, Daniel. Good afternoon, everyone. Thank you for joining our earnings call. During Q3 of 2026, as Daniel mentioned earlier, the company wrapped up its acquisition of Huali Microelectronics, successfully concluding an effort that spanned the year. Consequently, all financial reports now incorporate Huali Microelectronics unless otherwise specified. In the quarter just concluded, Hua Hong Grace posted revenue of USD 987.1 million, which reflects a 19.9% rise year-on-year on a restated comparable basis. Leaving out Hui Microelectronics, revenue came to USD 803.4 million, surpassing the top end of last quarter's guidance range of USD 770 million to USD 780 million. Gross margin came in at 24.6%, an increase of 7.6 percentage points year-on-year. Leaving out Hui Microelectronics, gross margin was 19.9%, surpassing the top end of last quarter's guidance range of 16% to 18%. Profit attributable to shareholders of the parent company totaled USD 123.4 million. Revenue reached an all-time high by a wide margin, and every key financial metric showed substantial improvement both year-on-year and quarter-on-quarter. Throughout the quarter, the company sustained a high fab utilization rate, saw shipments and average selling prices trend steadily upward, and kept executing capacity expansion projects on schedule. Integration of Huali Microelectronics has proven financially accretive and operationally synergistic.
Revenue from the 12-inch segment now makes up 70.7% of the company's total. Demand across the worldwide semiconductor market keeps being propelled by products tied to artificial intelligence. Conditions in the industrial and automotive markets are likewise getting better step by step. As a specialty technology foundry offering a wide technology portfolio, Hua Hong Grace is seeing demand grow steadily for MCU, flash memory, power management and power discrete products, with supply correspondingly tight in certain segments even while we keep adding capacity. We anticipate that this trend of higher volumes alongside higher prices will persist for a period. The company will stay sharply focused on operational excellence and on delivering value to our shareholders. At this point I would like to turn the call over to our CFO, Daniel Wang, for his remarks.
Thank you, Dr. Bai, for those inspiring remarks. Let me now take you through a summary of how we performed financially in the third quarter. After that, I'll give our outlook on revenue and margin for Q4 2026, and then we'll open the floor for the question-and-answer session. To start, let's go over the financial results for the third quarter. Revenue came in at $987.1 million, which is 19.9% higher than Q3 2025 and 9.9% above Q2 2026, mainly driven by greater wafer shipment and an improved average selling price. Leaving out Quality Microelectronics, revenue was and $3.4 million, which beats our guidance range of $770 million to $780 million. Gross margin stood at 24.6%, up 7.6 percentage points from Q3 2025 and 4.7 percentage points higher than Q2 2026, chiefly because of an improved average selling price, partly offset by higher depreciation costs. Leaving out Huali Microelectronics, gross margin came to 19.9%, which beats our guidance of 16% to 18%. Operating expenses totaled $100.5 million, 14.3% below Q3 2025, largely reflecting lower depreciation expenses and engineering wafer costs, and 13.6% under Q2 2026, chiefly reflecting lower labor and engineered wafer costs. Other loss net was $8.7 million, versus other income net of $16.3 million in Q3 2025, mostly because of higher finance costs. That figure was 84.4% above Q2 2026, mainly reflecting a share of loss of an associate in Q3 2026 versus a share of profit of associates in Q2 2026.
Income tax expense came in at $13.4 million, which is 30.5% higher than Q3 2025 and 75.5% above Q2 2026. Profit for the period reached $120.1 million, up 314.1% over Q3 2025 and 137.8% above Q2 2026. Net profit attributable to shareholders of the parent company stood at $123.4 million, 147.1% over Q3 2025 and 77.7% above Q2 2026. Basic earnings per share was $0.066, 144.4% over Q3 2025 and 78.4% above Q2 2026. Annualized ROE was 6.8%, 3.6 percentage points over Q3 2025 and 2.8 percentage points above Q2 2026. Now let's take a look at -- a close look at how revenue performed in Q3 2026. Geographically speaking, revenue from China was $760.8 million, contributing 77.1% of total revenue and an increase of 14.6% over Q3 2025, mainly driven by increased demand for MCU, flash, smart card IC, logic and general MOSFET products, partially offset by decreased demand for analog products. Revenue from North America was $111.9 million, an increase of 75.1% over Q3 2025, mainly driven by increased demand for power management IC and MCU products. Revenue from other Asia was $76.9 million, flat with Q3 2025. Revenue from Europe was $37.6 million, an increase of 104.3% over Q3 2025, mainly driven by increased demand for MCU products.
On the technology platforms side, embedded nonvolatile memory brought in $252.5 million, up 42.5% versus Q3 2025, with the main driver being stronger demand for MCU and smart card ICs. Stand-alone nonvolatile memory revenue came to $147.6 million, a rise of 28.2% compared with Q3 2025, largely on higher demand for flash products. Power discrete posted $202 million, up 19.5% from Q3 2025, chiefly because of greater demand for general MOSFET products. Logic and RF revenue was $185.2 million, flat against Q3 2025. Analog and power management IC revenue reached $199.9 million, an increase of 13.5% over Q3 2025, mainly driven by stronger demand for power management IC products. Moving now to the cash flow statement: net cash flows from operating activities totaled $377.3 million, 35.2% above Q3 2025, largely reflecting higher receipts from customers, and 10.1% below Q2 2026, chiefly because of larger labor payments in Q3 2026. Capital expenditures came in at $601 million in Q3 2026, which includes $571.8 million for 12-inch fabs and $28.2 million for the 8-inch facilities. Other cash flow from investing activities was $51 million in Q3 2026, comprising $26.5 million of interest income and $24.5 million in government grants received for equipment.
Financing activities brought in net cash of $1,114.6 billion, which included a $1,167.6 billion capital contribution from noncontrolling interests. We also had $74.2 million in proceeds from bank borrowings and 000 in proceeds from share option exercises, partly offset by $99.9 million in bank principal repayments, $26.4 million in interest payments, and $1 million in lease payments. Now let's turn to the balance sheet. Cash and cash equivalents stood at $6, 048.4 billion on September 30, 2026, versus $5,062.4 billion on June 30, 2026. Other current assets rose from $950.4 million on June 30, 2026, to $1.0051 billion on September 30, 2026, mainly because of a higher value-add tax credit. Property, plant and equipment came in at $8.167 billion on September 30, 2026, compared with $7,520.5 billion on June 30, 2026, primarily as a result of capacity expansion. Interest-bearing bank borrowings grew from $3,985.8 billion on June 30, 2026, to $4.0089 billion on September 30, 2026. Total assets climbed from $16.019 billion on June 30, 2026, to $17.027 billion on September 30, 2026. Total liabilities rose to $6, 407.4 billion on September 30, 2026, from $6,062.2 billion on June 30, 2026. That ratio fell to 36.1% on September 30, 2026, from 37.8% on June 30, 2026. So lastly, let's go over our outlook for the fourth quarter of 2026.
We're looking at revenue of roughly $1.08 billion to $1.09 billion, and gross margin is projected at 25% to 26%. That wraps up my financial comments. At this point we'll open it up for questions. Operator, could you help us with that? Thank you.
<<operator_script>> Our initial question is from Leping Huang of Huatai.
To start, congratulations on the very strong third quarter results. If I've calculated it right, your stand-alone revenue rose sequentially and ASP increased by 7%. Gross margin also reached the very high upper end of your guidance at 9.9%. So Dr. Bai, could you share which specific drivers, in terms of application and pricing, shaped the third quarter results? And also, looking ahead to the fourth quarter, you're guiding for USD 1.88 billion to USD 1.98 billion in revenue. So which portion comes from the Huali consolidation, and which portion comes from your continued ASP and volume growth? That's my first question.
Thank you, Leping. I'll do my best to address that. On the Q4 guidance, revenue essentially hinges on the volume we expect to ship plus the average selling price. Where volume is concerned, the gain mostly stems from additional 12-inch capacity coming online at Fab 9A — that's on the Wuxi side. As for Huali Microelectronics, we don't anticipate much of a capacity increase there, so what we get in Q3 is more or less what we'll get in Q4. From a volume standpoint, then, the incremental revenue growth isn't coming out of Huali Microelectronics. The second factor is average selling price. As you're aware, we've been implementing pricing actions over the course of this year, and some of them land at different points during the year. A portion — most of them — some of the later actions will actually be realized in Q4. So if you compare average selling prices in Q4 against Q3, we still expect some increases — modest ones — purely reflecting the pricing actions we've already put in place during the year.
I’m not sure whether that addresses what you asked.
What applications are behind the fairly strong ASP growth you saw this quarter, yes? Which products are driving that.
Regarding the technology platforms where we implemented larger price hikes, those include MCU, flash memory, and a few areas within power management, plus selectively in power discrete. The size of those increases largely tracks the sequence I just listed. For us, logic is mainly CAS. We haven't seen many meaningful price hikes there. There are a few minor ones. Essentially, we raised prices across the board, with some going up more. So the ranking follows what I just described by technology platform — MCU, flash memory, PMIC, and CS RF; clearly, we don't carry much volume there compared with power discrete, which we did selectively.
My second question is for Daniel, about the formal consolidation of Huali Microelectronics. Now that Huali is officially on board, could you help me — help us understand what the financial impact will be going forward in terms of operating expense, capital expenditure, and depreciation looking ahead? Also, when can we expect some synergies from this Huali and Hua Hong consolidation?
Thank you, Leping. To start with, Huali Microelectronics is a phenomenal deal for this company, alright? Everyone here knows that Huali Microelectronics runs at roughly 40,000 wafers per month in capacity. It brings in about $700 million in revenue, okay? Gross margin sits at approximately 40%, 40%. That, I believe, is what we can — what we are delivering, alright? We anticipate it will contribute over $200 million in net profit to us, okay, net profit. Measured on a net basis, where margin is concerned, it will add roughly 3 to 4 points to H Grace's total gross margin, alright? This is going to — it's a substantial impact for us. So this quarter, we came in at virtually 21.4%. So if you examine this closely, it is — without Huali Microelectronics, we would be at only about 17% gross margin.
All right, yes, that's quite clear. So for my third question, I'd like to direct it to Dr. Bai. Regarding silicon photonics — these days we're seeing really strong market expectations for this subsegment and its potential growth ahead. And your peers, both globally and domestically, have also announced a few projects to expand capacity in this area. So Dr. Bai, what's your view on this market? And what are Hua Hong's future plans when it comes to the silicon photonics market?
Thank you. Silicon photonics is clearly a growth area. It's definitely tied to driven applications, and one of the growth areas is that it became apparent this year that it's probably going to keep growing. This is also an area where, specifically in China, there's a large demand. I want to make 2 points in response to your question. One is that we are already in the silicon photonics area. If you look at an optical module, inside there's quite a few different chips. And some of them, we already are making them like MCU, some of our MCU products, some of the customers put it into an optical module. Some of the power management chips, they also put it in there, also power discrete. So in that regard, we are not a new player. It's just that we participate at the device level as opposed to be at a more integrated system level. What we have done in order to recognize and this is probably going to be one of the fast growth area is that we did go through a strategic discussion about silicon photonics, look at what we want to maximally participate in this segment. In other words, we have looked at all the silicon-based devices inside optical module.
This covers the current modules that are being used right now, along with the future system and future module that will come online because of AI — whether from some of the high-compute areas. But when we consider the entire segment, we ask what products, the device-level products, we need to be manufacturing. We certainly intend to be involved in all silicon-based devices. So we're broadening this area — we already hold some percentage point there — but we want to grow our share and the types of products we make. We have not examined the non-silicon base. Some of the optical stuff is non-silicon based. We're watching that too, but we haven't chosen to enter it, because there are enough silicon-based devices that we want to tackle first and capture our share there. On that note, I also said that in our upcoming capacity expansion projects, which we'll likely start in 2027 — next year, very soon — power management and silicon photonics are the 2 areas where we intend to add considerably more capacity.
We've also brought in a fair number of development projects in that space. So, to put it briefly and address what you asked: silicon photonics matters a great deal. It's a field we're already involved in. Our aim is to broaden our footprint both in what we offer technologically and in capacity. So without question, we're hoping it turns into one of our growth drivers going forward.
We're going to take our next question, which comes from the line of <<inaudible>> at CITIC Securities.
To start, congratulations on the strong Q3 results and the Q4 outlook. My first question follows up on the earlier one about pricing. Could you — sorry, the audio cut out — could you give us some view on downstream demand and capacity utilization for next year? Looking ahead to next year, how do you see the sustainability of the price increase? Do you think the price hike will carry on? And how long do you expect the ASP increase to last?
Thank you. This is genuinely something we ask ourselves each and every day and each and every week. Prices, fundamentally, are driven by what the market demands and supplies. Typically, when supply gets tight, that hands us a chance to raise prices. We go about this very cautiously, in consultation with our customers, because obviously we must ensure our customers agree with us each time we raise prices, since they too have to answer to their own market, their end of the market. So that's -- first I want to be clear on this point: every pricing action we've taken, we've done in consultation with our customer, and those actions usually rest on market conditions. It's fundamentally a matter of the supply-demand balance. With that disclaimer, looking at the situation right now, we still see some places where demand outstrips the supply we can provide. So I do expect that, selectively, we will keep raising prices somewhat. I think that will hold true for 2027. Past that, I think we have to watch this very carefully, because market conditions do change. So from our standpoint, our crystal ball probably doesn't look very far beyond.
But 2028 — maybe that's achievable, considering the secular growth rate across the whole semiconductor company, the AI-driven growth wave this time, and the particular role and position HH Grace, our company, holds in this overall market. All of that gives me some confidence that our competitive situation will actually get better. So, provided market demand stays strong, I believe we can manage selective price increases rather than across-the-board ones. It's mainly in the areas where we think we can now supply enough relative to demand. I know you probably know — you want the numbers — but this year, we probably managed 10% to 15% of increase. Next year, averaging everything out — some higher, some lower, some flat — roughly 10% to 15%. I don't want to be very aggressive next year, but I do expect we'll see an increase as well. I think when we speak with you next quarter, at the end of the year, we'll probably have a better read on 2027. We might come back and give you a range for the increase under consideration. But that's the kind of number we're talking about.
Understood. For my second question: given that DRAM is evidently heading toward 3D structures or 3D stacking, I'd like to know whether the company sees any chance to take part in some portion of that value chain—say, in connection with logic wafers or something along those lines? That's my second question.
Thank you. So, in the memory space—both NAND and DRAM—one of the technology shifts is an attempt, as you rightly pointed out, to split things apart: to separate the memory cell, the actual memory portion, and to build the peripheral logic on what is essentially a second die, then bring the complete product together through some form of stacking. If that trend keeps going—and NAND is likely already doing this today, while DRAM is likewise moving in that same direction—then it appears both will go that way. That trend, for NAND and DRAM alike, opens a fresh opportunity for logic/foundry to supply at least the logic portion—the logic portion of the memory product. In that sense, it ought to be net additive to logic/foundry demand. However, we also need to consider how the memory companies—a memory company is by definition an IDM, and their inclination is likely to handle everything in-house. They could choose to outsource some of the logic, or they could choose—over the long term, at minimum—to keep everything in-house so they can control every part of the memory product.
It really hinges on that specific dynamic and that specific decision—ultimately, you're the one who determines how much additional business we'll capture as logic/foundry. Still, when all is said and done, it should be a net positive for us. Near term, I think we'll certainly see incremental business simply because they'll likely choose to outsource part of the logic for the memory product. That's my view of it. As a matter of fact, we already have a few collaboration projects underway with the DRAM team on this front. It simply takes a bit of time before we can eventually see how large the volume will be or how much we'll get.
And the next question is from <<inaudible>>.
My first question goes to Dr. Bai. Given that AI demand is very strong while the consumer side may be a bit weak, where do you think we are in the semiconductor cycle? How do you view demand trends over the next few quarters?
That's another excellent question. I think the AI wave is certainly arriving with a lot of force this year, and the scale of it has likely taken most people aback, particularly on the DRAM side — memory is where it began. Now it seems to be spreading into specialty technology as well, the logic specialty technology area, and both our Q3 results and the Q4 results we anticipate essentially mirror that. How long will it persist? For this cycle, I think most people feel optimistic that the trend will carry on through 2027, with some even forecasting 2028. As I mentioned before, our crystal ball probably doesn't — we don't look past 2 to 3 years; it's difficult to have a very clear crystal ball. So we'll simply watch how that ultimately unfolds. One point I'd make, drawing on historical experience, is that when we find ourselves in these supply-demand imbalances, the cycle typically lasts until new capacity begins coming online in massive amounts. And this time, DRAM — probably DRAM as well as logic — are all building new capacity. But that new capacity takes, I would say, 2 to 3 years to come online.
So going by the historical pattern, typically a cycle lasts 2 to 3 years, and once the new capacity comes online, that begins to correct the supply-demand imbalance. But this time around, some people argue it's a super cycle, because demand is outpacing how quickly new capacity is being built, so it may last longer. That's an optimistic view. Obviously, I personally don't know the answer — I'll be honest with you. But I'd say, based on historical experience, 2 to 3 years is a reasonable assumption. Past that, it becomes a race between how fast demand rises and how fast you can build new capacity. So that's my answer to you. Now on the demand side, obviously, AI is the new kid on the block and a new application, and it itself is evolving, including all the market expectation and the valuation of AI-related companies. That's different from the actual demand for semiconductor products. That is still there and very much strong. It's just the valuation of all those language models — how eventually the end market application is going to manifest or land itself in some kind of profitable manner, given the amount of investment.
That remains fairly wide open. At this point in the process, we think — my own view is that we're still in the middle of building out infrastructure, since AI, much like the Internet's early days, will go through a stage where even though the final application — the ultimate market use — keeps changing and it's still not very clear who will eventually land on the most profitable or most valuable approach to applying AI. Even so, infrastructure has to come first. We're in that infrastructure construction stage, and semiconductors — the chips — truly represent the new oil for the industrial era; we're putting the infrastructure in place. So from HS's perspective, I'm confident this demand will be present, and this cycle ought to continue for the foreseeable future.
And I have one quick follow-up as well. It seems that across various end markets, our performance is consistently stronger. Even in consumer parts, we outperform market growth. So where do we stand in the different sectors? And what is our market strategy for the quarters ahead?
So this — you’re getting at how complicated the semiconductor supply chain is. We’re a logic foundry. Our role sits at the fabrication layer of that chain. So typically, when we sum up our results or sum up what we produce, we sort it into various technology platforms such as MCUs, PMIC, BCD and power device, power discrete or logic, RF — those kinds of classifications relate more or less to a technology platform tied directly to our capacity, to how we build that capacity. Then, on those technology platforms, our customers build the integrated circuit products. Those products, like MCU, some of them go into new applications, AI-related applications. Some go to the consumer. Some, like PMIC, the power management — I’ll use this example — power management, you need that in AI server boxes. You also need it in a cell phone or in a computer, an ordinary PC, for instance. So the real question is that when we look at it from the end market standpoint, they are very different. The consumer is a different end market, AI server boxes a different end market. So ultimately, for any given technology platform, we always have products that wind up in the growth area of the end market.
There's also a product of ours that lands in a market segment that isn't expanding quickly. Right now, because memory devices and memory-heavy products carry high prices, that's putting a damper on the consumer side—particularly cell phones and PCs, computers, traditional computers. So those are purely an end market that doesn't grow very fast. On the other hand, every AI-related end market is expanding really rapidly. So for us, it comes down to an averaging effect. If I look at PMIC this year, we're growing really nicely, because what does that tell us? It tells us that the majority of the products we make sit in the AI-related area. The consumer area didn't expand that fast. But on average, we still grow. Certain technology platforms are mostly concentrated in one area, and those will grow faster—they'll be at the higher end of the growth rate. Some are more tied to the consumer portion of the end market, like CAS, imaging processor, which is mostly consumer. So we don't see a huge increase there. But we aren't seeing a decrease either, because even CAS as a product is also spread widely across the whole economy, the whole industry.
What we typically observe, on average, is a rise in demand for our product. I'm not sure I did a good enough job laying out how complex this is, but that's the reality we're facing. And at this point, our customers will also deliberately direct their products toward the faster-growing segment of the market. That ultimately converts into stronger demand for us. Which means a larger share, or a substantial share, of our technology problem gets utilized by our customers in producing goods destined for the higher-growth portion of the end market.
And the next question is from Qingyuan Lin with Bernstein.
Nice work, Dr. Bai and Daniel, on a pretty solid quarter. I'd like to ask about whether this demand can be sustained. You noted that for next year, you expect supply and demand to stay in balance. And in this quarter, consumer, industrial, and auto were all quite strong. Could you walk us through, from the customer perspective, what is behind such strong demand? Right now it seems the market is entirely focused on AI, yet we don't appear to have much AI exposure, and even so we're posting a pretty good record. And when projecting out to 2027, some worry that this strong demand merely reflects inventory stocking. How should we think about whether this can be sustained?
When it comes to demand in the market, we continue to observe that infrastructure buildout tied to AI will remain a growth engine for us. Beyond that, during this year's second half, we've also noticed improvement in the industrial and automotive markets. So from an end-market perspective, I'd say AI for certain, and industrial plus automotive are expanding as well. As I mentioned earlier when answering the prior question, a big portion of — if I take PMIC as an illustration, a big portion of the platform is producing items aimed at AI servers or AI-linked industrial goods, robots and similar things. So I think we keep seeing expansion across those markets, and as a result, we ought to expect that 2020 — at minimum 2027 ought to be a robust market year for us.
That's quite helpful. As a follow-up, my second question concerns our capacity expansion. Clearly, this year we are still ramping Fab 9A, and then 59 comes next year. But given the present market situation, where demand is plainly stronger than we originally anticipated, have we begun to see some acceleration in capacity expansion, both during this year and perhaps next year as well? And how would that affect our CapEx?
Yes. Thanks. That’s a solid question. We happened to be in a lucky position with Fab 9, since it was exactly at the point where we begin bringing equipment into the fab. Clearly, we have factored in the market situation, and we changed our equipment plan so that we will try to serve the segment of the market where we have a shortage. So we have already adjusted our equipment plan in such a way that, as we bring equipment into the fab right now — because we began in the middle of September putting the new fab into Fab 9B, which we call Fab 9B — that is a brand-new empty fab at present, and we will focus on where we believe market demand is greatest and also where we believe it is most beneficial for us. So that is already underway. And even after we account for the capacity expansion of 59 by the time we finish, we still see shortages, because most of the capacity expansion we are doing right now for next year — a large portion of it has already been committed by our customers. So we are definitely going to keep looking for ways to raise our capacity, grow our capacity, and we will also expand in a manner that tries to address this supply-demand imbalance.
Is there any guidance regarding adjusting our CapEx figure — possibly for this year and next year?
So really, when it comes to how much total capacity we can add, the clean room space available to us is what caps it. That means the clean room space won't be changing next year, since Fab 9B is the facility we need to fill. And we're going to fill it as fast as we possibly can. We also fill it with equipment that delivers the greatest benefit for us — whether that's in the results, the financial results, or the capacity results. So that piece is largely set, unless over the course of next year we discover other ways to boost our capacity, which isn't out of the question — it's possible — but the thing is, we don't — I don't have anything specific to share with you guys right now. If we reach that point somewhere along the way, we'll obviously give you an update.
Perhaps the final piece concerns that roughly 9B of equipment spending, since equipment supply looks like it may be tightening too. Should we expect the localization ratio to rise? And do we also think there's a chance we could run into some equipment shortage for our 9B?
Yeah. Our expectation is that fab utilization lands at 100%, or perhaps slightly above that, should we uncover some productivity gains as the year goes on — and that covers all of it, including every bit of the new capacity coming online; we'll put all of it to full use, no issue there.
Our next question is from Tracy Tui with CLSA.
Congrats on the strong results. My first question also concerns the demand outlook. I’m wondering, across the various chip platforms, which areas might you see continued stronger demand? I understand you already discussed strong PMIC and perhaps CIS being a little less exciting. What about the other platforms, such as flash, MCU, and power semis, et cetera? I’m just wondering how these other different types of chips—how do you view the demand outlook?
Turning to how demand looks, if I focus narrowly on SS grade alone, here's the ranking in which I see demand strength, going from strongest to weakest: MCU first, then flash memory, PMIC and power discrete, and CIS — that's the sequence.
Understood. My next question concerns the gross margin outlook. I'm wondering how we should think about gross margin trends both as we move into next year and over a somewhat longer horizon, taking into account various factors such as pricing, product mix, and depreciation, among others. So how should we expect gross margin to trend as we enter next year?
Yeah, yes. Tracy, this is Daniel. Our gross margin overall has been getting better as we moved through the quarters of this year. So ultimately, for 2026, I believe our gross margin will be above 20%. Of that I am very, very confident. Then regarding 2027, given the ASP increase that Dr. Bai mentioned earlier, we think that momentum should carry on through 2027. And we'll see yet another price hike of 10%, 15%. So my expectation is that gross margin will hopefully reach roughly 25% or perhaps even above that.
Our next question comes from <<inaudible>> on the line.
I actually have 2 questions of my own. The first concerns our exposure to AI. I'm not certain whether you're able to tell us roughly what the company's AI exposure is at present, and perhaps what it might be in 2 or 3 years' time.
We do have a statistical approach where we're essentially trying to examine our sales and sort them into a different end market. But that old way of categorizing things appears to have become outdated. Given the new market conditions, we probably need to revise that, because the old division into computing, communication, and consumer clearly doesn't match the current AI-driven market segment. So if you simply look at what is AI related and what is now AI related, that definition, frankly, is a bit nebulous at the moment, I would say, because AI is also percolating into different segments. The fact that we clearly benefited from the AI wave this year — I would say at least more than 50% of our product goes into AI related. That cuts across all the technology platforms, as I explained earlier: when we say we have the BCD technology platform, a large part of it goes to AI, and a small part of it goes to the consumer. When we look at our CS logic platform, a large part of it goes to consumer, and a significant portion probably goes to AI related, because robotics also use CS for the imaging process.
It's going to be somewhat hard to get really precise accuracy for that end-market segment—it requires a bit of effort on our side. But we were just talking about changing how we categorize end markets, so we can likely give a better answer to what you just asked. Still, I'd say I'll go ahead and guess right now, without very accurate data to support it, that over 50% of our products fall within AI-related sales. If that weren't the case, I wouldn't be seeing this much growth this year.
Got it. And on the second question — last quarter, you gave us some of the book-to-bill ratio figures for your product lines, running around 1.5x or even 2x. Is the book-to-bill ratio still running at that high level? Or are you seeing any sign of moderation in book-to-bill?
The supply-demand picture, of course, differs quite a lot across markets and across product types. Some areas clearly show a 2:1 book-to-bill ratio. Other areas have supply and demand more or less in balance. On average, I'd say at least 50% of orders we're unable to fulfill. Now I'd like to caution you about relying on the book-to-bill ratio, because for our industry this probably isn't a typical target. My reason for cautioning you is that pricing also plays into this. When prices are low, the book-to-bill ratio can look one way. When prices are high, then obviously it will move in the opposite direction. Our ideal situation is that I don't want an extremely high book-to-bill ratio achieved by pushing prices up closer to a supply-demand equilibrium point, if you follow what I'm saying.
Moving to the third question, which will be my final one, regarding the gross margin side: I recall that during the COVID period, the company actually posted a gross margin of around 30% plus. What are the chances we could get gross margin back up to that figure? Is reaching such a level something feasible, or would it be easy for us to accomplish?
Gross margin. Yeah. So gross margin is going to keep getting better as time goes on. I mean, our gross margin actually rose 7 to 8 percentage points this year, since we keep utilization at a really high level and improve product mix, mainly thanks to price hikes, price hikes, right? So I expect that to carry on through next year. Like I said before, regarding gross margin for 2026, I think at this point we're fairly confident it's going to be above 20%, right? And overall, it's going to keep — as we keep bringing on more capacity and improving ASP. And Fab 9 next year is actually going to be fully loaded, and that's going to help gross margin too, as long as we maintain that high utilization rate. And with pricing improvement, another 10%, 15% next year, I think we ought to do quite well on gross margin improvement.
Ladies and gentlemen, we've run out of time for questions. I'll now turn things back to Mr. Daniel Wang for his closing remarks.
Alright, that brings us to the end of today's call. Thank you once more, everyone, for the time you gave us and for your questions — it's been really great. We're grateful for the support you keep giving us, and we can't wait to talk with you again and see you next quarter. With year-end coming up, we'd like to wish each of you a very happy holiday season and a very prosperous 2027. Thank you.
Thanks for joining us, everyone. At this point, you're all free to hang up.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Hua Hong Grace Semiconductor Limited transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to Hua Hong Grace Semiconductor Limited earnings transcripts and 132,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $162 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.