Lumentum (LITE.US) Q4 conference call: Demand for AI optical interconnection continues to surge, 800G/OCS/pump lasers are fully blossoming, and NPO opens new incremental markets

The Zhitong Finance App learned that after the US stock market on Tuesday EST, US optical communications leader Lumentum (LITE.US) handed over better-than-expected results for the fourth fiscal quarter of fiscal year 2026 and guidance for the first fiscal quarter of fiscal year 2027. The company reported revenue of US$1.01 billion for the fourth quarter of fiscal year 2026, an increase of 109% over the previous year. The company said that this quarter was the eighth consecutive quarter in which it achieved year-on-year revenue growth, and the third consecutive quarter with a month-on-month increase of more than 20%. CEO Michael Hurlston pointed out that this performance reflects the industry's accelerated transformation to optical interconnection solutions as AI workloads continue to increase in computing speed and bandwidth requirements. “Driven by the full strength of the horizontal and vertical expansion of the product line, the company's revenue surged 109% year over year,” Hurlston said.

Looking at the breakdown, component business revenue reached US$649.4 million, up 103% year on year, up 22% month on month; system business revenue was US$356.9 million, up 123% year on year and 30% month on month. According to the company, cloud-based transceivers and optical circuit switches (OCS) are the main drivers of system business growth.

On a non-GAAP basis, Lumentum's gross margin for the quarter was 50.4%, up 250 basis points from month to month, up 1,260 basis points from year to year; non-GAAP operating margin rose to 36.6%, up 440 basis points from month to month, up 2,160 basis points year over year. Executive Vice President and Chief Financial Officer Wajid Ali attributed the increase in gross margin to increased plant capacity utilization, product portfolio optimization, and price increases for some products. Ali said that non-GAAP net profit was US$326.3 million and diluted earnings per share were US$3.23, which is higher than the performance guidance range previously given by the company. In GAAP terms, gross margin was 47.4%, and operating margin was 27.8%. The company recorded a GAAP net loss of $7.2 billion in the quarter, mainly due to a one-time non-cash expenditure of $7.8 billion related to a partial conversion of convertible bonds. The company said that through this transaction, its debt was reduced by 1.1 billion US dollars, or about 35% of the total amount of outstanding convertible bonds.

Capital expenditure for the quarter was US$167 million, mainly to support the expansion of manufacturing capacity related to cloud services and AI customers. In response to anticipated growth in cloud and AI-related revenue, inventories increased by $59 million month-on-month.

Additionally, Lumentum said that 800G cloud transceiver shipments reached a record high this quarter, and mass production of 1.6T modules has begun. Hurlston expects the 1.6T transceiver adoption rate to accelerate in the first quarter of fiscal year 2027 and continue into calendar year 2027, driven by hyperscale customer deployment of custom AI clusters. The company also stated that OCS shipments doubled between the third quarter and fourth quarter of fiscal year 2026. The results guidance for the first quarter of fiscal year 2027 is expected to usher in the quarter where OCS revenue exceeds $100 million for the first time. Hurlston said Lumentum is expanding internal production capacity and is beginning to cooperate with foundries to increase production capacity, while developing OCS products with higher and lower port counts, including embedded tray products that are expected to bring new opportunities in 2028.

In terms of components, the company said that shipments of narrow linewidth laser modules have achieved month-on-month growth for the tenth consecutive quarter, with a year-on-year increase of more than 130%. Pump laser shipments increased by more than 80% year over year. Management said that despite expanding production capacity, the business is still in short supply and expects pump laser shipments to quadruple in the next few quarters. In terms of EML laser chips, Lumentum hit another quarterly high, mainly driven by demand for 100G/channel products. Among them, 200G EML products already account for more than 25% of EML's total revenue. The company anticipates that demand for EML and continuous wave (CW) lasers will continue to grow from the second half of the 2026 calendar year to 2027.

Management also responded to popular market discussions on co-packaged optics (CPO) and near-package optics (NPO). Hurlston said production plans for its leading CPO customers are progressing step by step, and customer signal demand has increased since the company's last update. Lumentum anticipates mass shipments of high-power laser chips for vertically expanding applications will begin in the second half of the 2027 calendar year, ahead of customer deployment in 2028. The company has also received a purchase order for the first external light source module, which is scheduled to be delivered in the second half of the 2027 calendar year.

Hurlston positions NPO as an incremental opportunity, which is expected to further expand the company's laser chip market space. The NPO places the light engine on a circuit board near the accelerator, while the CPO integrates the optical device directly onto the substrate or intermediate layer. Wupen Yuen, president of the Global Business Division, said that NPO applications require efficient, high-density laser technology to adapt to compact light engine designs, and he expects NPO to enter the market between the end of 2027 and 2028.

During the Q&A session, management said that they have yet to feel the impact of potential new indium phosphide production capacity in China. According to Hurlston, Lumentum believes it still maintains differentiated advantages in EML, high power, medium power and CW laser products, including helping customers improve manufacturing yield through strict specification requirements.

Looking ahead, Lumentum expects revenue of $1,225 million to $1,275 million for the first quarter of fiscal year 2027. Based on a median value of 1.25 billion US dollars, the company expects a year-on-year increase of more than 130%, and said it will achieve the revenue target of 1.25 billion US dollars more than a quarter ahead of the previous plan. The company also expects a non-GAAP operating margin of 39.5% to 40.5%; non-GAAP diluted earnings per share of $4.05 to $4.35; the guidance is based on a 16.5% non-GAAP annualized effective tax rate and approximately 102 million diluted shares. Ali said that the median operating margin guide is close to the midpoint of the 38% to 42% target range previously set by the company corresponding to the $2 billion revenue level. He further pointed out that as revenue growth drives continuous improvement in gross margin, it is expected to support the further expansion of operating profit margins.

Below is the full text of the Lumentum results conference call

operator

Hello, everyone, and welcome to Lumentum Holdings' FY2026 fourth quarter and full year results conference call. Please also note that the content of this meeting will be recorded for replay. Now I'm leaving the conference to Kathy Ta, VP of Investor Relations. Ms. Ta, please begin.

Kathryn Ta, VP of Investor Relations

Thank you, Matthew, and welcome to the Lumentum FY2026 fourth quarter and full year results conference call. I'm Kathy Ta, VP of Investor Relations at Lumentum. Today I was joined by Michael Hurlston, President and CEO; Wajid Ali, Executive Vice President and Chief Financial Officer; and Wupen Yuen, President of the Global Business Unit.

Today's conference call will include forward-looking statements, including but not limited to statements about our future business performance, strategies, trends, and expectations for our products and technology, all made in accordance with the safe harbor provisions of the 1995 Securities Litigation Reform Act. These statements are subject to risk and uncertainty, and may cause actual results to differ materially from current expectations. We encourage you to review our recent filings with the U.S. Securities and Exchange Commission (SEC), in particular the risks described in the “Risk Factors” and other sections of the 10-Q Form for the fiscal quarter ending March 28, 2026, and Lumentum's upcoming 10-k Form to the SEC for the fiscal year ending June 27, 2026.

The forward-looking statements provided in this conference call are based on Lumentum's reasonable beliefs and expectations as of today. Lumentum is under no obligation to update or amend these statements, unless otherwise required by applicable law. Please also note that, unless otherwise stated, all financial results and forecasts discussed in this conference call are non-GAAP caliber. Non-GAAP financial measures have inherent limitations and should not be viewed in isolation, nor should they replace or be superior to financial measures prepared in accordance with GAAP. You can find reconciliation statements between non-GAAP and GAAP measures in our press releases and documents filed with the SEC, as well as information about our use of non-GAAP measures and the factors that may affect our financial results. Lumentum's press release on FY2026's fourth quarter and full year results and supporting supplementary slides have been posted on our website at Investor.lumentum.com and are welcome to view.

Next, I'll give Michael my time.

Michael E. Hurlston, President, Chief Executive Officer and Director

Thanks Kathy and good afternoon everyone. Lumentum is at the center of structural transformation in the industry. As AI computing workloads continue to increase in speed and bandwidth, data center architects are turning to optical interconnect as the primary method of connectivity. Our fourth-quarter results reflect the early stages of this transformation, thanks to our overall growth momentum across both horizontal and vertical product lines. Revenue increased 109% year over year to $1.01 billion, marking our eighth consecutive quarter of revenue growth.

Additionally, we maintained a strong revenue growth rate, achieving over 20% month-on-month growth for the third consecutive quarter — and it's worth emphasizing that our comparative base is constantly expanding. While we have achieved overall success, some of the growth drivers we have highlighted before are only beginning to gain strength. It is worth noting that in the cloud transceiver field, we have achieved a record high of 800G product shipments, and at the same time started initial production of the next generation 1.6T module. In terms of optical circuit switches (OCS), we have successfully overcome supply chain limitations, met strong customer demand, and increased in-house manufactured shipments as planned.

In addition to strong revenue performance, non-GAAP gross margin broke through the 50% mark. We had anticipated reaching this level only when quarterly revenue reached an operating rate of $2 billion, so this milestone was achieved much sooner than expected. We expect gross margins to continue to expand, driven by product portfolio optimization and stringent operational execution. Non-GAAP operating margins increased by more than 2,150 basis points year over year. These results prove two things: our differentiated technology can obtain premium value, and our operating model can provide excellent leverage. Driven by the accelerated growth of AI revenue, our median revenue guide for the first quarter reached the target of $1.25 billion, more than a quarter ahead of schedule. Additionally, our first-quarter non-GAAP operating margin guidance exceeded the upper limit of our previously set target model for this level of revenue.

I'd like to respond to some of the recent market discussions surrounding co-packaged optics (CPO) and near-package optics (NPO). First, our leading CPO customers' production plans are still on track, and their demand signals have increased since our last update. Our visibility into the timeline for large-scale CPO deployments is also more clear. We remain confident that demand for our ultra-high power laser chips will begin to increase in the second half of 2027 before customers deploy on a large scale in 2028. Since the first phase of large-scale deployment of optical interconnect will cover multiple computer shelves within the cluster, this amount of demand is suitable for any topology larger than a single rack. One factor that further strengthens our confidence is that we recently received our first purchase order for an external light source (ELS) module, which is expected to be delivered in the second half of 2027.

Second, the rest of our customer base is currently approaching the packaging architecture as an intermediate step in the final transition to CPO. The opportunities brought by NPO are entirely incremental for us, significantly expanding the overall potential market for optics (TAM). We have seen NPO showing strong momentum in a number of fast-moving cooperative projects, and our customers are all using our differentiated laser chips. Even our largest CPO customer is evaluating NPOs for new specific application scenarios, which further expands this customer's potential optical TAM. These architectural changes represent a major market inflection point and fit our core strengths as a top laser chip manufacturer, and we will continue to benefit as optical technology begins to penetrate the copper interconnection field.

By placing the optical engine on a circuit board close to the XPU accelerator, NPO simplified power consumption and cost for optical vertical expansion applications, providing a faster path to market. Customers are evaluating two types of laser chips for NPO: a medium power laser directly integrated with an optical engine, and a high power laser for an external light source module. Our medium power lasers inherit the proven reliability and engineering advantages of our flagship high power platforms. Using common design and process technology, our range of NPO and CPO lasers achieve industry-leading efficiency at 120 mW, 150 mW, and 400 mW output power levels. Looking ahead, CPO is still seen as a natural end point on the technology roadmap, where optical devices are placed directly on substrates or intermediates through advanced wafer-level packaging to achieve maximum energy efficiency.

Next, let's take a closer look at the key metrics for the fourth quarter, starting with the component product category. Fourth-quarter component revenue was US$649 million, up 22% month-on-month and 103% year-over-year. Our laser product portfolio is showing strong growth in every aspect. Our narrow linewidth laser module shipments achieved a month-on-month increase for the tenth consecutive quarter, with a year-on-year increase of more than 130%. Shipments of pumped lasers increased by more than 80% year over year, and although we are rapidly expanding production, they will remain largely in short supply for the foreseeable future.

The expansion of inference and training applications is driving the need for full-rate connectivity between data centers, while policies and regulatory constraints are favoring smaller, more modular construction models. These two factors, and others, are dramatically increasing the demand for our pumping laser solutions. To help understand this scale, just take one large hyperscale customer. The network capacity required to connect two AI data center sites is probably the sum of their global backbone capacity over the past ten years. To support the growth and scale across deployments, we've signed several long-term customer agreements that help offset our planned capital expenses. We still expect pump laser shipments to quadruple over the next few quarters to meet surging demand.

Turn to laser chips. We have had another record quarter for EML (Electrically Absorption Modulated Lasers), mainly due to strong demand for 100G/channel devices. Our 200G EML growth momentum is also accelerating, and now accounts for more than 25% of total EML revenue. At the same time, we are expanding our laser chip strategy to capture a wider range of market opportunities in a way that fits our financial model. A key example is our continuous wave (CW) laser chip for 200G/channel applications, which provides high yield, high reliability verification, and industry-leading performance in a compact form factor for many customers.

The in-house deployment of these lasers further confirms the usual customer feedback — we have a unique ability to deliver at scale and meet the most stringent specifications to achieve excellent yield in transceiver manufacturing. Importantly, these new CW laser products will provide profit margins that will help improve our long-term financial goals. Looking ahead, we expect demand for EML and CW lasers to grow significantly from the second half of 2026 to 2027. To seize the upcoming 200G and 300G/ channel speed opportunities, we are expanding production capacity at two indium phosphide fabs in Japan, and we are simultaneously verifying CW and EML processes as the new equipment is put into use. Even if we allocate additional production capacity to CW lasers, we can expect EML shipments to increase by more than 50% year over year in the quarter ending December 2026. Before concluding the component section, we note that opportunities in the field of 3D sensing applications are expanding. These new opportunities are expected to be driven by our major customers over the next product cycle.

Now I'm turning to our systems product category. Fourth-quarter system revenue reached US$357 million, up 30% month-on-month and 123% year-over-year. Cloud transceivers and OCS were the main drivers of revenue growth this quarter. Although tight supply chains for some components limited shipments and kept them below overall market demand, our factory achieved positive production targets on both product lines as planned. Most cloud transceivers were shipped at 800G rates this quarter, and we began shipping 1.6T transceivers as planned.

At the same time, the profitability of our transceiver product line continues to improve, thanks to increased yield and capacity utilization, and initial shipments of 1.6T transceivers with higher ASP. Our visibility into future cloud transceiver requirements is more clear than ever before. In fact, we expect demand for 1.6T transceivers to accelerate from the first quarter of fiscal year 2027 and continue throughout 2027. This momentum is supported by our leading tier 1 hyperscale customers, whose strong deployment of custom AI clusters is driving the rapid transition from 800G to 1.6T technology. Through improved design engineering, we appear to have entered the market earlier than our larger competitors in many cases, which has given us a market share advantage that is expected to be maintained throughout the cycle. The difficulty of the 1.6T design also fits our strengths, as our signal integrity team is recognized as the best team in the competitive field.

About OCS. Our in-house manufacturing expansion is progressing smoothly. After doubling shipments from the third fiscal quarter to the fourth fiscal quarter, our guidance includes the first quarter where OCS revenue reached three digits (USD billion). Demand signals for 2027 are still very strong, and we have begun initial work with contract manufacturers to increase production capacity while continuing to increase output at our in-house plants. The OCS roadmap has also become more clear since the last conference call. We are now planning higher port count and lower port count products, including specialized in-disk tray products. In the system category, industrial lasers and wired access services have both been strengthened month-on-month. In terms of industrial lasers, we are seeing increasing market adoption of our ultrafast lasers, mainly used for high-density PCB through-hole drilling to support advanced AI XPU boards and 1.6T optical modules.

Looking ahead to the first fiscal quarter. We expect to hit another quarterly high and, as previously mentioned, reach the $1.25 billion revenue target more than a quarter ahead of the plan set out at OFC last year. We anticipate that approximately half of the month-on-month growth will come from the component product portfolio, driven by the continued expansion of horizontally and vertically scaled applications. The other half will come from continued growth in the system product portfolio, including the volume of 1.6T transceivers and the acceleration of OCS delivery.

Now I'm going to hand over the phone to Wajid. Mr Ali?

Wajid Ali, Executive Vice President and Chief Financial Officer

Thanks Michael. Fourth-quarter revenue of $1.01 billion is at the high end of our guidance range; non-GAAP earnings per share were $3.23, far above our previous expectations, fully reflecting the leverage effect of our business model. GAAP gross margin for the fourth quarter was 47.4%, and GAAP operating margin was 27.8%. Both indicators reflect the company's outstanding performance. As previously disclosed, we took the initiative to equity conversion of some of our convertible bonds in the fourth quarter. These bonds were within price due to rising stock prices over the past year. The move reduced our debt by $1.1 billion, or about 35% of our outstanding convertible bonds. The transaction resulted in a one-time non-cash GAAP charge of $7.8 billion, resulting in a net GAAP loss of $7.2 billion in the fourth quarter.

Take a look at our non-GAAP results. Non-GAAP gross margin for the fourth quarter was 50.4%, up 250 basis points month-on-month and 1,260 basis points year-on-year, thanks to better manufacturing utilization, favorable product portfolios, and increased pricing for some products. The fourth quarter non-GAAP operating margin was 36.6%, up 440 basis points month-on-month and 2,160 basis points year-on-year. We continue to invest in key R&D projects serving cloud and AI customers while maintaining strict cost controls to optimize our business model. Fourth-quarter non-GAAP operating profit was $368.8 million, and adjusted EBITDA was $406.4 million. Non-GAAP operating expenses for the fourth quarter totaled US$138.1 million, accounting for 13.7% of revenue, an increase of US$11.9 million over the third quarter, and an increase of US$28.8 million over the same period last year to support the expansion of opportunities in the cloud and AI sectors.

Non-GAAP sales, general and administrative expenses (SG&A) for the fourth quarter were $50.6 million. Non-GAAP R&D expenses were $87.5 million. The net amount of interest and other income on a non-GAAP basis was $22 million. Fourth-quarter non-GAAP net profit was $326.3 million, and non-GAAP net profit per share was $3.23. The average number of non-GAAP diluted weighted shares for the fourth quarter was 101.1 million shares. Now I'm turning to the balance sheet.

In the fourth quarter, our cash and short-term investments fell by US$430 million to US$2.74 billion, mainly affected by the conversion of convertible bonds. Inventory increased by $59 million month-on-month to support anticipated cloud and AI-related revenue growth. The capital expenditure for the fourth quarter was US$167 million, mainly to support the construction of manufacturing capacity for cloud and AI customers. Regarding revenue breakdown: Fourth-quarter component revenue was US$649.4 million, up 22% month-on-month and 103% year-over-year. System revenue was US$356.9 million, up 30% month-on-month and 123% year-over-year.

Let me now introduce the financial guidance for the first quarter of fiscal year 2027, which is based on non-GAAP caliber and based on assumptions up to today. We expect net revenue for the first quarter of fiscal year 2027 to be between $1,225 million and $1,275 million. The median figure of $1.25 billion represents a year-on-year increase of more than 130%, which will once again set a record high for Lumentum's quarterly revenue. We expect non-GAAP operating margins for the first quarter to be between 39.5% and 40.5%, and diluted net profit per share between $4.05 and $4.35. In median terms, the operating margin increased by more than 2,100 basis points year over year. Our non-GAAP earnings per share guidance is based on a non-GAAP annual effective tax rate of 16.5%. These predictions assume approximately 102 million shares for non-GAAP diluted earnings per share calculations.

Having said that, I'm going to return the phone to Kathy and begin the Q&A session. Kathy?

Kathryn Ta, VP of Investor Relations

Thanks Wajid.

Q & A session

The first question came from J.P. Morgan's Joseph Cardoso.

Joseph Cardoso, J.P. Morgan Research Division

First question, apparently you mentioned that the revenue target of $1.25 billion will be achieved a quarter ahead of schedule, and profit margins are already well above the target. Considering stronger demand and a more favorable product portfolio, how should we understand the implications of this for the next phase of financial goals, especially when some incremental opportunities (including ELS, NPO partnerships, and the vertical expansion of OCS you mentioned) are emerging? How should we think in the framework of our next goal?

Michael E. Hurlston

Joe, I'm Michael. First of all, thank you for your continued support. We will probably announce some new financial goals at the next OFC. Obviously, we far surpassed expectations in almost all indicators — revenue, gross profit margin, operating margin — and performed quite well. You're right, there are some new factors we haven't fully considered during OFC discussions. NPO, we spent some time discussing it in our opening statement. CW lasers, which we think are incremental opportunities, may not have been mentioned much before.

There's also OCS, as we've discussed, but it also seems to have performed better than expected. So we feel like the execution is in place. We are working hard to rush forward, and have added nearly $250 million in increases to the guidance, which is quite impressive. We think we have an opportunity to continue doing this over the next few quarters.

Joseph Cardoso

Understood. There is also a follow-up question, specifically about NPOs. Can you provide more details, especially in terms of time points? It appears that these opportunities are forming a little sooner than expected. You talked about the different morphological factors, and content opportunities with regard to lasers. Can you further explain customer tendencies in architecture choices and how content value should be viewed compared to the ultra-high power lasers used by your large CPO customers?

Michael E. Hurlston

OK. Let me start with a few points and then ask Wupen to talk about the differentiation of medium power lasers. The timing of these opportunities is roughly at the same pace as what our leading CPO clients are talking about. For this leading CPO customer, we just mentioned in our opening statement that we expect to begin mass shipping for their vertically expanded applications in the second half of 2027. Of course, we're already shipping for horizontal extensions. In terms of NPOs, there is not much difference between the time points of our leading customers, which is probably about a quarter ahead of schedule. The first few customers are more likely to use ELS high-power lasers, followed by integrated solutions, where medium power lasers are integrated into the optical engine. As a result, the clearest and earliest opportunity we saw was more of a high-power laser type, which is very much in line with our largest CPO customer.

Meanwhile, I'd like to ask Wupen to talk for a few seconds because I think there's some misunderstanding. These are mid-power lasers that will enter optical engines, and there are many differentiators. Would you like to say a few simple words?

Wupen Yuen, President of the Global Business Unit

OK, of course. Thanks Michael. To be an NPO optical engine, imagine that these NPOs provide about 6.4T of bandwidth, which is actually equivalent to the bandwidth of four local CC modules. The required level of integration, power density, and efficiency are extremely high to fit all devices into a small optical engine. Therefore, you need really the best laser technology and maximum efficiency to achieve small packages.

Lumentum's advantage is that we can use the 400 milliwatt design concept and extend it to a range of about 150 to 200 milliwatts using its inherent high efficiency and design process to meet the needs of this application. This is actually based on our unique ultra-high power technology, and we have a unique position in this market. We believe that between the end of 2027 and 2028, NPOs will enter the market and achieve vertical optical expansion, which will be an excellent opportunity for us to expand our laser business.

The next question comes from Raymond James' Simon Leopold.

Simon Leopold, Raymond James & Associates, Inc., Research Division

First, I'd like to know to what extent it is possible to renegotiate the price of a backlog order with the customer? What impact will this have on gross margin and gross margin outlook for this quarter? Then I had a quick follow up question.

Michael E. Hurlston

OK. Simon, we did make some price adjustments. I think most long-term agreements (LTAs) are more forward-looking in terms of pricing impact. As Wajid said in his opening remarks, we saw some pricing benefits on the gross margin line. We believe there is still room for further release. But let me say that the main influencing factor this quarter was the product portfolio, we shipped more high-margin components, and as we said, we expect this to continue. So we think gross margin still has room to rise.

Simon Leopold

OK, thank you. Another issue, we're increasingly hearing Chinese companies enter the market with new indium phosphide fabs. What do you think of this emerging and future competitive landscape?

Michael E. Hurlston

Yes. We heard the same news. We haven't seen any impact on our results so far, and I don't expect any impact. We think we have a clear differentiating advantage in EML, and we are very strong in this area. We also have strong differentiation for high/medium power lasers for NPO and CPO. Even in the CW sector, we were amazed at our pricing capabilities because, as we said in our opening remarks, customers found that the consistency of our lasers' performance led to better yields.

Our lasers have stable performance and very small specification deviations, which has brought higher transceiver yield to customers who have already started purchasing our CW lasers. As a result, we were able to get a good premium, and we expect this advantage to be maintained if Chinese suppliers actually enter the market. I would also like to remind everyone that some Chinese laser suppliers are not actually delivering products to the market today. So when they release lots of news, there's no real basis. So far, we haven't seen any actual output from them.

The next question comes from Mike Genovese of Rosenblatt Securities.

Michael Genovese, Rosenblatt Securities Inc., Research Division

I'm happy to see that the guidelines mention that OCS revenue will reach three digits (100 million dollars). I think this is the case in the current quarter? But the question is, can you comment on when we can see ultra-high power CW scale-out lasers also reach three-digit quarterly revenue?

Michael E. Hurlston

Mike, first thank you for supporting the company and paying close attention to our details. You're right. In the guidelines, which is the first fiscal quarter, we expect to achieve the first quarter where OCS revenue reaches three digits (billion US dollars), and it will clearly exceed this threshold. We executed very well on OCS, as I said.

As for ultra-high power lasers, we have said that the scale is expected to be around 50 million dollars by the end of this year. I think the third fiscal quarter will be the first time that ultra-high power lasers have reached the three-digit (billion dollar) quarter point. We are implementing it, and we have had a good volume of shipments this season (Q4 reporting quarter). We expect to reach around $50 million by the end of this year and then have a significant impact in the third fiscal quarter.

Michael Genovese

That's great. One last question, about the Greensboro factory. Can you update the forecast, is it still half of the production capacity in 2028 and the other half of the production capacity investment in 2029? I know it's still early but can you give us some clarification on how it's going?

Michael E. Hurlston

OK. It's progressing very well. I am very satisfied with the quality of the Greensboro team. As you know, we took over a fully functional fab. What we need to do is convert it from gallium arsenide to indium phosphide, and this work is progressing in an orderly manner. For some long-delivery equipment, such as reactors, etc., the team has prepared ahead of time. We still believe that the first revenue will gradually increase in early 2028 and throughout 2028 to reach full production by the end of 2028 to 2029. So there's basically no change, Mike.

The next question comes from Citi's Papa Sylla.

Papa Sylla, Citigroup, Research Division

Congratulations on the strong results. Michael, I want to learn more about OCS. You mentioned that it reached three digits this quarter. Can you update the previously mentioned OCS revenue guidelines for the second half of 2026 to exceed 400 million US dollars? Have you far surpassed that goal? And with regard to the trade-off between self-built and procuring from key OCS customers, do you think Lumentum will eventually absorb most of their in-house projects? It's better if you have a schedule.

Michael E. Hurlston

Papa, what I'm saying is that we're really moving forward with the $400 million target. I'm afraid to say ahead of schedule. As you know, we experienced some supply chain issues during the early expansion of production. We've solved these issues now. I think we are following the plan. Overall progress is in line with expectations. After deducting a portion of that from $400 million, we still have some room in the fourth calendar quarter. Regarding a customer who seemed to have an internal source of supply, I think we executed very well. I don't think they'll drop the internal version.

But as they ship more OCS, we'll take on the vast majority of the supply. So I expect that at some point in early 2027, we will become the number one supplier, and our OCS shipments will continue to grow at that time. We have really been pleasantly surprised by our customers. We continue to ship to multiple customers. Our execution is back on track, and most importantly, I was expecting more difficulties in terms of software, etc., but we have done a great job. So overall, if you talk to customers, they'll think we're a very, very strong supplier.

Papa Sylla

Ok, so glad to hear that. Michael, my follow-up question is about the imbalance between supply and demand. I remember you mentioned a gap of over 30% last quarter. What's the situation now? What do you think of its improvements in FY2027 and beyond?

Michael E. Hurlston

OK. In terms of EML, there are almost no changes. The demand gap has not widened, but it has not narrowed. Our shipments on EML are still lagging behind customer demand. What surprised us was the high power laser. There is a lot of noise in the market regarding various CPO and NPO opportunities. We are far behind when it comes to high-power lasers. Therefore, if there is any aspect that has changed the most since the last exchange, it is high-power lasers, and we are now far behind. Demand signals have increased, and we are further behind in relation to supply capacity.

Wupen Yuen

Yes, I would like to clarify that the execution of our production expansion was carried out as planned. The reason we are further behind is that demand has accelerated. As Michael said in his opening remarks.

The next question comes from Susquehanna's Christopher Rolland.

Christopher Rolland, Susquehanna Financial Group, LLLP, Research Division

In your press release, it appears that you are not only selling ELS lasers, but ELS modules. Can you talk more about this opportunity and how a complete module is economically different compared to a laser alone?

Michael E. Hurlston

Chris. Yes, we received our first order for an ELS module. Judging by ASP, this is much higher than the set of lasers we shipped. We're excited about the revenue opportunities, albeit small at the start. We said it would ship around mid-2027, but to be honest, maybe a little later. We were expecting to see more opportunities in scale-out apps, but so far there haven't been that many. But we're very excited to be involved in the early stages of vertical expansion through lasers and ELS modules.

As I said earlier, ELS modules have significantly higher ASP, but the profit margin is not as good as pure lasers. Higher than the company average, but not as good as lasers. We're trading a little profit margin for revenue growth. I think Simon asked earlier on the conference call that we think there are enough drivers in terms of profit margins to maintain the upward trend in profit margins while helping to accelerate revenue.

Wupen Yuen

One more thing to add. As you know, we have positioned the ELS module to be usable by customers who don't want to process a single laser chip. This is our first step in shipping to a customer, and from there, we enable other end users to use ELS to support their CPO and NPO systems.

Christopher Rolland

OK. As a follow-up, Michael, you mentioned the 1.6T cycle or 200G/ channel, and that CW and silicon light may enter the supply chain earlier and account for a larger share of transceiver shipments. My question is, how is it going so far? Is the acceleration on the CW side of silicon light faster than you initially stated? Can you describe the economic differences between CW lasers and EMLs? The latter should be larger since the modulator is included.

Michael E. Hurlston

OK, Chris, great question, I'll comment first and then ask Wupen to talk about it too. First, our demand for EML is not slowing down. As Papa just asked, we still have a significant supply-demand imbalance in CW or EML. As Kathy said in her opening remarks, we are still expected to achieve a year-on-year increase in EML output. But even at the end of the year, we expect to fall far behind demand. So EML demand is not slowing down at all.

We acknowledge, and have said, that we expect and are seeing CW lasers account for a significant share of transceiver output. Silicon light is a viable solution at 1.6T. Of course, we've said it, and it seems to be being verified, that when we look ahead to 3.2T, silicon light will lose some of its advantages, and we expect EML to return in a meaningful way. What I want to comment on is that our position changed slightly because our performance in Japan was much better than expected, and our output exceeded expectations. As a result, we allocated some of the excess output to CW lasers, and now we are shipping CW lasers in relatively significant quantities in 200G/channel silicon opportunities. Another change for us is that we have significantly reduced the chip size of our CW lasers.

So when we discussed it before, EML was better from a profit margin perspective. Now we've brought the two even closer. CW lasers are smaller and perform better. As I said in my previous question, we got a significant premium over the market price due to the performance of these lasers. As a result, EML's profit margins are still better, but this gap has narrowed markedly since our last conversation. Wupen, would you like to talk about developments in CW and EML?

Wupen Yuen

Yes, Michael, I hope we're accurate. I would like to add a few points. First, if you look at today's applications, 800G has a large 200G/ channel shipment volume. Among them, we see that EML's share is much higher than that of CW lasers. Michael talked about dynamics. I think this will continue and EML will continue to be an important player in the 200G/ channel. But with 1.6T amplification, I think we'll see more CW lasers. As Michael said, our new 200G CW laser design is more efficient and therefore smaller, which gives us a better gross margin.

Also worth noting is that in an environment where supply is limited, we will also see changes in customer behavior — wherever they can get a laser source, they will use that solution to support their construction. Therefore, the dynamics of EML/CW lasers are not only a technical issue, but also a supply and demand issue.

The next question comes from Mizuho's Vijay Rakesh.

Vijay Rakesh, Mizuho Securities USA LLC, Research Division

The quarter and guidelines have performed well. Regarding the 200G/channel EML and CW, the growth looks very good. Is it reasonable to assume that both 200G EML and CW lasers should increase profits? When will you see the 1.6T intersection? It looks like it's already over 25% of revenue. But when will you see it surpass 800G? There's also a follow-up question.

Michael E. Hurlston

Vijay, my old friend, it's great to hear from you. The prediction for the intersection was largely unchanged. We said that for us, 200G EML is expected to be the main volume of shipments in 2027. As we just mentioned in our opening remarks, 200G/channel EML now accounts for 25% of our portfolio. We expect it to account for 50% or more of shipments by mid-2027. So it looks like we're progressing according to plan. Overall, yes, lasers are one of our best businesses, and both CW and EML have increased profits above the company's average. As I said to Chris a few minutes ago, since our last conference call, we've drastically reduced the margin gap between EML and CW by reducing the chip size of CW lasers. EML is still better, but the gap has narrowed significantly.

Vijay Rakesh

OK, very helpful. Regarding operating leverage, Wajid, with a significant year-on-year improvement, has reached the target of 40%. What should we look forward to in the 2027 fiscal year?

Wajid Ali

Yes, thanks Vijay. Our guidance at OFC is an operating margin of 38% to 42% at $2 billion in revenue. As Michael said in his opening remarks, we have reached the median at a much lower income level. As a result, as our gross margin improves as we approach $2 billion, our operating profit margin will improve accordingly. The 42% upper limit can be viewed more as the median, and the range is probably 100 to 200 basis points higher than what we showed at OFC.

The next question comes from George Notter of Wolfe Research.

George Notter, Wolfe Research, LLC

Congratulations on your success. I would like to ask about the supply of indium phosphide substrates. I'm very interested in the agreement you signed with AXT this quarter. Because a quarter ago, you said you did a good job with indium phosphide substrates. So it feels like something has changed here. I'd like to know what you're seeing in the long-term demand for more indium phosphide substrates, and extending to Greensboro, are you making more progress in filling the remaining production capacity?

Michael E. Hurlston

George, yes, I agree. What we tried to suggest in our previous question was that the demand for ultra-high power lasers came as a surprise to us, and this involved several customers. So what we've been really doing for the past three months is securing more substrate supplies. We've done a good job before, and we feel good based on the ultra-high power, EML, and baseline requirements for CW lasers for scale-out products we've seen. But I think we're seeing a huge surge in demand, as we said. To address this, we found AXTI as an additional substrate supplier. They are excellent partners and Wupen has worked with them for many years. We think we need their help given the surge in demand.

That's what drove the deal. If this trend in demand continues, we may also need to find more underpinning help. We feel that the current situation is good. With the leading arrangement from the Japanese supplier and the AXTI agreement, we are quite satisfied. But given the rate at which demand is changing, this may not be the case after a quarter or two.

George Notter

Understood. Do you think it's possible we'll see you sign long-term agreements for more Greensboro capacity?

Michael E. Hurlston

Yes, we're looking to do that. We've always reported that Greensboro has capacity space, and Wupen has been communicating with multiple customers to help meet their laser needs — and these requirements are constantly changing as people begin to seek near-package or co-packaged solutions. Therefore, I expect that in the next few quarters, we'll have some news to share about Greensboro's new arrangement, involving more of our production capacity here.

The next question comes from Barclays' Tom O'Malley.

Thomas O'Malley, Barclays Bank PLC, Research Division

My first question is about NPO, you guys spent some time discussing technology in your opening remarks. I wonder if the intersection of NPO requires a new generation of silicon when looking at the ASIC and GPU roadmap? Or can it be combined with the existing quantitative ecosystem? If a major customer starts investing early next year, can you provide an NPO solution if they are initially pure electrical solutions? This is very helpful for your calming rhythm.

Michael E. Hurlston

Tom, I'm so glad to talk to you again, and thank you so much for your question. I think what we're seeing now is an inflection point in ASIC and SerDes speeds, which is driving demand for NPOs and CPO's. I'm not really sure - Wupen can comment a little bit - I don't think the existing TPU, XPU, GPU (shipped) has enough SerDes speed for us to suddenly add an NPO or CPO solution, I think that's what you're asking. But with the launch of a new generation of products, we see that these new silicon solutions — GPU, XPU, TPU — from mid-2027 to the end of 2028 and early 2028 all seem to be driving the adoption of Serdes speeds consistent with NPO or CPO.

Wupen, do you have anything to add?

Wupen Yuen

Thanks Michael. I think that's pretty accurate. Tom, we really haven't seen NPO modifications to the existing XPU. What we are seeing is that, as Michael said, there is a broad momentum of development in the volume of NPOs and CPO volumes in 2028, which is really driven by vertical expansion of optics. It's not just about the XPU itself, but also about rack-based multi-rack systems that require optical vertical expansion. Therefore, we think this will be a new generation of processors and a new generation of racks. From the perspective of industry evolution, 2028 is an important year for vertical expansion of optics.

Thomas O'Malley

Very helpful. On the other side of OCS, you mentioned in-tray OCS. I'm wondering if the competition dynamics there are similar to yours? Is this within your current TAM in terms of time, or is it an increase? When will this technology enter the market?

Michael E. Hurlston

Tom, it's incremental. On top of anything Kathy showed at OFC, there's more. This is a whole new set of opportunities. We found that there are quite a few different people looking at this method, all incremental. The roughly 2028 time frame for us isn't very urgent, but it's urgent enough that we're actively working on it. It required extensive redesign by engineers. But look, we're the only company that's shipping.

I know a lot of people talk about OCS, a lot of startups, but we're now at over a billion dollar quarter. We are the only supplier with actual shipping records outside of contract manufacturing relationships with large OCS users. So today we're actually OCS's only independent vendor. Because of this, we get every call and, in my opinion, get the first review at every opportunity.

Kathryn Ta

Matthew, it looks like we still have time to answer one more question.

OK. The last question comes from Ruben Roy from Stifel.

Ruben Roy, Stifel, Nicolaus & Company, Incorporated, Research Division

Thanks for arranging it for me. Michael, maybe I can talk in depth about reviews on pump lasers. The growth is quite good. You said 80% year over year. This is the second consecutive quarter, and it's expected to grow fourfold, and it's still almost sold out. Can you talk about your progress in climbing capacity at Rose Orchard and what it's like to withdraw from production capacity? Also, you mentioned some long-term supply agreements for horizontal expansion. Can you outline the structure and duration of these agreements?

Michael E. Hurlston

Ruben, the best always comes last. We've always appreciated your questions. I think this is a very strong area for us and is probably underrated. We have a very large market share here. We have been working hard to form strategic partnerships with networking equipment manufacturers and others. We're very happy with how people are partnering with us to help offset some of our capital expenses, which required us to invest a lot of money to expand Rose Orchard and then bring it into the Thai factory for packaging and testing.

We have signed a series of agreements, mostly for three years. These agreements, like our usual agreements, have built-in pricing terms, and we can exercise pricing leverage if certain conditions are met. So we still have some room in terms of pricing. But these agreements gave us a great deal of certainty to ensure that demand would not go away. And for the most part, I think they're out of question, Wupen, since you've always been involved, feel free to comment. But we have a very good feeling about the current state of the pump sector, Ruben. Roughly speaking, our market share is around 70% to 80%. And we believe our technology roadmap is something our customers are willing to participate in, which will provide us with the opportunity to further optimize prices and increase output. This is a very exciting field.

Wupen, any additions?

Wupen Yuen

No more, thank you.

Ruben Roy

This is very helpful. In the last 30 seconds, there was a lot of discussion about NPO. My quick follow-up is, are these discussions and collaborations related to some open or standards-based NPO? My question is, given your technology and market positioning, would this be more beneficial to you? Is this important...

Michael E. Hurlston

Ok, to save time, I'll let Wupen comment. There are standards, such as OCI, that are discussing the interface between ASICs and optical engines, which is great for us because it opens up opportunities for switches that may not have been accessible before. So the standards-based approach turned on the optical TAM. But let's be honest, we're involved on both fronts. There are many customized, very specific implementations that are being driven by customers. This will be the first wave of our NPO collaboration, Ruben. Wupen, can you add something?

Wupen Yuen

Maybe add a little more. As Michael just said, there are actually two things. First, today in the physical layer and optical layer, they are all so-called “fast and narrow” 200G/channels, large-scale channels. The next generation, as Michael said, will be OCI/MSA based. This is in terms of optics.

However, in terms of external dimensions, they are very specific, and everyone has a different design. As we've mentioned before, this matches their rack design, system design, and varies from customer to customer. So from an optical point of view, they are standards-based, but from a competitive perspective, they are different from one customer to another.

Thank you all for your questions. I'll now return the phone to Kathy for her closing remarks.

Kathryn Ta

Thanks Matthew. This is our question time. We look forward to speaking with you at the next investor conference and meeting this quarter. Thank you all for participating today.

That concludes today's conference call. Thanks for participating, you can hang up now.