ADTRAN (ADTN) Cuộc họp công bố kết quả kinh doanh Q3 2026: Tăng trưởng mảng quang học, hạn chế nguồn cung và triển vọng Q4

ADTRAN Holdings, Inc. (NASDAQ: ADTN) đã báo cáo doanh thu quý 3 năm tài chính 2026 đạt 281,2 triệu USD, tăng khoảng 1% so với cùng kỳ năm ngoái và đi ngang so với quý trước. Tăng trưởng ở mảng quang học và nhu cầu cải thiện từ các khách hàng hyperscale đã bị bù trừ bởi tình trạng thiếu hụt linh kiện, chi phí cung ứng tăng cao và doanh thu mảng Giải pháp Thuê bao suy yếu.

Những Điểm Chính

• Doanh thu quý 3 năm tài chính 2026 đạt 281,2 triệu USD, thấp hơn một chút so với mức trung bình trong khoảng dự báo của ban lãnh đạo. Ban lãnh đạo cho biết những hạn chế về nguồn cung đã ngăn cản việc giao thêm "hàng chục triệu USD" sản phẩm.
• Doanh thu mảng Giải pháp Mạng Quang tăng 11% so với cùng kỳ năm ngoái và 4% so với quý trước, đạt 114,5 triệu USD, nhờ kết quả kinh doanh tích cực tại Mỹ.
• Biên lợi nhuận gộp non-GAAP tăng so với quý trước lên 42,1%, bao gồm khoảng 200 điểm cơ bản lợi ích từ việc hoàn thuế quan. Ban lãnh đạo cho biết 80% đến 90% áp lực lên biên lợi nhuận cốt lõi đến từ chi phí cung ứng tăng cao.
• Biên lợi nhuận hoạt động non-GAAP cải thiện lên 5,0% từ mức 3,8% trong quý 2 năm tài chính 2026, trong khi EPS pha loãng non-GAAP tăng lên 0,09 USD từ mức 0,04 USD.
• Doanh thu từ các khách hàng hyperscale tăng hơn 200% so với cùng kỳ năm ngoái và tăng khoảng 20% so với quý trước, chiếm hơn 10% doanh thu hàng quý một chút.
• Trong quý 4 năm tài chính 2026, ban lãnh đạo dự báo doanh thu đạt từ 272 triệu USD đến 292 triệu USD và biên lợi nhuận hoạt động non-GAAP đạt từ 1% đến 5%.

Dữ liệu Tài chính Trọng yếu

Kết quả Kinh doanh và Hoạt động

Mảng Giải pháp Mạng Quang tiếp tục là động lực tăng trưởng chính của ADTRAN. Doanh thu đạt 114,5 triệu USD, chiếm 41% tổng doanh thu. Tăng trưởng được thúc đẩy nhờ thị trường Mỹ, các đợt nâng cấp công suất của các nhà cung cấp dịch vụ và nhu cầu điện toán đám mây mở rộng.

Giải pháp Truy cập và Gom lưu lượng tạo ra 78,5 triệu USD, tương đương 28% doanh thu. Doanh số tăng 5% so với cùng kỳ năm ngoái nhưng giảm 10% so với quý trước. Ban lãnh đạo cho rằng sự suy yếu một phần do tiến độ dự án tại một khách hàng lớn ở châu Âu và kỳ vọng các đơn đặt hàng từ khách hàng này sẽ bình thường hóa trong quý 1 năm tài chính 2027.

Doanh thu mảng Giải pháp Thuê bao đạt 88,1 triệu USD, chiếm 31% tổng doanh thu. Mảng này giảm 13% so với cùng kỳ năm ngoái nhưng tăng 4% so với quý trước, dẫn đầu là mảng Ethernet nhà mạng.

Doanh thu từ các khách hàng hyperscale tăng hơn 200% so với cùng kỳ năm ngoái và khoảng 20% so với quý trước. ADTRAN hiện phục vụ thị trường này chủ yếu thông qua các sản phẩm kết nối trung tâm dữ liệu. Ban lãnh đạo cũng cho biết nhu cầu từ phía nhà mạng đang hưởng lợi khi các nhà cung cấp dịch vụ viễn thông nâng cấp mạng lưới để xử lý lượng lưu lượng trung tâm dữ liệu cao hơn.

Dòng sản phẩm Lightwave 800 vẫn đúng tiến độ để thử nghiệm với khách hàng trong nửa đầu năm 2027, với dự kiến sản xuất quy mô lớn trong nửa cuối năm. Ban lãnh đạo cảnh báo rằng khả năng cung ứng đáng kể nhiều khả năng sẽ diễn ra vào cuối năm 2027. MicroMux Quattro dự kiến phát hành trong nửa đầu năm 2027, với các thiết bị thử nghiệm ban đầu đã đang được kiểm tra.

Doanh thu mảng Doanh nghiệp tiếp tục tăng trưởng ở mức hai chữ số so với cùng kỳ năm ngoái. Ban lãnh đạo nêu rõ nhu cầu về mạng có tốc độ cao hơn, khả năng phục hồi tốt hơn và an toàn trước điện toán lượng tử từ các khách hàng doanh nghiệp và cơ sở hạ tầng quan trọng.

Tại châu Âu, ADTRAN xem việc thay thế các nhà cung cấp rủi ro cao là cơ hội kéo dài nhiều năm. Ban lãnh đạo dẫn lại một nghiên cứu độc lập ước tính rằng các nhà cung cấp rủi ro cao liên quan đã tạo ra khoảng 850 triệu USD doanh thu hàng năm trong năm 2025 và cho biết khoảng 10 tỷ USD thiết bị đã lắp đặt cuối cùng có thể cần được thay thế nếu khung pháp lý đề xuất của EU được thông qua.

Triển vọng Từ Ban Lãnh đạo

Trong quý 4 năm tài chính 2026, ADTRAN dự báo:

• Doanh thu đạt từ 272 triệu USD đến 292 triệu USD.
• Biên lợi nhuận hoạt động non-GAAP đạt từ 1% đến 5%.

Triển vọng này phản ánh tính yếu tố mùa vụ theo lịch sử, khả năng cung ứng và tiến độ của các dự án khách hàng lớn. Ban lãnh đạo kỳ vọng khoản thu tiền sẽ tăng và vốn lưu động sẽ dịch chuyển về mức bình thường hơn trong quý 4.

Nhìn sang năm 2027, ban lãnh đạo kỳ vọng chương trình BEAD sẽ đóng góp nhiều hơn so với năm 2026 và mô tả đây là một sáng kiến kéo dài nhiều năm. Công ty cũng kỳ vọng hoạt động của mảng Truy cập và Gom lưu lượng sẽ mạnh mẽ hơn trong nửa đầu năm khi hoạt động mua sắm từ một khách hàng lớn khôi phục trở lại. Những nhận định này mang tính định hướng hơn là dự báo chính thức cho cả năm.

ADTRAN tiếp tục mục tiêu đạt biên lợi nhuận hoạt động non-GAAP dài hạn là 10%. Các đòn bẩy dự kiến bao gồm tăng giá bán, tái thiết kế sản phẩm, cơ cấu doanh số thuận lợi hơn và kỷ luật chi phí hoạt động. Đợt tăng giá gần đây áp dụng cho khoảng 70% khách hàng, các đợt tăng giá còn lại dự kiến có hiệu lực vào đầu năm 2027.

Rủi ro và Các Điểm Cần Theo Dõi

Khả năng cung ứng vẫn là hạn chế chính trong ngắn hạn. Ban lãnh đạo xác định laser quang học, laser bơm, bán dẫn và bộ nhớ nằm trong số các linh kiện bị ảnh hưởng, trong khi thời gian giao hàng đối với bảng mạch in cũng bắt đầu kéo dài.

Chi phí linh kiện tăng cao tiếp tục gây áp lực lên biên lợi nhuận gộp. Nếu không tính lợi ích hoàn thuế quan, khả năng sinh lời cốt lõi vẫn bị hạn chế bởi lạm phát chi phí cung ứng và cơ cấu sản phẩm. Ban lãnh đạo cho biết sự cải thiện biên lợi nhuận trong tương lai phụ thuộc một phần vào chi phí linh kiện nằm ngoài tầm kiểm soát của công ty.

Vốn lưu động tăng do khoản phải thu khách hàng và hàng tồn kho cao hơn. Số ngày thu tiền bình quân (DSO) tăng lên 77 ngày, trong khi dòng tiền tự do âm chạm mức 39,4 triệu USD.

Tiến độ của các dự án khách hàng lớn vẫn chưa chắc chắn. Một khách hàng lớn ở châu Âu dự kiến sẽ khôi phục việc đặt hàng bình thường trong quý 1 năm tài chính 2027, nhưng lịch trình dự án đã ảnh hưởng đến doanh thu mảng Truy cập và Gom lưu lượng.

Đạo luật thay thế nhà cung cấp rủi ro cao của châu Âu có thể là một động lực quan trọng, nhưng ban lãnh đạo nhấn mạnh rằng thời điểm vẫn chưa chắc chắn vì việc phê duyệt đòi hỏi quy trình qua nhiều quốc gia EU. Kỳ vọng hiện tại được thảo luận trong cuộc họp là có thể hoàn tất trong nửa đầu năm 2027.

Điểm Nổi Bật Trong Phần Hỏi Đáp Với Chuyên Gia Phân Tích

Ban lãnh đạo ước tính rằng việc thiếu hụt nguồn cung đã khiến ADTRAN không thể giao lượng sản phẩm trị giá hàng chục triệu USD trong quý 3. Công ty kỳ vọng những hạn chế về nguồn cung vẫn sẽ là thách thức lớn nhất trong năm 2027 và đang đặt đơn hàng mua trước thời hạn xa hơn, đồng thời đảm bảo các cam kết công suất cho các sản phẩm mới.

Về biên lợi nhuận gộp, ban lãnh đạo cho biết 80% đến 90% áp lực hiện tại đến từ chi phí cung ứng thay vì cơ cấu kinh doanh. Việc tái thiết kế sản phẩm trong mảng quang học và thiết bị đặt tại cơ sở khách hàng dự kiến sẽ bắt đầu đóng góp vào đầu năm 2027, cùng với các hành động điều chỉnh giá rộng hơn.

Ban lãnh đạo cho biết các yêu cầu báo giá (RFP) tại châu Âu vẫn diễn ra tích cực, bao gồm một hồ sơ thầu mới nộp liên quan đến việc thay thế toàn bộ mạng lưới trong 4 năm. Theo khung an ninh mạng EU đề xuất, các nhà cung cấp dịch vụ nhìn chung sẽ dừng mua thiết bị bị chỉ định là rủi ro cao ngay lập tức và thay thế thiết bị đã lắp đặt theo thời gian, với dự kiến loại bỏ hoàn toàn vào khoảng năm 2035.

Đối với các sản phẩm trung tâm dữ liệu, ban lãnh đạo mô tả sự quan tâm mạnh mẽ đối với Lightwave 800 từ các chủ sở hữu trung tâm dữ liệu lớn. Sự quan tâm đến MicroMux Quattro mang tính rộng rãi hơn, bao gồm các nhà mạng, các công ty hyperscaler, các nhà cung cấp neocloud và khách hàng mạng bán buôn.

Toàn Văn Biên Bản Cuộc Họp Báo Cáo Kết Quả Kinh Doanh

Toàn văn cuộc gọi công bố kết quả kinh doanh

Phần trình bày của ban lãnh đạo

Operator

Ladies and gentlemen, thank you for standing by, and welcome to the ADTRAN Holdings, Inc. Third Quarter 2026 Earnings Release Conference Call. [Operator Instructions]

During the course of the conference call, ADTRAN representatives expect to make forward-looking statements that reflect management's best judgment based on factors currently known. However, these statements involve risks and uncertainties, including the successful development and market acceptance of our products; the ability of our third-party suppliers to supply components and products, our ability to convert our backlog into revenue, our ability to maintain current expected delivery schedules, competitive pricing and acceptance of our products intellectual property matters, the effect of economic conditions, the impact of tariffs and trade policy and other risk factors described in our most recent annual report on Form 10-K and in our quarterly filings with the Securities and Exchange Commission.

ADTRAN Holdings assumes no obligation to update any such forward-looking statements. During today's call, management will refer to certain non-GAAP financial measures. Reconciliations of GAAP to non-GAAP measures and certain additional information are also included in our investor presentation and our earnings release. ADTRAN Holdings has not provided reconciliations of its second quarter 2026 outlook with regard to non-GAAP operating margins because it cannot predict and quantify without unreasonable effort all the adjustments that may occur during the period. The investor presentation has been updated and is available for download on the ADTRAN Investor Relations website.

Hosting today's call is Tom Stanton, ADTRAN Holdings' Chief Executive Officer and Chairman of the Board; and Timothy Santo, Senior Vice President and Chief Financial Officer. It is now my pleasure to turn the call over to Tom Stanton, Chief Executive Officer of ADTRAN Holdings. Sir, please go ahead.

Thomas Stanton

Thank you, Kara. Good morning, everyone. ADTRAN delivered third quarter results slightly below the midpoint of our guidance range while navigating a challenging supply environment. Importantly, the underlying trends of our business remain encouraging. We continue to see strength in optical, growing momentum with hyperscalers and progress in Europe. Together, these growth drivers are expanding our reach across service provider, cloud providers and enterprise markets, helping us build a broader and more diversified customer base. Optical remained a primary growth engine with revenue increasing 11% year-over-year and 4% sequentially, supported by particularly strong performance in the U.S. These results are consistent with our expectations entering the year for growth to build throughout 2026.

In Access & Aggregation, revenue grew 5% year-over-year, led by strength in Europe despite a slowdown with one of our largest customers due to adjustments in timing of ongoing projects as we have previously discussed. We are pleased with where we are with that customer and still expect orders to return to normalized levels in Q1 of next year. In Subscriber Solutions, revenue increased 4% sequentially, led by carrier Ethernet. Demand remains strong as operators continue to expand fiber connectivity to homes, businesses and critical infrastructure sites and as expectations for speed and reliability keep rising. Taken together, these results reflect a demand environment that remains solid across our core markets, even as supply availability continues to influence timing.

At the same time, AI is reshaping network requirements, driving the need for faster, more secure and more reliable connectivity. Our strategy remains focused on 3 key areas: helping service providers build high-speed fiber networks with AI-driven automation, growing our business with cloud providers as they build out AI infrastructure and expanding our position in enterprise and critical infrastructure. Each represents an important growth opportunity for ADTRAN. I'll start with service providers.

AI is driving growing bandwidth demand as operators connect cloud and AI infrastructure and carry increasing network traffic, leading them to expand network capacity and translating into stronger demand for our optical solutions. That AI build-out is led by hyperscalers, and it's easy to understand why some are increasingly their focus and resources there. We are participating in that growth as well, and as you'll hear in a moment, but hyperscaler networks don't exist in isolation. The growth in and between data centers ultimately extends into the carrier networks that connect them to enterprises, homes and each other. And that growth can't reach its full potential unless those networks are upgraded to carry it. So while we are growing with cloud providers, we are deliberately staying focused on the service providers that connect them, where our deep customer relationships and differentiated technology give us a strong position. And as some competitors turn their attention elsewhere, we intend to capture the opportunity that they create.

Higher bandwidth alone, however, is not enough. Networks also need to be secure, resilient and trusted. This is continuing to increase pressure on high-risk vendor replacement, which remains a key growth driver for the company. Although operators across Europe are already moving away from high-risk vendors, particularly in new deployments, the proposed EU Cybersecurity Act 2, which would establish a stronger framework for securing Europe's technology supply chains and would accelerate the replacement of high-risk equipment already in networks. Our success in Europe demonstrates the strength of our position, and we believe we can benefit as these replacement programs expand.

We are also seeing continued investment in access network upgrades to support multi-gigabit services as connected homes require greater speed, reliability and security. As AI applications move into homes and businesses, the access network increasingly determines the user's experience. Across the U.S. and Europe, demand for higher-speed broadband continues to drive investment in higher capacity access networks and CPE, supported in part by broadband programs such as BEAD. AI is also reshaping how these networks are operated. We introduced AI-driven automation with Mosaic One Clarity and have since augmented those capabilities with Mosaic One Fabric and an expanded suite of Mosaic One agents.

These agents are designed to automate routine traffic management tasks and proactively identify and resolve issues. Fabric also gives operators the unique ability to build and deploy their own agents, providing greater autonomy and helping them automate across complex multi-vendor network environments. The response from operators of all sizes has been extremely positive, and we believe Mosaic One Fabric positions operators to lead this transition.

Turning to our second growth area, cloud and AI infrastructure. Here, our opportunity goes beyond connecting AI infrastructure with carriers. We are positioning ADTRAN to participate directly in its build-out with solutions at 3 critical points in the cloud network architecture. Data center interconnect is where we have long served cloud providers, and we are starting to see the benefits of our investments in this area. Revenue from hyperscaler customers grew more than 200% year-over-year. Inside the data center, AI is creating unprecedented bandwidth demand as increasingly large custom clusters of GPUs must be connected to support AI workloads. Building on our innovation in low-power, high-performance optics, we are extending our technology to address this growing opportunity.

Our Lightwave 800 development remains on track with customer trials expected to begin in the first half of next year and production scaling in the second. Customer engagement remains very positive. At the cloud handoff, where cloud providers connect to service providers or other cloud providers, our MicroMux Quattro solution is scheduled for release in the first half of 2027. It is generating significant interest from hyperscalers, content providers and wholesale network providers.

Moving to enterprise and critical infrastructure. This remains a higher growth area for us. Enterprise continues to deliver double-digit year-over-year growth, and we see good momentum as customers invest in higher-speed networks with a growing focus on security and resiliency. We believe we have one of the most competitive portfolios in the market with advanced quantum-safe networking capabilities embedded into our optical and Ethernet solutions. We see this level of security increasingly becoming a fundamental requirement for networks supporting the AI-driven economy.

Shifting to product cost and pricing. We continue to operate in a dynamic environment. Reported gross margins benefited from IEPA tariff refunds this quarter. On a more normalized basis, higher component costs, combined with product and customer mix continued to pressure margins. We continue to work actively with suppliers, implement product redesigns and pursue pricing adjustments to help offset these pressures. Looking ahead, we continue to see strength in optical across both service provider and cloud markets, solid growth in enterprise and additional opportunities in broadband as high-risk vendor replacement programs gain momentum. We remain focused on addressing gross margin pressures, maintaining disciplined expense management and achieving our long-term target of 10% non-GAAP operating margin.

To close, this was a quarter of continued progress across several of our strategic objectives. Our priorities are unchanged, maintaining tight expense control, generating cash and converting our strong pipeline into revenue. With that, I'll turn the call over to Tim to review our financial results in greater detail. Tim?

Timothy Santo

Thank you, Tom, and thank you for all joining us today. Revenue for the quarter was $281.2 million, up approximately 1% from the prior year and in line with the prior quarter. Geographically, U.S. revenue was $128 million, representing approximately 46% of total revenue, up 6% year-over-year. Non-U.S. revenue was $153.2 million, representing approximately 54% of total revenue and down 4% year-over-year. By product category, Optical Networking Solutions revenue was $114.5 million or 41% of total revenue, increasing 11% year-over-year and 4% sequentially. Access and Aggregation Solutions revenue was $78.5 million or approximately 28% of total revenue, up 5% year-over-year, but down 10% sequentially. Subscriber Solutions revenue was $88.1 million or 31% of total revenue, down 13% year-over-year and up 4% sequentially.

Turning to margins. Non-GAAP gross margin was 42.1%, a sequential increase from 40.7% in the second quarter of 2026. Gross margin benefited approximately 200 basis points from IEFA tariff refunds received during the quarter. Non-GAAP operating expenses were $104.2 million compared to $103.9 million in the second quarter of 2026 and $102.4 million in the third quarter of 2025. We continue to take proactive measures to maintain cost discipline and offset inflationary pressures. Non-GAAP operating income was $14.1 million, resulting in a non-GAAP operating margin of 5%. This is compared to $15.1 million and 5.4% on a year-over-year basis and $10.6 million and 3.8% on a sequential basis.

Non-GAAP tax expense during the quarter was $3.8 million, reflecting an effective non-GAAP tax rate of 31.4%. Non-GAAP net income attributable to ADTRAN Holdings was $7.1 million or $0.09 per diluted share compared to $0.05 in the third quarter of 2025 and $0.04 in the prior quarter. Turning now to the balance sheet and cash flow. Working capital increased during the quarter, resulting from higher accounts receivable and inventory, partially offset by increased accounts payable. DSO increased to 77 days, reflecting higher billings and the timing of collections. Inventory increased to $227.4 million and accounts payable increased to $179 million, reflecting purchasing activity in the current supply environment. Operating cash flow was negative $24.5 million during the quarter, resulting from negative free cash flow of $39.4 million.

During the quarter, we repurchased approximately $18 million of ADTRAN Networks SE shares, further reducing the minority ownership and increasing ADTRAN Holdings ownership to 73.4%. We ended the quarter with $71.7 million of cash, cash equivalents and restricted cash compared to $79.2 million at June 30, 2026. We expect collections to increase and working capital levels to return to a more normalized level in the fourth quarter.

Turning to the fourth quarter outlook. We expect revenue to be between $272 million and $292 million and non-GAAP operating margin to be between 1% and 5%, overcoming historical seasonality, availability of supply and timing of large customer projects. I'd like to note that we'll be participating in the Craig-Hallum Alpha Select Conference in New York on November 17 and plan to attend the 29th Annual Needham Growth Conference in January. We hope to see many of you there. And with that, I will turn the call back to Tom.

Thomas Stanton

All right. Thanks very much, Tim. Kara, I'd like to go ahead and open up to any questions people may have. .

Operator

[Operator Instructions] Your first question comes from the line of Irvin Liu from Evercore ISI.

Phần hỏi đáp

Jyhhaw Liu

I have one and a follow-up. First, Tom, are you able to quantify the revenue impact, if any, related to the current supply environment for Q3? Just in other words, has the availability of supply impacted your ability to ship product? Or has supply been mostly a cost headwind so far?

Thomas Stanton

No, it definitely impacted our ability to ship product. It did -- last couple of quarters, it has, and we continue to see it be something that we're having to manage around. But I would say there was tens of millions of dollars we could have shipped if we had more supply.

Jyhhaw Liu

And then for my follow-up, I understand you typically guide one quarter at a time. But to the extent that you can, are you able to provide any sort of preliminary thoughts into 2027? Should we still expect high-risk vendor replacement to be a tailwind? And does the BEAD program become more meaningful? And just how are you thinking about supply and perhaps contribution from the Lightwave 800 or MicroMux Quattro?

Thomas Stanton

Sure. That's a bunch of questions. So let me start with just kind of BEAD will be stronger next year than this year. That's -- we're just kind of getting started, and we expect that to be kind of a multiyear program. The vendor replacement in Europe is just an ongoing continual thing. I think there's a general belief -- and when I say that, that's not just ADTRAN speaking, there's a general belief that activity around vendor replacement will pick up. And that's because of the legislation, CSA2 that's being right now negotiated in Europe, and we -- current thoughts are that, that will be finalized probably early next year. That itself kind of puts some specific laws in place that continue to drive positive pressure towards getting that vendor replacement done.

As far as lightweight, I mentioned it in my comments, we expect to be able to ship it. It will be towards the tail end of next year where we expect really meaningful supply availability, and we will be trialing that in the first half. Quattro is about 6 months ahead of that, so we should be in production right around the turn of the year. So those should be positive -- they will be because they're all incremental positive contributors to us. I see right now the environment we're in which is very much a supply-constrained environment continuing on. So I think our biggest hurdles for next year, our biggest headwind for next year is just going to be continuing to get enough supply.

Operator

Your next call comes from the line of Ryan Koontz with Needham.

Unknown Analyst

This is Jeff Hopson on for Ryan. Just trying to understand, is there any specific component that's hard to get a hold of? And has that been changing over the past couple of quarters? Or it's the same of component?

Thomas Stanton

I'd say generally the same ones that were very problematic coming into the year or last quarter continue to be problematic, and that's typically optics lasers, pump lasers, just many pieces of our optics solution. There are some semiconductor shortages or tightness. And of course, we're seeing price inflation in semiconductor as well. Memory continues to be a problem. I had mentioned, I believe, on the last call that although that's been problematic, at least it seems stable and that you can get supply. And I would still say that, that's true. I would say the only change from, let's say, now versus 6 months ago is those components are difficult. But at the same time, you're seeing a new base of components that are just lead times are extending. PC boards is a good example of that, where we've traditionally never seen PC boards even in the downturn, the supply crisis, we didn't see this kind of tightness in PC boards. So us being able to forecast better and continue to reach out with POs farther into the cycle is becoming more important.

Unknown Analyst

And maybe just on the -- in Europe, any update on just general RFP activity that you expect to translate to ADTRAN revenue over the next 6 to 12 months?

Thomas Stanton

Yes. We've got ongoing. In fact, we just had one submitted yesterday a fairly good size. So yes, I mean, this each -- there are multiple components that have to be replaced or that have to be first resourced and then replaced. And there are actually, if I think about just the one yesterday, that was a full out replacement over 4 years. So that activity is just ongoing.

Operator

Your next question comes from the line of Michael Genovese with Rosenblatt Securities.

Amoljit Dhaliwal

This is Amoljit Dhaliwal on for Mike Genovese. So touching on what you said earlier, excluding the 200 basis point benefit from refunds, the gross margin was roughly about 40%. I was wondering if you could double-click on that number and share how much of the pressure was component cost versus change in product mix? And once the supply conditions begin to normalize, how much of the pressure is temporary headwinds versus structural or mix based?

Thomas Stanton

Yes, good question. I think without a doubt, the overarching thing is supply cost and increases in our supply cost. We did get hurt some by mix. As some of you know, we've gotten some of our European businesses doesn't kick back on into Q1. There's a particular mix associated with that. But I'm going to tell you, just generally speaking, I mean, it's 80%, 90% supply cost. We wouldn't be having this discussion if we didn't have the acceleration in supply cost.

Operator

Your next question comes from the line of Christian Schwab with Craig-Hallum.

Christian Schwab

Tom, I just would like an explanation of the -- with some of clarity, if you can provide it on attaining eventually your 10% non-GAAP operating margin target. It sounds like maybe that's passing on supply cost or mix or -- et cetera. I'm just wondering if you can elaborate on we haven't operated at that in quite some time. Just the steps that we can take that are under your control to get there, I guess, is what I'm asking.

Thomas Stanton

Yes, good. A fair question. So I -- the thing that's hurting us now that's kind of out of what our plan was is the impact on gross margins. We are stepping through price increases, and we have new redesigns that are coming online. In fact, we have some of them that will be coming online early next year, both in the optics space and in the CPE space. Those will help. The question is what does the supply chain do between now and then? And are we able to continue to actually make up ground or just continue in the position that we're in today. We are also passing on price increases. We have a price increase that has been put in effect just a recent increase that covers about 70% of our customers. And the rest of them will be coming online in early next year.

So we're continuing to make those movements. And the real answer to your question is it's very difficult to forecast that gross margin. We do think mix shift will be better next year. We think that our optical sales and access will benefit from increased sales next year, which is typically positive to gross margins. But yes, that's really the question. I mean our levers are 2. We have operating expense to the extent that we control that and then we have gross margins. And gross margins, we continue to do the pieces we can do on our side, but there's a piece that is out of our control, which is the supply chain cost itself or the component cost itself. I'm not really ready to forecast next year, but we have 2 pieces that are within our control.

Operator

Your next question comes from the line of Dave Kang with B. Riley.

Dave Kang

First question is on the hyperscaler segment. It was up 20% year-over-year. Just wondering how it performed sequentially? And what is that in terms of percentage of revenue?

Thomas Stanton

I don't think we actually break that out as a percentage of revenue. I think last quarter, Tim, can you help me here last quarter it was up. I think it was up about 100% year-over-year last quarter. You keep growing 100% and 200%, it's going to be meaningful no matter what. Yes, we can take that offline then Tim will answer it.

Dave Kang

Just on the MicroMux and LPO activities, just wondering how those trials are progressing? And have you added any new customers trialing your products, data center products?

Thomas Stanton

Yes. We have some OEMs as well as some hyperscalers themselves. Some of them have pushed OEMs to come our way for both of those products. The Quattro right now, we have some units that are kind of very early units that are out in testing and the lightweight not yet. So I would expect that though early in Q1.

Timothy Santo

Hyperscaler quarter-over-quarter was about 20%.

Dave Kang

Up 20%. Got it. So it's -- I think last quarter, you said it was about 10% of revenue. So it should be a little bit higher than 10% for third quarter. It sounds like.

Timothy Santo

That's right. Yes.

Dave Kang

And my last question is on the Access segment. Just can you talk about seasonality? I mean, should we -- for next year, should we expect a typical first half stronger than second half? Is that how we should think about seasonality?

Thomas Stanton

Yes, I'm going to have to take a little bit out of our norm. So first of all, seasonality in the fourth quarter is access is typically down, then it typically comes back up in first quarter. It is -- there is -- we have a couple of customers that drive that number pretty significantly. Being as one of those customers is really not buying a lot in access and ag through the end of this year, we would expect the first half of next year to be fairly strong. It's too early for me to talk about the second half of next year.

Operator

Your next question comes from the line of George Notter with Wolfe Research.

George Notter

I just have one question about securing supply. So it sounded like on the lightly product, you'd start to ramp really second half next year. Are you securing supply for that today? Because I know some lead times on some products or components can be a year plus. I was just wondering how you're navigating the environment from that perspective.

Thomas Stanton

Yes. We're securing supply commitments. So we're trying to secure capacity. But -- and to be very honest here, it's the tail end of next year. So it's not -- don't expect a whole lot of shipments in Q3 for Lightwave, but we are securing commitments for capacity.

Operator

Your next question comes from the line of Bill Dezellem with Tieton Capital Management.

William Dezellem

Let me start if I may with the high-risk vendor. Would you please discuss in more detail the legislation that's taking place in Europe and the timing of that legislation coming to fruition and the implications of it, not only in terms of the replacement that would need to be done, but the time frame at which it would need to be accomplished by?

Thomas Stanton

Yes. So the legislation does a few things. First of all, it's been in the works now for well over a year. So hopefully, we're coming to the end of the process. There is some proposed language that you can find. I think if you do a search on the Internet right now, you can see some of the proposed language that's currently contemplated. There are a few things that are still yet undetermined that aren't agreed upon. But in general, the concept is you stop buying and then you have a period of time by which you have to remove and replace all of the offending vendor equipment in your network. The legislation itself and getting that put into law, I think, of course, is a positive catalyst just for ongoing business. The dates associated with full replacement, like I said, aren't decided yet. But right now, I think it's generally contemplated to be around 2035 to actually have all the equipment ripped out. But no new equipment, it would be effectively right away.

William Dezellem

And given the pace of legislation can move, even though it's a year into the process, is the implication that there may be another year or more? Or do you see momentum behind this? And apologies if I haven't close of attention to just how this is picking up or not.

Thomas Stanton

Yes. No problem. I would tell you all of our investors that this is important legislation for ADTRAN. I mean we are the biggest beneficiary of this legislation being approved. You're right about the -- it being a -- it is an EU rule, right? So it's not just one country. So there are multiple countries that have a vote. The current thought is, and I would say that that's kind of the way that the participants are viewing it is that it will be done early next year. But having said that, this is multi-government legislation. So it's not -- it's something you wouldn't want to bet on. That's kind of -- but that is literally the current time frame is actually first half of next year.

William Dezellem

And taking this one step further, the stop buying the high-risk vendors, how much revenue do the high-risk vendors currently earn each year? And really the spirit of that question is trying to understand the magnitude that we're talking about here at this point in, I guess, in 2026 and '27.

Thomas Stanton

Yes. So we actually commissioned an independent study to actually look at that. And I think it was -- yes, in 2025, that number was somewhere in the mid-$800 million directly related to our equipment on an annual basis. And I don't think that, that's -- I would still have to think we're in the same ballpark at this point. And the bigger piece, of course, is if they do get this legislation or when they do get this legislation approved, there's a solid $10 billion worth of equipment that's got to be replaced.

William Dezellem

And then given that we're only 35 minutes into the call, I am going to ask you to dive into more detail, if you would, on the data center activity in terms of what you all are doing today and then the magnitude of the opportunity that your new products could lead to if they are as successful as you think they will be?

Thomas Stanton

Yes. So what we're doing today is mainly data center interconnect, right? So it's actually connecting data centers to data centers. So I guess that would be interconnect. And that's using our ongoing optical products. I will say that it's being -- it's also impacting our carrier sales. So we're seeing some really good activity here in the U.S. and in some portions of Europe, where you have carriers that are really trying to make sure that they upgrade their network to be able to meet data center traffic requirements and one of them here in the U.S. is very vocal. And our business is definitely picking up because of that. But we also are in data center upgrades where the hyperscalers themselves are upgrading, and we have one customer that, of course, is materially larger than the rest.

That's ongoing. We expect that actually -- that business to continue to grow meaningfully. The new intersection point for us is these 2 chips. The LPO is huge. We have strong interest in that product. And I'm trying not to overhang myself here, but we have strong interest in that product. And that product is not singular. We have different versions of that product that will be coming out over the next probably 3 years or so. But we've got very good interest on our very first. So it's anything that has to do with hyperscalers typically -- it's typically tens to hundred millions of dollars. So it's -- it would be very beneficial for the company when we're successful there.

William Dezellem

And one additional question, and I do promise I'll step away. The interest that you said is very strong. Is that coming from 1 or 2 potential customers? Or is it much more broad-based than that?

Thomas Stanton

On the lightweight, it is really large data center owners. So there's a finite number of them. On the Quattro, it is pretty much across the board. It's carriers, it is hyperscalers, it's neoscalers, it's MOPHY-type customers. So it's across the board. It's a much broader product. It's a much broader customer set.

Operator

I think we are -- I see the question queue is empty at this point. So thank you very much for everybody joining us, and we look forward to talking to you next quarter.