高塔半導體 (TSEM) 2026 年第三季法說會：矽光子成長與第四季 6 億美元財測指引

重點摘要

• 高塔半導體 (Tower Semiconductor) 公布 2026 年第三季營收創下 5.26 億美元的歷史新高，較 2026 年第二季成長 14%。淨利達到 1.2 億美元，淨利率為 23%。
• 矽光子 (SiPho) 業務創造了約 2.4 億美元的季度營收。管理層預期，年底的投片產能將達到支援第三季出貨量水準的約兩倍。
• 公司預測 2026 年第四季營收中位數約為 6 億美元。若達到該水準，將使 2026 全年營收達到 20 億美元，並奠定 24 億美元的年化營收跑率 (run rate)。
• 高塔半導體正在執行一項 9.2 億美元的矽光子 (SiPho) 與矽鍺 (SiGe) 產能計畫。截至第三季已支付約 65%，管理層預期投資將在一年內獲得完全回報。
• 客戶的產能承諾現已延伸至 2029 年。管理層表示，其已宣布的 200 公釐 (8 吋) 與 300 公釐 (12 吋) 擴產計畫所涵蓋的產能，均已透過已簽署的協議及處於後期談判階段的合約被預訂一空。
• 高塔半導體更新後的長期商業模型目標為年營收 36 億美元、毛利率 45%、營業利益率 38% 以及淨利 12 億美元，惟須視產能利用率、定價、驗證與執行等假設條件而定。

核心財務業績

業務與營運表現

矽光子 (SiPho) 仍是高塔半導體的主要成長動力。該業務在 2026 年第三季貢獻了約 2.4 億美元。產能正於四個製造據點同步擴展，預計年底產能將達到支援第三季營收水準的約兩倍。

管理層預期更高的投片率將轉化為 2027 年第一季與第二季的顯著出貨量放量。包含 Track-1 300 公釐 (12 吋) 擴產計畫在內，矽光子投片產能目標是在 2027 年第三季前達到支援 2026 年第三季出貨量水準的三倍。

高塔半導體也正與 10 多家主要客戶合作開發近封裝光學 (NPO) 產品。管理層表示，NPO 透過結合發射與接收功能，能增加高塔半導體的矽含量，並可擴大對矽鍺 (SiGe) 跨阻抗放大器 (TIA) 與驅動器的需求。大多數 NPO 流片均在 300 公釐 (12 吋) 平台上進行，包括 16 通道與 32 通道設計。

公司正將更多材料整合至其矽光子平台上。為 NewPhotonics 進行的初期生產透過晶粒至晶圓 (die-to-wafer) 貼合技術整合了磷化銦雷射。高塔半導體也正將該製程導入包含磷化銦調變器與鈮酸鋰在內的技術，目標鎖定下一代單通道 400 Gbps 的產品。

矽鍺 (SiGe) 的需求由三座 200 公釐 (8 吋) 晶圓廠與一座 300 公釐 (12 吋) 晶圓廠進行供應。高塔半導體表示其 H6 SiGe 技術的設計導入 (design-in) 活動強勁，並在本季度宣布與 Semtech 達成產能協議。

跨 200 公釐與 300 公釐平台的 BCD 製程技術需求年增約 10%。管理層將此成長歸因於 AI 與雲端資料中心，以及行動裝置與汽車應用。與戰略客戶簽訂的長期協議預計將使 2027 年的 BCD 綜合平均售價比 2026 年高出 20%。

位於義大利阿格拉特 (Agrate) 的 300 公釐產能走廊創下了 RFSOI 季度營收歷史新高。高塔半導體也報告了 Fab 10 產品流片數量創紀錄。在 CMOS 影像感測器 (CIS) 與顯示器領域，公司開始生產用於高階攝影機的新型快門感測器，並正為預計於 2027 年放量的矽基 OLED (OLED-on-silicon) 準備產能。

所有 200 公釐晶圓廠在第三季的產能利用率達 85%-90%。高塔半導體的 300 公釐晶圓廠則持續保持滿載，高於公司 85% 產能利用率的長期模型設定。

管理層業績展望

高塔半導體預測 2026 年第四季營收中位數約為 6 億美元。若達成此目標，2026 全年營收將達到 20 億美元，高於 2025 年的 15.7 億美元。

管理層表示，公司的需求與投片計畫並未呈現典型的第一季季節性放緩。相反地，高塔半導體預期將持續成長，並以 2027 年開局營收高於第四季展望所隱含的 24 億美元年化營收跑率為目標。

2026 年第四季矽光子的年化營收跑率預計將突破 11 億美元，並目標在 2027 年第一季進一步成長。

針對邊際毛利 (incremental gross profit)，財務長 Oren Shirazi 維持 55%-58% 的長期預測。他提醒道，第三季較高的邊際毛利率並非可持續的基準，因為隨著產能投入營運，新增設備將開始提列折舊。

高塔半導體更新後的長期模型目標如下：

該模型假設在 9.2 億美元的矽光子與矽鍺投資全面裝機與完成驗證後，晶圓廠以 85% 的產能利用率運作。該模型亦取決於未來的晶圓價格、設備與製程驗證、客戶產品驗證、成本以及完成時間表。

風險與關注焦點

• 隨著 9.2 億美元投資計畫下的其餘設備投入營運，新增的折舊費用將降低邊際利潤率。
• 長期模型仰賴設備的順利且即時安裝、製程驗證以及客戶產品驗證。
• NPO 出貨成長取決於市場採用率。管理層的目標是在 2027 年下半年達成顯著的 NPO 出貨，但也強調時間點仍具不確定性。
• 新的矽光子與異質整合機會（包括整合雷射）未包含在已披露的基於產能之營收框架中，且仍取決於客戶採用情況與量產執行的進度。
• 矽光子與矽鍺領域的競爭正在加劇。管理層預期隨著市場擴大，將有多家供應商參與，不過其認為截至 2029 年的客戶協議將支撐高塔半導體的市場地位。
• Track-2 的建設預計將於 2028 年開始，目標是在 2029 年釋出實質產能。管理層表示提前一個季度完成是有可能的，但無法保證。

分析師問答環節亮點

管理層表示，一旦產能達到支援 2026 年第三季出貨量水準的三倍，已宣布的矽光子擴產計畫可支援約 7.2 億美元的季度營收跑率。基於 2027 年第三季的投片產能，這相當於略高於 28 億美元的矽光子年化跑率，且此數據不包含整合雷射機會與其他潛在的新增項目。

高塔半導體拒絕將其 36 億美元的營收模型歸類為保守。執行長 Russell Ellwanger 將其描述為一個積極的目標，但表示管理層對達成該目標充滿信心。

在定價方面，管理層表示長期合約包含價格目標。由於 NPO 具備更高的整合度與更先進的技術，預計其將重置價格曲線，起始價格點會高於傳統的可插拔 (pluggable) 產品。

高塔半導體計劃隨著時間推移，將其日本 300 公釐 (12 吋) 晶圓產能從每月約 10,000 片擴大至約 40,000 片。管理層指出，一片 300 公釐晶圓的面積約為 200 公釐 (8 吋) 晶圓的 2.25 倍。

公司表示，其 NPO 客戶均為現有的可插拔收發器 (pluggable-transceiver) 客戶，目前正擴展至更先進的項目。管理層預計，可插拔與 NPO 的需求都將使高塔半導體受惠，因為公司在這兩種架構中均已建立穩固地位。

法說會逐字稿全文

完整財報電話會議逐字稿

管理層陳述

Operator

Good day, and thank you for standing by. Welcome to the Tower Semiconductor Third Quarter 2026 Earnings Conference Call. [Operator Instructions] Please be advised today's conference is being recorded. I would now like to hand the conference over to your first speaker today, Noit Levy, Chief People, Communications and Investor Relations Officer. Please go ahead.

Noit Levi-Karoubi

Hello everyone and thank you for joining us. Welcome to Tower Semiconductor Third Quarter of 2026 Financial Results Conference Call. With us today are Mr. Russell Ellwanger, Chief Executive Officer; Dr. Marco Racanelli, President; and Mr. Oren Shirazi, Chief Financial Officer.

Before we begin, please note that certain statements made during today's call may be forward-looking and subject to risks and uncertainties that could cause actual results to differ materially. These risks are detailed in our SEC filings, Form 20-F and 6-K as well as filings with the Israeli Securities Authority, all available on our website. Tower assumes no obligation to update forward-looking statements.

Our third quarter 2026 results are prepared in accordance with U.S. GAAP. Some data presented may include non-GAAP financial measures as defined under SEC Regulation G. Reconciliation to GAAP figures and full explanations are provided in today's press release and financial tables. For your reference, a supporting slide deck is available on our website and integrated into this webcast.

With that, I'll turn the call over to our CEO, Mr. Russell Ellwanger. Russell?

Russell Ellwanger

Thank you, Noit. Hello, everyone. Thank you for joining our call. We have much to discuss. We will speak not only to the strength of our current performance, but very importantly, on what comes next, namely, a great business and financial strength with which we will be opening 2027, the visibility we have into the future demand, the realization of investments to fully capture that demand, 2028-2029 further customer commitments and our significant growth in customer line value creation activities for 2029 and beyond. Indeed, a long-term value creation campaign with boots on the ground, providing real-time measurable financial and strategic achievements.

During the third quarter, we continued to deliver record results while building upon our technology leadership and expanding capacity with increasing customer commitments to support a significantly higher level of revenue and earnings in the quarters and years ahead. At the center of this growth are our Silicon Photonics platforms where accelerating AI infrastructure demand, expanding customer program and long-term capacity commitment continue to increase Tower's scale and strategic position. This momentum is complemented by strong execution across our broader portfolio of differentiated analog semiconductor technologies. We are entering the strongest growth phase in our history. Our third quarter performance demonstrates that the strategic decisions we have made are translating into record revenue, expanding earnings power and significant visibility into the years ahead.

Third quarter revenue was $526 million as compared to $396 million in the third quarter of 2025 a 33% year-over-year growth, resulting in a net profit of $120 million or 23% net margin, up from 14% net margin a year ago. Our fourth quarter 2026 midrange revenue guidance is $600 million, which achievement would result in a 2026 full year revenue of $2 billion against the full year 2025 revenue of $1.57 billion. Achieving a $2.4 billion revenue run rate is most significant as our demand and start plan do not indicate the typical industry's first quarter seasonality. To the contrary, we expect to see continued growth. Hence, the starting point of 2027 is targeted to be higher than the $2.4 billion annual begin run rate and greater than 45% increase as compared to the first quarter of 2026 revenue starting point. Later in the call, we will share details that underscore the strong quarter-over-quarter growth we are driving in 2027.

Moving into our businesses. Please refer to Slide 4 as referenced for the Q3 revenue breakdown by technology. The third quarter was another quarter of strong operational execution by our teams, both in the expansion and re-purposing of 4 of our 5 factories towards substantially higher Infrastructure mix with a background of growth activities in all of our core businesses. Our Silicon Photonics business is experiencing exceptional growth. To support accelerating customer demand, we are expanding capacity across 4 manufacturing sites, a business that contributed approximately $240 million of Silicon Photonics revenue in Q3 2026 alone. With substantial capacity coming online this quarter, we expect year-end capacity to be approximately double the level that supported the Q3 2026 Silicon Photonics revenue. This positions us for significant revenue ramp in the first and second quarters of 2027 as the new start rates move through the fab and ship, establishing a new substantially higher, sustainable baseline for the quarters ahead and thereby also achieving a full ROI within 1 year for the previously announced $920 million CapEx investments.

Additionally, the execution of our released Track-1 300-millimeter expansion is progressing ahead of schedule with all tools ordered and slotted with full production starts targeted for Q3 2027, including the Track-1 300-millimeter expansion, our Silicon Photonics start capacity is targeted to increase by 3x versus the level that supported the Q3 2026, 250 million SiPho or 240 million SiPho shipments. Importantly, this capacity is spoken for. Through a combination of signed agreements and contracts in advanced stages of negotiation, we have secured customer demand commitments for both 200-millimeter and 300-millimeter production and now having extended it to include 2029, at which time our incremental Track-2 plan capacity ramp is well underway, ultimately, quadrupling our 300-millimeter Japan capacity. Of note, all expansion is planned to continue to be self-funded.

In addition to serving the rapidly growing market, largely of pluggable scale-out transceivers, we continue many activities with Tier 1 customers that take advantage of new opportunities expanding our served market. The most significant of these is the change in form factor towards near packaged optics or NPO that promises to add scale up opportunities to our clearly established leading position in the scale-out market. In the view of most analysts, NPO scale-out can double the optical transceiver market. To this end, we are prototyping NPO products with more than 10 of our top customers. NPO can more often integrate, transmit and receive functions in our SiPho die, expanding our silicon SAM versus the traditional pluggable market, where normally only transmit function is built into the SiPho die. NPO can also improve our position in the electrical IC SiGe market as most implementations of NPO will replace the DSP to improve cost, power and latency with SiGe TIAs and drivers and at that with greater footprint due to the increased incorporation of [indiscernible] functionality, expanding our SiGe market opportunity.

To this end, we have announced this quarter a substantial capacity agreement with Semtech and are seeing further unprecedented growth opportunities for our SiGe business. Longer term, we are building the capability to integrate new material into our SiPho platform to further expand our served available market. In one such example, we announced initial production for NewPhotonics with a new technology where we integrated Indium phosphide lasers within our silicon photonics die. This is done through a die to wafer attached in indium phosphide epi-chiplets, further processed in our factory to create a final die that incorporates indium phosphide lasers within our Silicon Photonics. As this technology ramps, we gain access to the laser market and additional multiples of increase in serviceable market.

This die-to-wafer technology is now being deployed with other materials such as [indiscernible] lithium niobate to provide 400 gigabit per lane next-generation differentiated products to our Tier 1 lead customers. Beyond the optical transceiver market, our SiPho platform, combined with heterogeneous integration capability emerged as a go-to solution for silicon photonics-based optical circuit switches. Silicon photonics-based solid-state OCS are emerging as a strong contender against MEMS or other niche solutions because of fast nanosecond level switching speeds, the CMOS manufacturing scale and compatibility with the CPO-NPO ecosystem for tighter system integration.

Finally, we have been a significant player in coherent SiPho products for [ Marvell ] used in data center connect applications, as we have discussed in the past. This technology is now being evaluated for inside data center scale-across applications, which could as well significantly expand the sector of the SiPho market we already have strong position and differentiated technology. Longer term, we are also investing in several scale-up and scale-in [indiscernible] technology, which promised high total bandwidth by creating large parallel links at lower speed through the use of micro LEDs and/or VCSELs. We have several active engagements ongoing with Tier 1 customers with a potential of high volume and a new facet of the optical communication market.

As stated earlier, NPO is providing an expanding market for silicon germanium TIA and drivers, resulting in unprecedented customer demand being served now from 3, 200-millimeter factories and 1, 300-millimeter factory. In tandem with our capacity expansion, we continue to advance our technology road map and are happy to report that our H6 next-generation silicon-germanium technology is seeing strong design-in momentum already in production for the most demanding products. The strength of our technology was also reflected recently when Semtech, a strategic silicon-germanium partner [indiscernible] for its optical interconnections solutions. Our 300-milliter capacity corridor in Agrate, Italy delivered its highest RFSOI revenue quarter to date. We also set new records for the number of product tape-outs for Fab 10 demonstrating the strength of our long-standing customer relationships and the exceptional execution by our Fab 10 team to deliver the technology road map, extending the design win momentum on our industry-leading 300-millimeter RFSOI platforms. We expect this trajectory to continue over the coming quarters.

Demand for Tower's BCD technologies continues to strengthen across both 200-millimeter and 300-millimeter platforms, with about 10% year-over-year growth. This growth is driven not only by the rapid expansion of AI and cloud data centers, but also by growing customer success in mobile and automotive applications. This momentum is reflected in increasing customer commitments, with several strategic accounts entering into long-term agreements at higher ASPs, resulting in forecasted 20% higher blended prices for 2027 versus 2026. In CIS and displays, we have started production of a new [indiscernible] shutter sensor for high-end video cameras with one of the DSLR MLC industry leaders. Together with our lead customers, we are driving next phase of business growth, building demand towards the significant additional capacity that will come online with the Uozu expansion. We're also working closely with our OLED-on-silicon display customers to provide capacity for their expected steep ramp in 2027.

Looking at utilization. During the third quarter, all of our 200-millimeter fabs operated at utilization levels of 85% to 90%. Our 300-millimeter 12-inch fab continues to be fully utilized well above our 85% utilization model. And as stated, substantial capacity is being added to all sites in accordance with extreme growing demand.

To summarize, we entered the second half of 2026 with strong business momentum supported by broad customer demand across our differentiated technology platforms and continued high utilization across our entire manufacturing footprint. Importantly, the strength we see extends beyond the current quarter, well beyond. Customer engagement, design activity and demand ability continue to reinforce our confidence in growth opportunities ahead, enabled by the unique significant value, which each of our specialized analog technologies provide. Our prospects are very good and even getting better with all of our key business units experiencing increased demand with expected revenue growth throughout 2027. SiPho and silicon-germanium are experiencing unprecedented growth having achieved majority share of pluggable scale-out market. We are not only enjoying the substantial market growth of this segment, but by having close partnerships and cooperations with the market leaders we are enjoying rich developments, increasing our served market, targeting to continue proliferating of our differentiated technologies into multifaceted applications. The number of advanced NPO engagements is just one of the many indicators of re-partnerships capturing the future.

As a summary point, our Q4 '26 SiPho annualized run rate is expected to exceed $1.1 billion with significant target further increases in Q1 '27. We will end this year in an extremely exciting position.

With that, I'd like to turn the call to our CFO, Mr. Oren Shirazi. Oren, please.

Oren Shirazi

Thank you, Russell. Hello, everyone. The third quarter marked another significant step-up in Tower's financial performance with record revenue, record profit and strong operating leverage across the P&L. Revenue increased 36% year-over-year to $526 million while gross profit increased by 92%. Operating profit increased by 2.5x and net profit reached $120 million representing 23% net margin or 2.2x as compared to the third quarter of 2025. These results demonstrate our accelerating growth in translating in substantial earnings expansion. They also reflect the increasing contribution of our differentiated, higher-value technologies and the strength of our operating model as we scale.

Importantly, we are entering this next phase of growth from a position of financial strength, evidenced by our strong balance sheet and cash position. The customer long-term contracts that secure our CapEx investments will yield quick ROIs by providing increased visibility into our customers' future demand and high capacity utilization. I will now review the quarterly financial performance, followed by our balance sheet, cash flow and capital investment plans and conclude with financial framework underlying our long-term business model.

Looking into the P&L. Revenue for the third quarter of 2026 was $526 million, a record, in the company's history, representing 14% quarter-over-quarter growth as compared to $460 million in the second quarter of 2026 and $130 million higher year-over-year or 36% year-over-year growth. Gross profit for the third quarter of 2026 was a record $179 million, an increase of 92% compared to $93 million in the third quarter of 2025. Operating profit for the third quarter of 2026 was $127 million, 2.5x the operating profit in the third quarter of 2025. Net profit for the third quarter of 2026 was $120 million, reflecting 23% net margin. Net profit is more than double the $54 million in the third quarter of 2025. Earnings per share for the third quarter of 2026 were $1.06 basic and $1.05 diluted, as compared to $0.48 basic, and $0.47 diluted in the third quarter of 2025.

Moving to our balance sheet. Prepayments received from SiPho customers mostly towards 2027 capacity reservation, including the $290 million previously announced prepayments received in the first quarter of 2026, are presented in the balance sheet of liabilities under short- and long-term customer advances and are included in the 2026 cash flow report as cash flow from operating [indiscernible]. Our balance sheet continues to be very strong, as evidenced by the following indicators and financial [indiscernible]. As of the end of September 2026, our assets totaled $4 billion, primarily comprised of $1.8 billion in net fixed assets predominantly fab machinery and $2 billion of current assets. Shareholders' equity reached a record of $3.2 billion, representing 80% of total assets. Our current ratio is strong at about 3.7x and our increased long-term asset balance mainly reflects the [ Newport ] fab lease prepayment as well as other prepayment.

Hedging. I would like now to describe our currency hedging activities. Regarding Japanese yen, since the majority of TPSCo's revenue is denominated in yen and the vast majority of TPSCo's cost costs are also in yen we have a natural hedge over most of our Japanese business and operations. To mitigate part of the remaining yen exposure, we execute Zero-cost cylinder [indiscernible] to hedge currency fluctuations. Hence, while the yen exchange rate against the dollar may fluctuate, there is a limited impact on our margins. Regarding the Israeli shekel, while we have no revenue in this currency, a portion of our cost in Israel is denominated in shekels, so we also hedge a large portion of that currency risk by entering into Zero-cost cylinder transactions to mitigate the exposure. And while the shekel exchange rate against the dollar may fluctuate, as seen over the past few quarters, during which the shekel appreciated strongly against the dollar, the impact on our margins is limited.

Now moving to CapEx investment plan. As previously announced, in order to support the increasing SiPho and SiGe demand, we are executing a $920 million investment plan to expand capacity and capability of SiGe and SiPho equipment across our 8-inch Fabs in Israel, Newport Beach and Texas as well as our 12-inch Uozu fab in Japan. These investments are progressing ahead of schedule in terms of on schedule tool, technology, process qualification, equipment arrivals and ran plans and are supported by customer contracts and prepayment. Approximately 65% of this $920 million CapEx investment has been paid to date and is included in our cash flow for investing activities for the reporting period through the third quarter 2026 while the remaining 35% is expected to be paid during Q4 '26 and H1 '27.

As Russell noted earlier, we expect 1 year full ROI from this investment. In addition, we are executing on the Track-1 Japan 300-millimeter CapEx plan by issuing purchase orders for 12-inch SiPho capacity and capability tools that will significantly increase our 12-inch SiPho capacity. These CapEx payments are scheduled for 2027 and 2028.

Business model. As recently announced, the company updated its business model which now includes the planned investment in our Arai facility, formerly Fab 6 to re-purpose it for 12-inch SiPho wafer manufacturing, thereby maximizing the company's 12-inch wafer output in Japan, supported by METI, Japan Ministry of Economic Trade and Industry. The updated model is based on forward-looking operational business and financial assumptions, including the assumption that all fabs will operate at 85% utilization post full installation and qualification of the $920 million in SiPho and SiGe CapEx investments. We continue to ramp up SiPho and SiGe capacity and capability in our Fab 2, 3, 7 and 9. And in addition, we are working to re-purpose the Arai facility and qualify the new equipment tools to be installed there.

Assumptions regarding forward-looking wafer selling prices for existing and future products and flowed and forward-looking assumptions regarding cost and completion dates for the successful qualification of tools to be installed, process technologies to be qualified and customer products to be qualified and ordered from the company. Under this updated financial model, we assume -- we target $3.6 billion in annual revenue, which is $760 million higher than the [indiscernible] model target, reflecting 27% revenue increase. $1.63 billion annual gross profit resulting in 45% gross margin as compared to 39% in the previous model. This annual gross profit is $510 million higher than the -- than the previous February 2026 model, representing 67% incremental gross profit derived from the incremental $760 million revenue, reflecting our enhanced product mix.

We also target $1.38 billion annual operating profit, resulting in 38% operating margin as compared to 32% in the prior model. And we target $1.2 billion in annual net profit resulting in a 33% net margin as compared to 26% in the prior model and 20% in the second quarter of '26 and as compared to 23% in the third quarter of '26. This annual net profit of $1.2 billion target is $450 million higher than the prior February 2026 model, representing 59% incremental net profit derived from the incremental $760 million revenue, reflecting our enhanced product mix. That concludes my prepared remarks.

Now I'd like to turn the call back to the operator so we can take your questions. Operator?

Operator

[Operator Instructions] We're going to take our first question from Cody Acree from StoneX Financial.

分析師問答

Cody Grant Acree

Congrats on the progress. It's great to see. Maybe, Russell, thanks for the exponential increases that you gave in capacity additions. Maybe if you can just help me to just work out the math there. Can you just level set us with the $920 million of CapEx program now firmly on track and ahead of schedule. Can you level set your fiscal '28 capacity support now for SiPho and SiGe from a revenue standpoint?

Russell Ellwanger

I would prefer not to give an exact target on revenue, but I can give you a multiplication against the previous quarter revenue as far as raw capacity. So as stated, the previous quarter, Q3 was $240 million of SiPho revenue. We will finish Q4 with 2x the start capacity that drove that revenue in Q3. So 2x the starts of the Q3 shipments. That revenue will become the run rate in Q2 when there's a full Q1 of that start rate and the December, January, February all shipped within Q2. And in Q3, we will add another 1x of that. So 3x of the Q3 shipments through the expansion that we are doing in the Fab 7 [indiscernible] factory. That will be plus/minus the SiPho capabilities off of investments that we have already announced. We are evaluating and will most likely be announcing shortly additional SiPho announcements where we've been able to identify further areas for expansion, predominantly in two of our 200-millimeter sites. And we have not yet released that. We've not yet talked about the amount of CapEx nor the amount of capacity that would increase, but it's another significant tranche on top of it.

But the capacity, if you go against the the multiplications that I talked about, that would give somewhere of a $2.85 billion to $2.9 billion SiPho run rate as of Q3 start rates. However, that is really not 100% accurate, because it does not include what we're doing with integrated laser. So that is outside of it and on top of it and a substantial forecast in and of itself. So if that's kind of level sets the baseline of SiPho revenue for everything that has been announced and will be fully invented on starts as of -- by plan as of the third quarter of 2027 is somewhere a bit above the $2.8 billion run rate.

Cody Grant Acree

That's very helpful, Russell. That's a great place to start. Maybe with all of that, then it sounds like your summer update for your business model, your $3.6 billion than to a $1.2 billion in net profit that Oren just ran through may already be too conservative. Would you agree with that?

Russell Ellwanger

No comment. No, I think it's an aggressive target. But certainly, we wouldn't be giving a target to -- the Street that we don't have very, very high confidence that we will meet.

Cody Grant Acree

And then lastly, if I may, for Oren, you had a very, very high incremental margin this quarter. First time we've seen it in the 62%. How sustainable do you think that level? Is that a new baseline? Or will that fluctuate as you get new tools in and start depreciating that capacity?

Oren Shirazi

Yes, exactly like you said, this quarter, still not -- all the tools started their depreciation schedule. From the fact that I said that 65% of the tools were paid, so you can maybe understand that approximately 65% of those $920 million tools started depreciation and the other not, and there is the second tranche of Track-1. So we will experience additional depreciation coming in later. And this is why the long-term model, I don't think has the 65% incremental gross margin that we saw this quarter and has like 55%, 58% annually. And so we repeat our same forecast of 55% to 58% incremental gross profit.

Operator

And the question from line of Vijay Rakesh from Mizuho.

Vijay Rakesh

Just a question on the SiPho, SiGe. Obviously, you're doing multiple versions of NPO. You have VCSELs, laser internal -- integrated lasers. Can you talk to how you're seeing the pricing and contracts line up as you look at some of the newer layouts in NPO going out? And a quick follow-up.

Russell Ellwanger

It's a very, very interesting question and a good question. Our long-term contracts include pricing targets. There's been maybe some inputs from time to time that curtails price increases, but it does the opposite as new players possibly come into the game. One way that they try to get into market is by deeply discounting. Now those that deeply discount typically do it because they don't compete technically. But what our contract do is that they set a bar of a starting position, not that if market really with demand, some lowering of prices that we wouldn't adhere to being reasonable with our customers. But our pricing is set very, very well in our long term contracts and pricing targets are within all of our contracts.

NPOs, I don't want to give an exact number for NPO. I don't think that, that would be a proper thing to do. But NPO certainly, days with our trend of any new platform resets a pricing curve as far as getting a higher ASP. So the way that, over time, anybody protects high margins is by being the fastest to deliver products and taking advantage of being the first in the market to help a customer grow their market by being the first with their specific end applications. So I just want to say that our pricing is solid, it's good and it's very sustainable through the way that we have gone about our learn term agreements we'll have a higher starting point than a pluggable and that makes sense as it's a more advanced technology platform. Our pluggable start point wasn't bad and isn't bad right now, as is seen by the incremental margin drop down.

Vijay Rakesh

That's very helpful. I think one of the things that were coming up at OCP U.S. Obviously, [indiscernible] are planning to talk more about [indiscernible], 400G per lane. As you look at your NPO demand backlog, are you seeing more engagements start to line up at 3.2T. Is that a higher dollar more complicated layout on SiPho, how do you look at that mix and the opportunity?

Russell Ellwanger

The answer to both parts of your question is yes.

Vijay Rakesh

Is the mix higher at 3.2T?

Russell Ellwanger

Yes.

Operator

And the question comes from the line of Simon Coles from Barclays.

Simon Coles

There's a lot of chatter about competition on SiPho given how well you guys are executing. You started ramping 12-inch yield, I think, late last year, early this year. Can you just give us an update on sort of that progress and how that's sort of helping you maintain your leadership?

Russell Ellwanger

I'm not sure I 100% understand your question. I'll answer what I think it is. And if you need further clarification, please ask. We indeed started shipping revenue volumes in the first quarter of 2027 at extremely high yields, really high yield. Within our 300-millimeter platforms, we have very special applications. We have an amazing [indiscernible] capability that I think is absolutely best in the world. So we've added multiple new types of platforms. Most of the NPO tape-outs that we're doing are being done at 300-millimeter. And those are 16 lanes, 32 lane applications right now.

And the pluggable that we're doing also just has incredible performance, substantially better on insertion loss than I think anything else can be -- even coming close to. So if that answers your question, the ramp that we're doing at 300-millimeter is quite substantial. We have stated that from where we'll be at in Q1 on 300-millimeter which on equivalents is somewhere about 10K, 200 millimeter. By the end of the year, we'll be increasing that by the third quarter of the year. We'll be increasing that by 3x. And then incrementally increasing as well, in 2028, ultimately, very, very big increase as the new shale [indiscernible] at the end of 2028. Did that answer your question? I hope it did.

Operator

And the next question comes from the line of Sahej Singh from Stifel.

Sahej Singh

Also, your framing was very helpful. So maybe let me build on that. You said 3Q revenue, SiPho revenue was about $240 million. That adds -- and that -- the 3Q '27 step adds another 1x on top of that capacity through Fab 7 and Arai.

Russell Ellwanger

Triples it.

Sahej Singh

But I guess by my math, that would be roughly $960 million a year. So that's $200 million more than the $760 million Track-1 adds to the '28 model. So first, if I'm understanding that correctly, is that all Track-1? Or does it include remaining capacity from the $920 million program. And then second, in those same units, has Track-2 ramps towards the quadrupling of Japan's 300-millimeter? How -- roughly how many multiples does Track-2 adds? Is that another 1x, 2x? Or would it be more than that as a floor?

Oren Shirazi

For the first question, what Russell said is that today, including -- I mean, today, the total revenue that the shipments of Q3 reflected is $240 million. And tripling it means $720 million, not $960 million, so $720 million per quarter run rate towards I think it was quoted answer -- a question about 2028, [indiscernible] at Q3 27 capacity baseline can support capacity $720 million a quarter. So time 4, it's $2.8 billion, which Russell mentioned, right? Not $960 million.

Russell Ellwanger

There's additional SiPho activities that are not included in that. So I'm not saying that we are gated at the $2.8 billion. There's integrated lasers or many things that do not include into that. If we're talking specifically about the quadrupling, we're quadrupling what we're talking of the present capacity, which is somewhere about 10,000 wafer per month in our Fab 7, 300-millimeter, and that will be quadrupled to 40,000 wafer per month, approximately. I mean we don't know the exact numbers as of yet, the build hasn't been done. But what that does then is an extreme increase if it was all on SiPho because you're putting 2.25 against each 300-millimeter wafer as you compare to 200-millimeter. So if you were going to 40,000 200-millimeter equivalents, you're close to 100,000 wafers.

Sahej Singh

That's very helpful. I guess, Russell, the [indiscernible] sort of math. You're pointing to the 2.25x die size increase on the 300-millimeter which makes sense. But Track-2 is also predominantly the NPO which means the content uplift is also there and that sort of changes the like-to-like math. I think that most investors are trying to do. Could you maybe give some color on that as well?

Russell Ellwanger

The price per wafer, again, had stated, as we go to NPO, it resets the price target, right? It resets the pricing curve. What bigger chips do as they create more silicon demand. So the NPO will -- the more lanes that you put into it, the more silicon you're going to need to ship the same amount of wafers. [indiscernible] change the cost of the silicon other than the higher the integration, the newer the technology, the higher the starting price point.

Richard Shannon

Yes, correct. I mean I'm on the same page. I think it's mostly about sort of sizing "higher". But one last one for me and it's maybe on timing. I guess you said Track-2 is well underway. And I believe Fab 7.

Russell Ellwanger

I said Track-2 will be well underway in 2028. '28, we'll start with buildings in '29, when we have our last contracts that we are signing now, again, we extended our capacity contracts to 2029, which is a very, very good thing for us. Tremendous visibility and commitment. So we were -- had stated the 1-year ROI for the $920 million investment that we did predominantly for 200-millimeter expansion. By having customer commitments, it really adds their confidence on short-term ROI. And the ROI is an extremely aggressive model.

Sahej Singh

The question just being Track-2, I think you take ownership of Fab 7 April 1. Could that start earlier, meaning, I guess, simply put, could Track-2 output come in sort of second half '28? Or is it almost all '29 driver?

Russell Ellwanger

Could as possible, what we're really believing is the amount of time to build we're quite confident that the '29 will be [indiscernible]. If it can be accelerated by a quarter, that's certainly our target, right? I mean there's any investment, the quicker that you can start shipping, the quicker you place your customers and the quicker you please your investor.

Operator

And the question comes line of Mehdi Hosseini from Susquehanna International Group.

Bastien Faucon-Morin

This is Bastien filing for Mehdi. I guess backing on the competitive landscape question into 2027. Given the ramp of one of your major competitors in terms of [indiscernible] capacity, how should we think about the market share in Silicon Photonics and silicon germanium for 2027?

Russell Ellwanger

Let me ask Dr. Racanelli to speak to that as he's directly interacting with driving the BUs and tremendous amount of customer contact interaction. And then if I have anything to add, I'll be happy to do so.

Marco Racanelli

Thank you, Russell. So it's hard always to pinpoint market share, particularly in the market that is growing as quickly and rapidly as this is growing, both on the silicon photonic and the silicon germanium side. And with a market that's growing at this rapid pace, there's obviously room for more than one player, and that's what we're seeing, as you described, a little bit of competition.

What I'd say is that we have -- we were the first to ramp in 200-millimeter to very high not only capacity but also very high yields with very good performance. And as was just talked about previously, we're also the first to ramp in 300-millimeter with very high yields and very high quality and performance. So that is giving us confidence in giving our customers the confidence to extend the LTAs as we talked about, into 2029. So we -- we're building capacity as quickly as we can and all that capacity is being spoken for almost immediately as we announce it, certainly well before it's even put in place with contracts from customers. So that's a tremendous position to be in.

We also continue to invest at the leading edge of technology with our customers, having many Tier 1 customers with us. We have incredible visibility over their needs of the future. And so putting in place technology that will pay dividends well into the future and not just today. So we anticipate not only short-term maintaining very good market share, but certainly, there will be others competing with us in the market. And with growth as it is, that's perfectly understandable and okay. But also in addition to the short term, our printing technology, 3.2T was talked about with both silicon indium phosphide, lithium niobate technologies and CPO, as we've talked about both on the Silicon Photonics side and also silicon-germanium side, putting technologies in place that will enable at least at the leading edge for us to maintain our competitive advantage and market share for the long term. .

Russell Ellwanger

Thank you, Marco. If I could just add, and I think this is a very important point to add. We really are engaged in billions of dollars of long-term agreements. Customers will never sign up to being the end of '26 now, not '27, not just '28 but 2029, unless they have faith that, number one, you can ship wafers on time at the highest quality. But secondly, that you will maintain a differentiated, more than competitive road map. What we are shipping now to lead customers for the most part will not be a very big portion of what we're going to be shipping in 2029. So the confidence that the customers have off of our relationship off of the programs that we're working on that are not just present, next, but generation plus 2 and in some cases generation plus 3 type programs, that's what gives us the assurance there will be competition. I mean it's a big market. It's a good market. It's a big enough market to afford competition.

But we're quite confident in maintaining our leadership position because we are leaders, working with the leaders on leadership programs. And I think that, that's the biggest thing that I can say what our market share will be. I mean, presently, I think it's extremely high. And one could check that with revenue. I mean, look at just the Q3 close to $1 billion run rate on SiPho revenue, I think there's -- you'd be hard pressed to find others that can come close to that. But breaking $1.1 billion. It's a pretty substantial place where we sit. But the biggest thing isn't the market share that we have now, it's the confidence in customers to sign up on long-term agreements through 2029 because those are committed. They're committed capacities. And they have to know us. We have to know them. There has to be a partnership and trust in order to go into those contracts. And that's what really for us gives us confidence of maintaining a very high market share, is that we are working truly with the best of the best on very advanced programs. That's it's a joke, I have, but it's really true. I mean, for [ Kevin Costner ] in the field of dreams, if you build it, they will come, this may be true. For us, we need very strong customer partners.

And if we have the best of the best in each market and we're working with them together on next-generation and generation + 2 technologies, when we build it, it will be sold because we are dealing with the leaders. And so again, [indiscernible] possibly. But to give a specific market share number, I know what I think it is right now to say what it will be at the end of '27, '28, '29, that would really be [indiscernible]. But to say that we will maintain a very high differentiated market share on high technological different products, I think that, that's very confident to be able to state.

Bastien Faucon-Morin

And I guess my follow-up was on another end market, the RF mobile. I think RF mobile is approaching the trough is in the low 50s this quarter. How should we think about the trajectory of RF mobile going into 2007? Given on the last call, you mentioned the tape-out activities.

Russell Ellwanger

Mentioned, what, I'm sorry, please?

Bastien Faucon-Morin

Sorry, I mean in terms of activities, how should we think about the trajectory of RF mobile?

Russell Ellwanger

For us, the trajectory is very strong. We had a bit of a dip in moving from Fab 7 to Fab 10, needing to re-qual multiple customers, because of re-purposing the Fab 7 really to focus on optical between the SiPho and the SiGe. It's going very well. We expect Q2, while Q1 was the highest shipments we've had. Q1 will be -- Q3 was Q4, Q1, Q2 should continue to grow, where Q2 will be absolute records. So I think it's -- for us, it's going extremely well.

Again, I don't have a specific feel of the exact market share percentage we are working with leaders and growing market share with the leaders. And so I believe we're in very good shape. I also am quite confident that on one of our programs isn't just a step function in capability. It's an absolutely opening of a new market capability itself. And that has not been formally released. We're working with at least one very elite customer on that type of figure of merit and it's really a breakthrough figure of merit. Marco, is there anything else you would like to add there?

Marco Racanelli

No, I think you said it. We don't have the market share that we do in SiPho and RFSOI. There are multiple players. But we continue to lead at the very leading edge. So the market share that we have tends to be the very leading edge, the cream of the creme, kind of piece of it, and that's what we're focusing on. We have limited capacity for RFSOI. Now we've dedicated Fab 10, and we use it for those leading components and leading technologies.

Operator

And I will go take our final question for today. And the question comes line of Richard Shannon from Craig-Hallum Capital Group.

Richard Shannon

Congratulations on great results again. Russell, let me follow up on a couple of comments you made on the call regarding SiPho and specifically in the laser attach side, where it's not encompass into the revenue model you've been talking about. Maybe you can talk about it in terms of -- in terms of a couple of different directions, one of which is how much content add do you have with laser integration or laser attach here? And over time, what percentage of your picks would you expect to have laser attach and what are, I think there's multiple ways you can do it. What kind of attach rate do you see over time?

Russell Ellwanger

It's really multiple directions in the market for integrated laser, it enters into a new served market that can be very big. There's multiple analysts that say that integrated laser will be a huge portion. There's others that really don't address it. I truly do not have a clear answer to it. Our elite customers there are very, very committed for growth, with very strong forecast. And again, those are numbers that are not yet within the model, but they can be very appreciable, especially as one could imagine, the -- there are appreciably higher ASPs, against a very good baseline to begin with. So it's a good market to get in.

The die attach though, is for more than just the indium phosphide laser. It's for an Indium phosphide 400G modulator that we've already plus/minus demonstrated, and for the lithium niobate modulators. So the die attach that we're doing encompasses many different things, and it could be a very big portion of the 6.4T million form factor of NPO or the pluggable with a 400 gigabit modulator.

Richard Shannon

Okay. I appreciate that perspective. My follow-on question is on the topic of NPO. You talked about greater than 10 customers on this call here. Maybe you can characterize kind of the customer set here across pluggable transceiver suppliers, historic ones, OEMs and even hyperscalers. And then how would you characterize kind of the competitive dynamics here? And ultimately, any qualitative commentary on market share that you expect in that product over the next couple of years? It would be great to hear.

Russell Ellwanger

The existing customers that we have for pluggables all of our NPO customers are a subset of that. So we're not doing NPO with anybody that we haven't done pluggables with. It's an extension or a newer capability for a broader market. So it's not a different customer set whatsoever. One of the reasons I believe that we have so much engagement is as stated, it's the confidence and trust that's been built up through the engagement with pluggables. So that was the first part of your question.

We expect to see an appreciable amount of NPO shipped in the second half of this coming year. That's our expectations and targets. I will it happen. It really depends on market adoption. But it's the same set of customers that are just doing more if you will, advanced programs with us, as stated earlier, Gen+1, Gen+2, Gen+3 program are readily available between us and our premier customers. Marco any color you'd like to add there?

Marco Racanelli

Yes. No, I think as you say, it will depend a little bit how big NPO gets for us in the short term will depend on market adoption. The beauty for us is that we have strong share, both the pluggable and NPO. So if it's not NPO and it's pluggable, that's still Tower wafer being built. So -- and yes, I think as Russell indicated, it's -- most of our existing customers that are building NPO products with us. We haven't put that out in the public domain, specifically with them, but we'll probably look to release some PR so that it's in the public the near future.

Operator

Dear speakers, there are no further questions for today. I would now like to hand the conference over to Russell Ellwanger for any closing remarks.

Russell Ellwanger

Thank you very much. Thank you for the questions. Just as a post-summary comment, we had a corporate theme that we implemented, still have a corporate theme, it was implemented in 2025, and it was Bold Growth, Limitless Impact, Infinite Reach.

We really adhere to that and try to breathe it in, in everything that we're doing. If you're old in your growth, if you believe that you can have limitless impact, if you try to have infinite reach, you truly inspire customers that you're working with as your partners. We had a very successful conference in Shenzhen, China customer conference. We have in 1.5 weeks or so our customer conference in Santa Clara, California Technology Conference, and it was a reunion of friends, and I can't overstate that. When you adhere -- and we really try to adhere to our values. We have what we call our value factors and its leadership, partnership impact and innovation. When you adhere to those values, and the definitions of them. I mean the true underlying definition of leadership, I think, is something that John F. Kennedy had stated, and that was "leadership and learning are indispensable to each other".

If you would adhere to a partnership, which means truly working with your customers to adapt ourselves to what the market demands are and quickly, really doing it quickly, it drives unbelievable relationships as you have success together. And that's what we see through our long-term agreements is a customer acknowledgment of the success that we've had together. We get a lot of joy in it, they get joining it as success begets success, and it can never be overstated. Impact is very, very important. An impact really it deals with giving your people the power capability and ownership to make change, to make things happen. A lot of our customers are amazed at the amount of young engineers that they interface with because we focus on the development of very high-talented, high-potential people, and their growth is really dependent upon working with customers that inspire them. There are three things that really enable success and the capability, environment and truly the motivation, your passion.

If you have passion, you have capability and you create an environment of partnership where the customer feels open and that they can tell you whatever they want, you won't get offended, they can criticize where it has to get better, but also where we can tell them what they have to do better, what a relationship that becomes. And what a truly blissful time it is an inspiring time. Where we talk about impact and innovation as well, and it's very, very critical. Our model has been and continues to be very, very little drop from gross margin to operating margin. I had stated it before, but at the $3.6 billion model that we put out, gross margin 45%; operating margin, 38% and a 7-point drop from gross to operating margin. And why is that critical for us. We want to stay lean on infrastructure because it allows us to be fast. And the faster we are, the more we delight our customers.

If the questions -- well, not if, the questions about NPO, the questions about market share in the future, they're very, very critical. But how do we ensure that it happens, it's fast execution. And that is within our model of innovation because within innovation, sometimes the greatest innovation is simplification and simplification reduces the drop of gross margin to operating margin.

So with that, I'd like to end the call. It's very, very exciting where we're at, where we're going. I really thank investors for their interest in the company. I very much appreciate our customers for their faith and trust as demonstrated by the LTAs. And our employees worldwide are amazing. They give and give and get rewards, I think, that are commensurate with what they're doing, but their biggest reward is the joy of going home at the end of the day, believing that you've been successful, having the tools to have been successful. We look forward to connecting with you at the upcoming conferences we'll be attending. Barclays 24th Annual Global Conference in San Francisco, December 10, Nomura at CES '27 in Las Vegas, 5th and 6th of January '27. The 29th Annual Needham Growth Conference in New York, the 12th and 13th of January. And as always, please analysts, investors contact our Investor Relations department. Any time you have questions or wish to follow up. We very much enjoy your inputs. We get better through them. Thank you very much.

Operator

This concludes today's conference call. Thank you for participating. You may now all disconnect. Have a nice day.