코히런트(COHR) 2027 회계연도 1분기 실적 콘퍼런스콜: AI 광학이 사상 최대 매출 견인

핵심 요약

• 코히런트(Coherent Corp., NYSE: COHR)가 발표한 2027 회계연도 1분기 매출은 사상 최대인 23억 7,000만 달러로, 전 분기 대비 16% 및 전년 동기 대비 50% 증가했다. 프로포마 매출은 전년 동기 대비 56% 증가했으며, 이는 2026년에 매각된 사업 부문을 제외한 수치다.
• Non-GAAP 매출총이익률은 40.6%에 달했으며, Non-GAAP 영업이익률은 23.2%로 상승했다. Non-GAAP 희석 EPS는 전년 동기 대비 80% 증가한 2.09달러를 기록했다.
• 데이터센터 및 통신 부문 매출은 전년 동기 대비 79% 성장해 전체 매출의 82%를 차지했다. 데이터센터 매출은 85% 증가했는데, 이는 AI 인프라, 800G 및 1.6T 트랜시버, 인듐 인화물(InP) 생산능력 확장에 힘입은 것이다.
• 포톤링크(PhotonLink)의 양산 확대가 본격화되기 시작했다. 경영진은 매출이 3월 분기에 1억 달러를 상회하고, 또한 2027년 말 이전에 분기당 2억 5,000만 달러를 넘어설 것으로 예상하고 있다. 또한 코히런트는 5년 만기 포톤링크 장기 공급 계약(LTA)을 체결했으며, 계약 규모는 약 200억 달러에 달하며, 관련 매출은 2028년 초부터 발생할 것으로 예상된다.
• 2027 회계연도 2분기 실적에 대해 경영진은 매출 가이던스로 25억 5,000만~27억 5,000만 달러, Non-GAAP 희석 EPS 가이던스로 2.28~2.48달러를 제시했다.
• 경영진은 2027 회계연도 프로포마 매출 성장률이 60%를 상회할 것으로 예상하며, 코히런트 최초로 2027 회계연도 말까지 분기 매출 30억 달러 달성에 이어, 2028 회계연도에는 최초로 분기 매출 40억 달러 달성을 계속해서 목표로 삼고 있다.

주요 재무 실적

매출총이익률 개선은 가격 최적화, 투입 비용 절감, 제조 수율 개선 및 6인치 인듐인화물(InP) 생산으로의 전환을 반영했다. 코히런트는 장기 목표치를 상향 조정했으며, 이제 매출총이익률이 다음 수준을 상회하는 첫 분기를 예상하고 있다: 2028년 말 이전 45%.

사업 및 운영 실적

데이터센터 및 통신

부문 매출은 전분기 대비 21% 및 전년 동기 대비 79% 증가해, 두 지표 모두에서 4개 분기 연속 성장 가속화를 기록했다. 해당 부문은 코히런트 전체 매출의 82%를 차지했다.

데이터센터 매출은 전분기 대비 17% 및 전년 동기 대비 85% 증가했다. 코히런트는 여러 고객사를 대상으로 1.6T 트랜시버 양산을 확대하는 한편, 800G 매출 성장을 지속하고 있다. 경영진은 800G 매출이 2027년까지 계속 성장하고, 1.6T 매출은 향후 몇 분기 동안 가파르게 증가할 것으로 예상하고 있다.

2027년까지의 수주 커버리지는 '이례적'인 수준으로 평가됐으며, 고객사 주문은 2028년까지 점차 연장되고 있다. 코히런트는 해당 분기 동안 통상 2020년대 말까지 이어지는 다수의 장기 계약을 체결했다.

포톤링크 및 집적 광학

포톤링크(PhotonLink)는 스케일아웃, 스케일업 및 스케일인 아키텍처 전반에서 CPO, NPO 및 칩 간 광학 링크를 지원한다. 코히런트는 이러한 애플리케이션이 대략 2030년까지 300억 달러 규모의 시장 기회를 나타낸다고 추산하며, 이에 따라 데이터센터 및 통신 부문의 총 시장 기회는 약 900억 달러에 달할 전망이다.

초당 100테라비트의 대역폭을 요구하는 반도체의 경우, 경영진이 추산하는 코히런트의 전체 포톤링크(PhotonLink) 포트폴리오 가치는 대략 반도체당 1만 5,000달러 수준에 달할 수 있으며, 이는 각 고객이 선택하는 제품에 따라 달라진다.

코히런트는 주요 AI 기업과 신규 5년 포톤링크 장기 공급 계약(LTA)을 체결했으며, 계약 규모는 약 200억 달러에 달한다. 이번 계약은 동사의 400밀리와트(mW) CW 레이저를 기반으로 하며 보다 광범위한 솔루션 세트를 포함한다. 매출은 2028년 초부터 본격적으로 확대될 것으로 예상된다.

이번 계약 및 앞서 발표된 엔비디아(NVIDIA) LTA 외에도, 코히런트가 발표한 협력 건수는 10건이 넘는 CPO 협력, 10건이 넘는 NPO 협력 및 5건이 넘는 칩 간(chip-to-chip) 협력에 달한다. 경영진은 초기 스케일업(scale-up) 구축에 인듐인화물(InP) 기반 고출력 CW 레이저가 사용되고, 2027년 말 또는 2028년 초경 VCSEL 기반 NPO 애플리케이션이 뒤따를 것으로 예상하고 있다. 스케일인(scale-in) 칩 간 애플리케이션은 이후인 2029년에서 2030년경으로 예상된다.

스케일아크로스(Scale-Across) 데이터센터 인터커넥트

통신 부문 매출은 전분기 대비 34% 및 전년 동기 대비 84% 증가했으며, 이는 전통적인 통신보다는 주로 스케일아크로스 데이터센터 인터커넥트에 의해 견인되었다.

스케일아크로스 매출은 2억 달러를 돌파하며, 전분기 대비 53% 증가했고 전년 동기 대비로는 거의 4배 급증했다. 해당 사업은 거의 연간 10억 달러의 매출 런레이트(run rate)에 도달했다. 경영진은 3월 분기부터 본격화되기 시작하는 멀티레일(multi-rail) 제품에 일부 힘입어, 분기 스케일아크로스 매출이 2027년 말까지 1분기(회계연도) 수준 대비 대략 두 배로 증가할 것으로 예상하고 있다.

인듐인화물(InP) 생산 능력

자체 인듐인화물(InP) 생산량은 당초 계획보다 한 분기 앞서 전년 동기 대비 두 배로 증가했다. 코히런트는 EML(전계흡수변조 레이저) 부문에서 분기 최대 생산량을 기록했으며, 2027년 말까지 자체 생산 능력을 다시 두 배 이상으로 늘리는 경로를 순조롭게 밟아가고 있다.

동사는 텍사스와 스웨덴의 6인치 시설이 3인치 라인의 수율을 웃도는 수준에서 EML, CW 레이저, 고속 포토다이오드를 생산하고 있다고 밝혔다. 취리히 생산 시설은 2027년 상반기에 가동을 시작할 것으로 예상된다.

경영진은 트랜시버 성장세가 추가된 자체 생산량을 흡수하면서, 자체 조달과 외부 조달 인듐인화물 부품의 현재 믹스가 지난 1년간 거의 변하지 않았다고 말했다. 동사는 향후 몇 분기에 걸쳐 믹스가 저비용의 자체 공급 쪽으로 전환될 것으로 예상하며, 이에 따른 매출총이익률 개선 효과의 대부분은 아직 본격화되지 않은 상태다.

OCS 및 산업 부문

광회로 스위치(OCS) 고객 협력은 스케일아크로스, 스케일아웃, 스케일업 애플리케이션 전반에서 확대되고 있다. 코히런트는 320x320 플랫폼 및 더 고밀도 구성을 위해 2개 사업장에서 생산을 늘리고 있다. 경영진이 추산하는 OCS 시장 기회는 적어도 2030년까지 40억 달러에 달한다.

산업 부문 매출은 프로포마(pro forma) 기준 전분기 대비 및 전년 동기 대비 보합세를 유지했다. 수주는 반도체 장비 고객사가 견인했으며, 경영진은 2분기(회계연도)에 성장이 재개될 것으로 예상하고 있다. 그 밖에 언급된 기회 요인으로는 AI 서버 회로기판 드릴링, 핵융합 에너지 애플리케이션, xPU 냉각 소재 및 양자 기술 등이 있다.

경영진 가이던스

코히런트가 데이터센터 및 통신 수요에 맞춰 생산 능력을 확장함에 따라, 경영진은 2분기(회계연도)에도 설비투자(CapEx)가 전분기 대비 다시 증가할 것으로 예상하고 있다.

2027 회계연도에 대해, 경영진은 프로포마(pro forma) 기준 매출 성장률이 다음을 상회할 것으로 예상하고 있다: 60%. 또한 회사가 예상하는 최초의 실적은 2027 회계연도 말 분기 매출 30억 달러 달성이며, 아울러 사상 처음으로 달성 경로를 보고 있는 목표는 2028 회계연도 분기 매출 40억 달러입니다. 경영진에 따르면 이러한 목표는 수주잔고, 장기 계약, 자체 설비 증설 및 공급 계약을 통해 뒷받침됩니다.

리스크 및 주목할 부문

• 생산능력 실행: 성장 목표는 인듐인(InP), OCS 및 기타 제조 생산능력의 빠른 확장에 달려 있습니다. 자본적 지출은 7억 7,600만 달러로 증가했으며, 회계연도 2분기에 다시 늘어날 것으로 예상됩니다.
• 내부 공급 전환: 예상되는 매출총이익률 개선 효과의 상당 부분은 코히런트가 자체 생산하는 인듐인(InP) 부품 비중을 늘려야 실현될 수 있습니다. 내부 조달과 외부 조달 간 비중은 아직 실질적으로 변하지 않았습니다.
• 집적 광학 도입 시기: 포톤링크(PhotonLink)의 성장은 CPO 및 NPO 도입 시점에 달려 있습니다. 경영진은 2027년(역년 기준) 말 또는 2028년 초에 스케일업 출하가 시작될 것으로 예상하며, 칩 간(chip-to-chip) 도입은 그 이후에 이루어질 것으로 전망하고 있습니다.
• 데이터센터 프로젝트 우려: 애널리스트들은 데이터센터 자금 조달 및 인허가 문제에 대해 질문을 제기했습니다. 경영진은 생산능력 투자를 뒷받침하는 근거로 고객사 선급금, 확정 주문, 다년 계약을 제시했습니다.
• 업계 공급 과잉 가능성: 한 애널리스트는 2028년(역년 기준)에 인듐인(InP) EML 생산능력의 초과 공급이 나타날 수 있는지 질문했습니다. 경영진은 트랜시버, 집적 광학 및 데이터센터 인터커넥트 제품 수요를 언급하며 해당 기간 동안 중대한 공급 과잉 위험은 보이지 않는다고 답했습니다.

애널리스트 Q&A 주요 내용

• 포톤링크 램프업: 경영진은 3월 분기에 1억 달러 이상의 매출을, 2027년(역년 기준) 말 이전에는 분기당 2억 5,000만 달러를 웃도는 매출을 기록할 것으로 예상하고 있습니다. 새로 체결된 200억 달러 규모의 장기 계약(LTA)은 2028년(역년 기준) 초부터 매출에 기여하기 시작할 예정입니다.
• 스케일어크로스 성장: 코히런트는 장비 공급업체와 하이퍼스케일러에 부품, 모듈, 라인카드, 서브시스템 및 완성형 솔루션을 공급하고 있습니다. 멀티레일(Multi-rail) 매출은 3월 분기부터 본격적으로 늘어나기 시작할 것으로 예상됩니다.
• 매출총이익률 잠재력: 경영진은 45%를 상회하는 목표치가 상한선으로 여겨지지는 않는다고 밝혔습니다. 신제품 양산 확대, 6인치 인듐인(InP)의 영업 레버리지 효과, 가격 최적화 및 비용 절감이 주요 동인입니다.
• 고객 자금 기반 생산능력: 코히런트는 거의 모든 LTA가 설비 증설 자금을 지원하기 위해 선급금이나 일회성 엔지니어링 비용(NRE) 지급과 같은 고객 선지급금을 포함하는 구조를 따른다고 밝혔습니다.
• 기판 확보 여건: 회사는 5곳 이상의 기판 공급업체를 두고 있으며, 이 중 다수와 장기 계약을 체결했습니다. 경영진은 2027년(역년 기준) 생산능력 확장에 필요한 기판을 확보했다고 밝혔습니다.
• 경쟁적 입지: 경영진은 추진 중인 루멘텀(Lumentum)과 마벨(Marvell)의 결합에 대해 논의하면서, 경쟁적 이해 상충 없이 하이퍼스케일러 및 반도체 고객사를 지원할 수 있는 독립적인 광학 솔루션 공급사로서 코히런트의 입지를 강조했습니다.

실적 발표 컨퍼런스콜 녹취록 전문

전체 실적 발표 컨퍼런스 콜 녹취록

경영진 발표

Operator

Greetings, and welcome to the Coherent Corp. First Quarter Fiscal Year 2027 Earnings Webcast. [Operator Instructions] Please note this event is being recorded.

It is now my pleasure to introduce your host, Mr. Paul Silverstein, Senior Vice President of Investor Relations for Coherent. Please go ahead.

Paul Silverstein

Thank you, operator, and good afternoon, everyone. With me today are Jim Anderson, Coherent's CEO; and Sherri Luther, Coherent's CFO. During today's call, we will provide a financial and business review of the first quarter of fiscal 2027 and the business outlook for the second quarter of fiscal 2027. Our earnings press release can be found in the Investor Relations section of our company website at coherent.com.

I would like to remind everyone that during our conference call, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company. These are subject to a number of significant risks and uncertainties, and our actual results may differ materially. For a discussion of factors that could affect our future financial results and business, please refer to the disclosure in today's earnings release, our most recent Forms 10-K and 10-Q and the reports that we may file on Form 8-K with the Securities and Exchange Commission.

All our statements are made as of today, November 4, 2026, based on information currently available to us. Except as required by law, we assume no obligation to update any such statements. During this call, we will discuss non-GAAP financial measures. You can find a reconciliation of these non-GAAP financial measures to GAAP financial measures in our earnings release and investor presentation that can be found on the Investor Relations section of our website at coherent.com.

Let me now turn the call over to our CEO, Jim Anderson.

James Anderson

Thank you, Paul, and thank you, everyone, for joining today's call. Q1 marked a strong start to fiscal 2027. Revenue increased 16% sequentially and 56% year-over-year on a pro forma basis. Non-GAAP EPS increased 80% year-over-year, reflecting continued gross margin expansion and operating leverage. Strong customer demand and our rapid capacity expansion are supporting this growth with near-term bookings and newly signed LTAs providing multiyear visibility. We remain on track to achieve our first $3 billion revenue quarter by the end of fiscal '27. We expect the full year pro forma revenue growth rate to exceed 60%, more than double our fiscal '26 pro forma growth rate. We also have a clear line of sight to our first $4 billion revenue quarter in fiscal '28.

Coherent is a photonics platform company, bringing together technologies across the optical signal chain from materials and components through subsystems and complete solutions. Our breadth positions us to support customers across Datacenter & Communications networks. We're at the center of an extraordinary expansion in optical networking infrastructure, driven by rapid AI growth, the transition from copper to optical connectivity and rising demand for bandwidth and energy efficiency across increasingly large and complex datacenter architectures.

Our established businesses continue to grow rapidly, with datacenter revenue up 85% year-over-year in Q1. At the same time, new growth areas are ramping as well. For example, scale-across datacenter interconnect revenue is growing rapidly with Q1 revenue roughly quadrupling year-over-year and reaching a nearly $1 billion annualized revenue run rate.

In addition, PhotonLink, our newly announced integrated optics platform, is also beginning to ramp, and we expect PhotonLink revenue to exceed $100 million in our March quarter. Accelerating revenue growth, continued gross margin expansion and operating leverage position us to grow profitability faster than revenue. We expect fiscal '27 to be an exceptional year for Coherent.

Moving to our Datacenter & Communications segment. Revenue increased 21% sequentially and 79% year-over-year, marking a fourth consecutive quarter of accelerating growth on both a sequential and year-over-year basis. The segment now accounts for 82% of Coherent's total revenue.

Near-term demand remains strong and our visibility into long-term demand continues to improve. Order coverage through calendar '27 is exceptional, and customer orders increasingly extend into calendar '28. We signed multiple new LTAs in Q1, and customer LTAs typically extend through the end of the decade. We continue to see no signs of attenuation in demand. Within the segment, Datacenter revenue increased 17% sequentially and 85% year-over-year, marking our fourth consecutive quarter of double-digit sequential growth. We expect another quarter of strong sequential and year-over-year growth in the current quarter.

In Datacenter transceivers, we're ramping 1.6T products with multiple customers, while continuing to grow 800-gig revenue. We expect strong transceiver revenue growth in Q2, driven by both 1.6T and 800 gig. We expect 1.6T revenue to continue ramping rapidly over the coming quarters, while 800-gig revenue continues to grow through calendar '27.

At ECOC in September, we demonstrated our progress beyond 1.6T with a 3.2T OSFP pluggable transceiver built with our 400-gig differential EMLs and photodiodes. We expect pluggable transceivers to remain the dominant form factor for scale-out networks over the coming years. Coherent is helping shape the industry ecosystem for emerging formats, demonstrating XPO at OFC and also supporting the new USFP form factor.

Turning to OCS. Customer engagement is broadening across scale-across, scale-out and scale-up networks, spanning our 320x320 platform and denser configurations. We're expanding production at 2 sites to meet demand. As capacity ramps, we expect OCS to become an increasingly meaningful contributor to revenue growth and margin expansion.

At ECOC in September, we launched PhotonLink, our new platform for integrated optics. It's designed to address the transition from copper to optical connectivity in AI datacenters, creating a significant opportunity for Coherent. We estimate that CPO, NPO and chip-to-chip optical interconnects represent approximately $30 billion of market opportunity by 2030, bringing our total market opportunity for Datacenter & Communications to about $90 billion in 2030.

PhotonLink brings together Coherent's photonic technologies and manufacturing capabilities in an end-to-end platform that spans the optical signal chain from generating and shaping light through transmission and detection to converting the signal to electrical form. The platform is designed for a wide range of architectures and applications, including CPO and NPO, scale-out and scale-up networks and scale-in chip-to-chip links. It includes lasers, laser modules, complex fiber assemblies and other optical components and supports indium phosphide-based CW lasers and VCSEL-based technologies. We believe this breadth and depth differentiates us. For an xPU or switch chip with 100 terabits per second of bandwidth, we estimate our content opportunity can reach approximately $15,000 per chip.

Customer engagement is validating the opportunity. In Q1, we signed another significant PhotonLink LTA with a major AI company. The total revenue value of this new LTA is approximately $20 billion over 5 years, with revenue beginning to ramp in early calendar '28. This is in addition to our previously announced LTA with NVIDIA. Beyond these 2 anchor customers, we have more than 10 additional CPO engagements, more than 10 NPO engagements and more than 5 chip-to-chip engagements.

We expect PhotonLink revenue to ramp quickly over the coming quarters. We expect revenue to exceed $100 million in our March quarter and to exceed $250 million per quarter before the end of calendar '27. Expected shipments in the current quarter include a range of CPO products, including our ultra-high power 400-milliwatt CW laser, our fiber assembly solutions and other products.

PhotonLink is initially being deployed in CPO scale-out applications. We expect to begin shipping PhotonLink products for CPO and NPO scale-up applications in the second half of calendar '27. Over time, we expect PhotonLink to become a multibillion-dollar product line for Coherent.

Expanding our indium phosphide capacity is a key enabler of growth in both indium phosphide pluggable transceivers and indium phosphide-enabled PhotonLink solutions. Our 6-inch indium phosphide expansion is progressing well. We doubled our internal indium phosphide output capacity year-over-year in Q1, 1 quarter ahead of our original plan. Q1 was our highest production quarter ever for EMLs, including both 100-gig and 200-gig devices. The ramp contributed meaningfully to datacenter growth, and we expect it to remain an important driver of revenue and gross margin.

Most of the gross margin benefit of our indium phosphide expansion is still ahead of us as the expansion will enable us to increase our mix of internally sourced indium phosphide components, which have a tremendous cost advantage. We remain on track to more than double our internal indium phosphide output capacity again by the end of calendar '27. We've secured the substrates and other critical inputs needed to support this ramp. And given the strength of customer demand are planning additional capacity expansion beyond 2027.

Our 6-inch lines in Texas and Sweden are producing EMLs, CW lasers and high-speed photodiodes with yields that continue to exceed those of our 3-inch lines. Production includes our 400-milliwatt CW laser for CPO, which is covered by our NVIDIA partnership and is expected to begin generating revenue this quarter. We also remain on track to begin 6-inch indium phosphide production in Zurich in the first half of calendar '27, further extending what we believe is a meaningful manufacturing advantage.

Moving to Communications. Revenue increased 34% sequentially and 84% year-over-year, and we expect another quarter of strong sequential and year-over-year growth in the current quarter. Traditional telecom continued to grow, but the primary growth driver was scale-across, which has quickly become the largest source of revenue within Communications.

Over the past year, we've expanded our datacenter interconnect portfolio to help customers connect datacenters with greater capacity and efficiency. Building on our leadership in pump lasers and our broad component portfolio, we've introduced higher level products, including modules, line cards, subsystems and our new multi-rail solution. These offerings are seeing strong customer demand and help customers make better use of existing fiber while scaling high-capacity links between distributed AI datacenters.

In Q1, scale-across revenue exceeded $200 million, up 53% sequentially and nearly fourfold year-over-year. That puts the business at a nearly $1 billion annualized revenue run rate, making it one of our fastest-growing businesses. With multi-rail beginning to ramp in our March quarter, we expect quarterly scale-across revenue to roughly double from the Q1 level by the end of calendar '27.

Turning to Industrial. Q1 revenue was flat sequentially and year-over-year on a pro forma basis. Bookings were strong, led by semiconductor capital equipment customers. And we expect growth to resume in the current quarter. Near-term growth is being driven by semiconductor capital equipment, AI infrastructure and fusion energy. In semiconductor equipment, our lasers, precision optics and engineered materials serve front-end applications and we've seen a strong uptick in bookings.

In AI infrastructure, our CO2 lasers and zinc selenide optics are used to drill advanced circuit boards for AI servers. Demand is also growing for our pulsed laser deposition systems used to produce high-temperature superconductor tape for magnetic fusion.

Over the longer term, we see significant opportunities in datacenter xPU cooling and quantum technologies. For AI datacenters, our Thermadite material can enable higher xPU performance and token generation. We're shipping prototypes to several strategic customers this quarter, and we expect revenue to begin ramping in the second half of calendar '27.

In quantum, we supply lasers and materials to leading quantum technology developers and are engaged in next-generation programs. Given the multiple growth areas, we expect our Industrial segment to return to growth and build momentum over the coming quarters.

In summary, Q1 demonstrated the strength of our growth momentum and the breadth of our opportunity. Demand remains strong, customer commitments continue to deepen our visibility and we are expanding capacity ahead of plan.

Established businesses continue to grow rapidly while new growth areas are ramping quickly. Our focus is on execution, expanding capacity, delivering for our customers and improving profitability as we grow. We believe this combination positions Coherent for sustained growth in fiscal '27 and beyond.

I want to thank our team for their outstanding work this quarter. I'll now turn the call over to Sherri.

Sherri Luther

Thank you, Jim. I'm extremely pleased with the outstanding Q1 financial results. Our performance in the first quarter represents a powerful continuation of our growth trajectory, characterized by accelerating top line momentum, durable structural gross margin expansion and significant operating leverage. The compounding impact of these operational drivers fuels non-GAAP EPS growth to significantly outpace revenue growth. We achieved this strong performance while preserving balance sheet flexibility and strategically directing capital towards targeted capacity expansions and making investments in our product road map to support robust AI Datacenter & Communications demand.

Let me now provide a summary of our results. First quarter revenue was a record $2.37 billion, representing a sequential acceleration of 16% and a substantial 50% increase year-over-year. This performance was driven by the continued AI infrastructure build-out across our Datacenter & Communications businesses. On a pro forma basis, excluding the financial impact of the Aerospace and Defense business, and the Munich, Germany, product division, both of which were divested in 2026, revenue increased 16% sequentially and 56% year-over-year.

Our non-GAAP gross margin for the first quarter was 40.6%, representing a 37 basis point increase sequentially and a 184 basis point improvement as compared to the year ago quarter. This continuous expansion reflects durable pricing optimization and structural cost improvements, particularly within our Datacenter & Communications segment.

The improvements in gross margin, both sequentially and year-over-year, were driven primarily by pricing optimization, lower product input costs and improved manufacturing yields. We have made significant progress with our pricing optimization strategy to ensure that we price our highly differentiated products for the value they provide to our customers.

Our gross margin profile is also benefiting from the transition to 6-inch indium phosphide as we continue to scale our production capacity. Given the progress that we have made over the past couple of years and our current trajectory, it is clear that we need to raise our gross margin target. We now expect to achieve our first quarter with gross margin above 45% before the end of calendar 2028. We believe the drivers of this gross margin expansion are continued pricing optimization, ramp of new growth platforms, full operational leverage of our 6-inch indium phosphide capacity and other cost improvements.

I also want to note that the new over 45% target is 300 basis points higher than the long-term target we provided to the investment community only 18 months ago.

First quarter non-GAAP operating expense was $412 million compared to $377 million in Q4 and $304 million in the year ago quarter. Non-GAAP operating expenses as a percentage of revenue decreased meaningfully to 17.4%, down from 18.4% in the prior quarter and 19.2% in the year ago quarter. This improvement reflects enhanced operational discipline and underscores the powerful operating leverage inherent in our business.

From an R&D perspective, we are strategically directing our capital toward high conviction long-term growth drivers, which Jim outlined in his remarks. This disciplined approach ensures our investments are concentrated where they generate the highest financial returns and drive sustainable revenue growth.

Our first quarter non-GAAP operating margin increased to 23.2% from 21.8% in the prior quarter and 19.5% in the year ago quarter. The sequential and year-over-year increases were driven by strong revenue growth, expanding gross margin and structural expense discipline. First quarter non-GAAP earnings per diluted share was $2.09, up 20% from the prior quarter and up 80% from the year ago quarter. Earnings growth continued to exceed revenue growth, demonstrating gross margin expansion and operational leverage accelerating our bottom line profitability.

Turning to the balance sheet. We closed the quarter with a healthy cash balance of $2.34 billion compared to $2.59 billion at the end of the prior quarter and $1.55 billion at the end of the year ago quarter. Our debt leverage ratio at the end of Q1 was 1.1x as we continue to maintain a highly resilient and flexible balance sheet. Consistent with our capital allocation priorities, we strategically invested in opportunities that will drive long-term growth and profitability.

Capital expenditures for the first quarter increased sequentially to $776 million, up from $556 million in the prior quarter and $104 million in the year ago quarter. These investments directly support the future growth of our Datacenter & Communications businesses with excellent financial returns. Our conviction to deploy this capital at an accelerated rate is backed by exceptional customer visibility, a robust pipeline of committed purchase orders and additional multiyear supply agreements.

Furthermore, this capacity offers significant fungibility as our infrastructure can be efficiently repurposed to support multiple product lines. Given the strong visibility and record backlog, we expect capital expenditures to increase sequentially again in the second quarter. With our strong balance sheet and profitability trajectory, we possess the scale and flexibility to continue investing capital toward high ROI growth initiatives while supporting customer demand.

I will now turn to our guidance for the second quarter of fiscal 2027. We expect revenue to be between $2.55 billion and $2.75 billion. We expect non-GAAP gross margin to be between 40% and 42%. We expect total operating expenses of between $430 million and $450 million on a non-GAAP basis. We expect the tax rate for the quarter to be between 18% and 20% on a non-GAAP basis. We expect non-GAAP diluted EPS to be between $2.28 and $2.48.

In summary, fiscal 2027 is off to a powerful start with strong financial momentum backed by deep demand visibility. Our focus remains on rapidly ramping capacity to meet demand, expanding our gross margin, capitalizing on the immense operating leverage in our model and executing a disciplined capital allocation strategy focused on maximizing returns that drive long-term shareholder value.

That concludes my formal comments. Operator, please open the line for the Q&A session.

Operator

[Operator Instructions] The first question today comes from Joseph Cardoso with JPMorgan.

질의응답

Joseph Cardoso

Really appreciate some of the disclosures that you provided this quarter. And I wanted to first start with PhotonLink and the initial visibility you provided there. I was hoping you could provide more color with respect to the ramp itself, especially just given the backdrop of the $20 billion agreement that you've highlighted? And then I have a follow-up.

James Anderson

Yes. Thanks, Joe. So first of all, yes, we're certainly really pleased with the momentum that we have around PhotonLink and kind of whether you measure it by customer engagement or LTAs, as you mentioned, or just the revenue outlook that I mentioned in the prepared remarks, we're just really pleased with the progress there.

And just as a reminder, PhotonLink is -- we kind of consider it a one-stop shop for integrated optics. It's basically our platform for allowing customers to basically have a platform that allows for the optical link from end to end. It's the complete optical signal chain and all the technology along the optical technology along that path. And really excited about the progress, as I mentioned in the prepared remarks, over 20 customer engagements on CPO, over 10 -- or sorry, over 10 on CPO, over 10 on NPO and then over 5 chip-to-chip engagement. So the customer engagement is very extensive and they continue to grow.

You mentioned or asked about the LTA. Really pleased about that major AI customer and that LTA is, as I said, $20 billion over a 5-year period. And the revenue for that would start at the beginning of calendar '28. But the PhotonLink revenue ramp is starting ahead of that. We'll ship a little bit of -- a small amount of PhotonLink revenue this quarter, but we expect to ship over $100 million in our March quarter. And then before the end of calendar '25 (sic) [ '27 ], expect to ship over $250 million a quarter. So really pleased with all of that progress.

And I guess the other thing that I would add is, if we look further out, we're expecting this over time to become a multibillion dollar product line. The initial shipments of PhotonLink are all around CPO in the scale-out network. But when we start to ship into a CPO and NPO into scale-up networks, I think that drives another huge leg of demand for this technology. I think that will be the much bigger adoption that we see. And that we expect to start to ship revenue into CPO and NPO applications in the scale-up domain towards the end of calendar '27. So yes, we're excited about the progress and expect this to become a major product line for us.

Joseph Cardoso

And maybe in a similar vein, I wanted to touch on the scale-across DCI opportunity. Particularly curious how we should think about the customers and products that are underpinning the revenue outlook to double by the end of the next calendar year.

James Anderson

Yes. Thanks again. This has been a really great growth area for us over the past year. In fact, it's surprised us a little bit how quickly this has grown. So overall, so DCI and scale-across that falls within our Communications part of our business, which is essentially everything outside of -- all communications outside of the Datacenter. And if we look at that segment overall, Communications grew 84% year-over-year. But while telecom grew most of that -- the vast majority of that growth was driven by DCI. And so what's going on there is you're seeing these large AI workloads that are spanning multiple datacenters. And so that's driving the need for much greater bandwidth and connectivity -- optical connectivity between datacenters. And so that's what's driving. That's the underlying market dynamic that's driving the growth.

And then you asked about customers and products. So customers would be either equipment makers. So we're certainly selling into multiple different equipment makers that -- whose equipment goes into scale-across applications, but also selling directly to hyperscalers as well. And so it's a mixture of both types of customers. And then the products, it's interesting, we sell -- in this part of the market, we sell products at multiple levels of the value chain. So we sell in components like, for instance, pump lasers that go into a wide range of different equipment. At one level up in the value chain, we sell subsystems and modules, so ZR/ZR+ modules, line cards, amplifiers, ROADMs, et cetera. And then one level up from that, in some cases, sell complete solutions to our customers as well. So we're selling across multiple levels of value chain. And so I think that's helped us really maximize the growth here as well.

And as I mentioned in the prepared remarks, if we look at our Q1 revenue, it was over $200 million. So we're basically at about $1 billion annualized revenue run rate now. And that's -- that Q1 growth had quadrupled year-over-year. And we expect that if you take that Q1 baseline, we expect the revenue to double again before the end of calendar '27. And so that growth that we've seen to date, that's ahead of multi-rail. So we expect multi-rail revenue, which is a -- multi-rails are great solutions, some great technology that we have there. We expect that to start -- that revenue to start to ramp in our March quarter. And so it's been a really good area of growth, and we expect this to continue to be a very good area of growth over the coming quarters as well.

Operator

The next question comes from Simon Leopold with Raymond James.

Simon Leopold

First, Jim, I want to thank you for the increased transparency on Slide 7. I know this is something I've been probably annoying you about. So thank you. The slide did not mention OCS. I know you talked about the product group in your prepared remarks. But could you update us on your progress and expectation for this product group? And then I've got a follow-up, please.

James Anderson

Yes, sure. Thanks, Simon. Thanks for the feedback, as always. We're always looking to improve. So I appreciate the feedback. But yes, in OCS, I did mention in the prepared remarks OCS falls within our Datacenter part of our business. And just reiterating, Datacenter was up 85% year-over-year in our Q1. OCS continues to make good progress. We've got really extensive customer engagement. We've seen that customer engagement, continue to expand across more applications and not just scale-out applications, but scale-across and scale-up applications as well. And that's across -- in terms of products, that's across both 320x320 systems, which are shipping as well as denser configurations. I think the denser configurations will be used more in the scale-up applications.

We're ramping production at 2 different sites in parallel, and that's really to meet the demand that we're seeing. We have a very differentiated technology. It's liquid crystal. It's nonmechanical technology, has better power efficiency and better reliability than MEMS-based solutions. So we feel really well positioned. We're ramping production. And I think it will be an increasingly larger contributor to our revenue growth in the quarters to come. And maybe last thing to add is we see the market opportunity for OCS as at least $4 billion by 2030. It could be larger than that.

Simon Leopold

And as a follow-up, I have a hard question, one that I imagine you anticipate getting is, what does the proposed consolidation of Lumentum and Marvell mean to you?

James Anderson

Yes. Thanks, Simon. No, I don't think that's a hard question at all. I think happy to answer that. I'd give a few thoughts on this. I think number one is it definitely demonstrates that optics and photonics is becoming an absolutely critical part of datacenter architecture.

As we look forward, as we convert the entire scale-up network and then even the scale-in chip-to-chip communication links, over time, all of that getting converted from copper to optical. Photonic technology is absolutely critical to making that happen to driving the innovation and datacenter architecture. So I think it's a proof point around just the criticality of optics and photonics.

And then I think for Coherent, I think it just -- it highlights the tremendous strategic value of the portfolio that is within Coherent. There's no other company, I believe, on the planet that has a broader and deeper set of photonic and optical technology. If you look at the entire optical signal chain from the generation of the light to the transmission through our fiberoptic cable that we make to the detection on the other side, I don't think there's any other company that's got a better portfolio of technology.

So I think it definitely highlights the strategic importance of our technology platform. And then look, I think from a customer perspective, I think that -- I think a lot of -- for a lot of the customers, our customers, I think it's really important that they have an independent photonic optical supplier and partner. And so I think it positions us really well to be that independent optical supplier that's going to support -- without any competitive conflict can support multiple different customers and support both hyperscaler customers, can support semiconductor customers.

And so I think it positions us really well to be that partner of choice in terms of optics for the industry. So look, we think we're really well positioned from both the technology standpoint, manufacturing standpoint and strategically. And I think we're on a really great growth trajectory. And I think you can see some of that already in the PhotonLink technology and the PhotonLink revenue ramp.

Operator

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

Thomas O'Malley

I guess the first part of the question for you, Jim, is why now in terms of giving the longer-term forecast? It's been something that you've kind of held off on doing for a while. You gave $3 billion and then $4 billion next year.

And then the second part of the question is, is that all backed with supply? How are you able to get visibility that far out? Is that already in the books today? Or perhaps you could give us like what percentage of that is backed with supply today? Or if you need to add any to get you to those numbers?

James Anderson

Yes. Thanks, Tom. I think it is a little bit different. We are providing a little bit longer-term visibility than we normally would. I think it's really backed by the -- just extraordinary level of visibility that we have on the business. When I look at the backlog, the bookings both -- so both near-term visibility and long-term visibility. If I look at just the backlog that we have, I mean, calendar '27 is basically booked out and the customers are booking out calendar '28 now. So from a near-term booking demand perspective, we have great visibility.

We've signed a number of new LTAs over the past quarter. I mentioned one of the big LTAs as part of the PhotonLink prepared remarks. But the LTAs also give us really good visibility because they include very high-quality forecast, long-term forecast. So I think that's -- given that really good visibility, that's certainly giving us confidence.

And then also, great that you mentioned supply. We've done a number of things within our supply base to make sure that we've got the right supply in place to support that as well. So yes, I would say that, that's -- those views that we gave, whether it's $3 billion this year and the $4 billion quarter next year is underpinned by supply agreements. Also, remember that we're highly vertically integrated. So a lot of that's underpinned by our own internal supply that we're ramping, right?

One of the constraints, if we look back over the past year or 1.5 years, one of the constraints has been indium phosphide. And I'm just really pleased with how the team has executed very, very well on the indium phosphide production ramp. We've doubled production capacity a quarter earlier than what we thought we would, and that gives me the confidence that they're going to achieve -- the team is going to achieve our plan to more than double capacity again by the end of calendar '27. So you put all that together and yes, we're feeling really good about the outlook and the visibility for the business.

Thomas O'Malley

Super helpful. And then just as my follow-up for that LTA. Obviously, $20 billion, a pretty robust number over that 5 years. Could you just give us any color just around that customer perhaps? And then if you can talk about the customer around some of the solutions that, obviously, it's PhotonLink, so it spans a large number of your products, but any more specificity would be really helpful, just given the size of that customer.

James Anderson

Yes, sure. So definitely a leading AI datacenter customer name you will definitely recognize, right? And so we're really proud of that partnership, excited about it. That's all -- that LTA for $20 billion is all around PhotonLink technology. I would call it -- I would say it's anchored by our laser, our 400-milliwatt CW laser. That's kind of the anchor of that LTA. But there's -- it's more than -- we're selling more than just that laser. So we're selling more, call it, a higher level solution than that. So more than just that laser.

So we're really, really pleased by that. As I mentioned, the revenue -- we expect the revenue to start to flow from that in early 2028 and to ramp over the course of the quarters and years after that. And so really pleased with that. And I think given the engagement that we have on PhotonLink across many different customers, many different solutions, we certainly expect to sign more long-term agreements and partnerships with PhotonLink over the coming quarters as well.

Operator

The next question comes from George Notter with Wolfe Research.

George Notter

I guess I just wanted to ask you a question about the 45% gross margin outlook before the end of '28. So I guess I'm curious like what does that assume about your transceiver business? I'm just wondering if, to hit that, you need to gate your transceiver supply or transceiver ramp? Or is this just a case where everything can ramp in units then naturally would ramp and you wind up at 45% gross margins. What's the view there?

James Anderson

Yes. Thanks, George. I'll start and see if Sherri has any color to add. But yes, we haven't assumed that we're kind of artificially gating the transceiver business. So there's no assumption on that. We're assuming that the transceiver business grows at what we think is a reasonable growth rate to expect from the transceiver business. But on top of that, we do have new products that are ramping on top of that. PhotonLink, a number of new products and datacenter interconnect, OCS ramping over time as well. And so you put all those products ramping over time. And those products are accretive to the company's margin. I think the new product ramp helps as well.

And then in addition to that, the other thing that I would add more on the cost side is a number of different cost activities that we're driving. But certainly, a big one of it is we expect additional leverage in terms of gross margin from our indium phosphide ramp. I think there's -- as I mentioned briefly in the prepared remarks, I think most of the benefit from indium phosphide expansion is still ahead of us. And so I think there's a good component of the margin expansion that's kind of underpinned by the indium phosphide expansion as well.

George Notter

Is it possible to get the company to higher margins than 45% over time? I mean, I guess I'm maybe looking at Lumentum as an example. Those guys obviously have had quite a bit of margin expansion. Obviously, the fastest-growing parts of that business are many of the areas where you're kind of looking for growth going forward. Is 45% just a stop in a bigger picture or a bigger number? What do you think?

James Anderson

I don't think there's any structural barrier to us driving gross margin above 40%. In fact, I think Sherri, in her remarks, I think she said over 45% before the end of calendar '28, I think, is the way she said it, right? So I -- yes, we certainly don't view just like our old target of 42%, where we said over 42%, we never viewed that as a stopping point. We viewed that as stop along the way. And I would say 45% is -- yes, stop along the way, and I don't see any structural reason why the company couldn't continue to drive gross margins beyond that.

Sherri Luther

Yes. I would just add that the target that we put out about 18 months ago at our Investor Day of over 42%, that was a 3- to 4-year target that we put out, but we pulled that in, right, in only 18 months now raising our target to over 45%. So as we get close to that target, we'll certainly reevaluate and look at what makes sense beyond that, but it's just a step along the way.

James Anderson

Yes. And George, maybe the other thing I'd add is we're certainly super focused on gross margin expansion. But the other thing that we're very focused on is operating leverage, right? And I think we're quite pleased to see that we're gaining operating leverage in terms of OpEx as a percentage of sales, and we expect that to continue as well. So I think you should expect continued operating leverage as well.

Operator

The next question comes from Gianmarco Conti with Deutsche Bank.

Gianmarco Conti

Yes. Great. Congrats on the quarter. Maybe starting with PhotonLink. I wanted to touch on the scale-in/scale-up opportunity with micro-VCSEL. You mentioned you had 5 engagements, I believe, chip-to-chip. I'm curious if you could talk a little bit more detail here whether there is anything you can quantify in terms of revenue to come online next year? Anything around micro-VCSEL?

James Anderson

Yes. So thanks for asking about micro-VCSEL. So certainly, when I look across the PhotonLink engagements, there is, I would say, many VCSEL, micro-VCSEL, we call it multi-core VCSEL because it -- the multi-core VCSEL attaches to our multi-core fiber that we produce. And so -- but multiple of those engagements involve our VCSEL technology. I mean maybe just as a reminder of the company's sort of depth in VCSEL technology. Number one is we've shipped over 700 million VCSELs into datacenter applications over the past year. So we have enormous manufacturing expertise in VCSELs.

Also at the ECOC conference, when we launched PhotonLink, we talked about our technology in VCSELs, micro-VCSELs, VCSEL arrays, et cetera. We showed that -- one of the things that we showed was that we had industry-leading energy efficiency around VCSELs of -- we showed a 1 picojoule per bit capability, which I think is very industry-leading capability. And so VCSEL is definitely a core technology, something that the company has a long history in and we have multiple engagements.

I think you'll see that deployed in NPO applications relatively in scale-up NPO applications in relatively short amount of time. I think you also asked about scale-in chip-to-chip. I think that's a little further out in time. I think PhotonLink and scale-in applications is more in the 2029 to 2030, but we'll definitely see, I think, VCSELs deployed in NPO applications well before that.

Gianmarco Conti

If I could just ask a follow-up. It's going to be a tough question, sorry to put you on the spot. But I've been receiving a lot of questions from investors around Chinese laser suppliers. I was wondering if you could provide any kind of thoughts on basically more supply coming in online? And whether you see them competing for ultra-high-powered lasers for CWs?

James Anderson

Yes. So I don't see that those competitors in terms of ultra-high power. If you're talking about like the 400-milliwatt CW laser, we're not seeing competition from really -- we really only have one competitor there as far as we can see, right? So really not seeing any other competitors out there with high-power seated laser capability.

Paul Silverstein

Operator, before we go to the next analyst, if we could ask all the other analysts to limit themselves to one question with the goal of trying to accommodate everybody on the call. All right. Thank you.

Operator

The next question comes from Dan Zhu with Bernstein.

Daniel Zhu

You mentioned that scale-up optics should start towards the end of calendar 2027. Anything you can share about whether the initial deployments are NPO or CPO? And also about how -- whether sort of this is primarily going to be on VCSELs or will this start with InP-based lasers?

James Anderson

Yes. Thanks, Dan. So I think what we'll see -- we'll definitely see CPO in scale-up. We certainly believe that. I think we'll see NPO scale-up soon after that as well. If not by the end of calendar '27, I think we'll certainly see deployment in the early part of '28. So I think in scale-up, both CPO and NPO within either that late '27 or early '28 time frame in the scale-up applications.

And then I think -- did you ask -- what was the second part of your question, Dan?

Daniel Zhu

How much of it is VCSEL versus InP?

James Anderson

Yes. I think the initial deployment is indium phosphide, the high-power CW lasers. That's for CPO in both initially scale-out and then in scale-up. And then I think VCSELs will happen in that NPO kind of early '28 time frame.

Operator

The next question comes from Ryan Koontz with Needham.

Jeffrey Hopson

This is Jeff Hopson on for Ryan. You noted ramping the indium phosphide supply for your own laser supply and that'd be able to lead to the margin expansion from weaning off external sources. I was just curious if there's any color on what the split is currently versus internal versus external? And kind of how -- or the time line of that over time as it relates to the margin expansion?

James Anderson

Yes. Thanks, Jeff. We haven't broken out the split, but I can tell you the split hasn't changed much in a year. We're using about the same mix of internal versus external now as we were a year ago. So although we've doubled our internal indium phosphide supply our -- the primary usage of our internal indium phosphide parts is in our datacenter transceivers.

And if you look at the growth rate of our datacenter business, datacenter transceivers, it's basically tracked the growth of our indium phosphide capacity, although -- basically one quarter later because when -- in the quarter we produce the lasers, the next quarter is when we ship the transceivers. And so essentially, our transceivers have been -- business has grown at the same rate as our internal indium phosphide, which means that our mix of internal versus external really hasn't changed much over the past year.

Now as we continue to expand indium phosphide now looking forward, more than doubling it by the end of '27, we do believe over the coming quarters, we'll start to shift more of the mix from external towards internal.

Now we always think -- we always believe we'll have some portion of our supply from external for our transceiver business. But we believe over the coming quarters, we'll start to shift more towards internal. And that's really because it's a tremendous cost advantage for us to use our internally produced lasers, especially the ones that are made on the 6-inch platform. And in my prepared remarks, when I said the majority of the benefit from 6-inch indium phosphide is ahead of us, that's actually what -- really what I'm referring to is that benefit of changing the mix towards internal is largely ahead of us and that's a much bigger cost structure improvement.

Operator

The next question comes from Papa Sylla with Citi.

Papa Sylla

Congrats on the very impressive results. Jim, I wanted to maybe double-click on the $15,000 of content per, I think, 100-terabit chip you mentioned earlier at ECOC. So assuming that is probably the full stack opportunity, which might be a longer-term figure. But I thought it would be helpful to also have a sense, even qualitatively, on what percentage, I guess, of that $15,000 you expect to capture in the near term versus longer term and how it may change kind of beyond '27?

James Anderson

Yes. Thanks, Papa. Yes, the $15,000 basically represents -- if you take an example of a chip that needed 100 terabit of I/O and if you were using our full PhotonLink portfolio, which would include not just the lasers, but all the other components, isolators, thermoelectric coolers, laser modules, our full fiber assembly kit, all the connectors, et cetera. So everything that we do, if you use all of that, the content per chip would be we estimated about $15,000 per chip.

Now it will depend on the customer and their particular application, how much of that content they utilize, whether they utilize a portion of that, all of it, that will really depend on the particular engagement. But that $15,000 is sort of the opportunity that's out there for us. I think we'll see over the next year or 2 as PhotonLink ramps, I think we'll see customers that leverage the vast majority of that. And I think we'll also see customers that leverage a subset of that. So I think we'll definitely -- we'll see a mix of that.

And we're happy to support our customers, whether they want the entire solution or they want a portion of it, we'll work with them on whatever works best for them. But regardless of the amount of content they take, for us, this -- we see this as all greenfield growth because most of PhotonLink adoption over time will be driven by scale-up. And again, that's all greenfield growth for us because that's us and the industry converting copper links to optical links over time. So we see it as greenfield revenue growth.

Operator

The next question comes from Cody Acree with StoneX Financial.

Cody Grant Acree

Sherri, maybe for you, if we go back to the gross margin discussion. Can you maybe help to stratify the contribution part of that? Can you talk about the indium phosphide 6-inch ramp yields increasing utilization rates, maybe even your mix thoughts? And just if you can also factor into that the competitive dynamics on your module, on your transceiver piece of more global Chinese competition and what's inherently just a more price-sensitive market?

Sherri Luther

Yes. Sure, Cody. I think you cut out a little bit, but I think you were asking for more color on what the drivers are to gross margin expansion over 45%. That's what I think your question is.

Cody Grant Acree

Yes, yes, that's right.

Sherri Luther

Yes. Okay. So if I were to rank order, I think that the largest contributors are going to be our new product ramps and that full operational leverage of our 6-inch indium phosphide capacity and then pricing optimization, followed by cost reduction. That's kind of how I would rank order it. And the timing of some of those things kicking in, right, will vary.

But certainly, new product ramps. I mean those are gross margin accretive and new products that Jim referred to earlier in his prepared remarks. Those are gross margin accretive. And so that will benefit our gross margin. As we continue to shift our supply from external to internal, as Jim talked about a little bit ago, that's going to drive significant opportunity in our gross margin.

And then pricing optimization, rather, I mean, that is an area where I think we will continue to have benefit there fully in our Datacenter & Communications segment, certainly in our Industrial too, but definitely in Datacenter & Communications, there's opportunity for that. We saw that to be a large contributor during the current quarter as well as in prior quarters. And so that will continue to be a benefit as well to drive to over 45% and cost reductions that will always continue to be something that we focus on, but that's how I would rank order them.

Operator

The next question comes from Karl Ackerman with BNP Paribas.

Karl Ackerman

Jim, you spoke about how customers are beginning to book out to calendar 2028 that gives you confidence in your growth forecast. At the same time, there is investor concern about the financing and permitting of data centers here in the United States. In light of this, how does the team align its laser production capacity plans with your major customers? And has the team considered customer prepayments for additional laser production capacity?

James Anderson

Yes. Thanks, Karl. The answer to the last part of the question is yes, definitely. Maybe I'll start with the end of the question. So we've signed a number of LTAs over the past quarters. The one public LTA that is the one with NVIDIA that we signed back in March, but we've signed a number of other LTAs. The LTAs follow -- all follow a general pattern. And one part of the LTA is typically some sort of upfront payment from the customer either -- it could be either a prepayment or NRE. There's different forms of payment, but there's some sort of upfront payment from the customer.

And what we do that for is for 2 reasons. Number one is it makes sure that the customers have skin in the game. But number two is we use that upfront payment to help fund the CapEx to build the capacity for that customer, right? We think it's important that if we're going to go build the capacity that the customers either fund that or help co-fund that. And so far, we've had customers to be very, very willing to do that. And I think that's worked really well for us. And so I would say almost every LTA that we've signed has that general pattern. And so that certainly gives us better confidence and helps offset the risk in terms of our own capital deployment.

But I guess on the first part of your question, we certainly worked very, very closely with our customers on planning out the demand. I think one thing that -- to keep in mind with indium phosphide capacity, is remember that it's not just being driven by datacenter demand. There's 2 things driving optics -- in general, optics demand, right? Number one is datacenter growth. But number two is in the case of integrated optics, it's the conversion of the copper links to optical over time. And that is a technology trend that in my view is independent of the market demand. That's a technology trend that will happen and regardless of the datacenter demand because that's necessitated by the ability to achieve the bandwidth and the link speeds that we need to achieve in the scale-up domain or chip-to-chip is we have to switch to optics over time. And so that's an independent technology trend. And that's certainly part of what's driving the indium phosphide capacity demand over the long term.

Operator

The next question comes from Meta Marshall with Morgan Stanley.

Meta Marshall

Maybe just a question on substrate availability. I think there have been some concern in the market around availability of 6-inch versus 3-inch. I know you noted in the prepared remarks that you had gotten capacity, but maybe if you could just speak about kind of 6-inch versus 3-inch substrate availability and how that's impacting expansion plans.

James Anderson

Yes. Thanks, Meta. So we have over -- on 6-inch in particular, but 3-inch as well, we do continue to buy 3-inch wafers. We do still have 3-inch production, but all of our expansion is mostly on 6-inch moving forward.

And we have over 5 different suppliers of substrates. We've signed long-term agreements with a number of those suppliers. And when I look at our next step of more than doubling the capacity by the end of calendar '27, we believe we've secured the necessary substrates for that capacity expansion. And then we've certainly -- we have plans in place to further expand capacity beyond '27, and we're certainly working with or already have in place, substrate capacity for that as well. So I think the summary is we have multiple different suppliers, and we've made sure that we have the substrate capacity for the expansion.

Operator

The next question comes from Sean O'Loughlin with TD Cowen.

Sean O'Loughlin

Great. I'll make it quick. Jim, can you just talk about how fungible your capital installed base is across indium phosphide versus gallium arsenide, if we maybe we end up in a world that is more VCSEL or multi-core VCSEL heavy than it is ultrahigh-power CW?

James Anderson

Sure. I would say there's some fungibility, but not -- it's not like you can easily switch a line from indium phosphide to gallium arsenide or vice versa. But I will note that we have a tremendous amount of gallium arsenide capacity. One of the stats that we shared at the PhotonLink launch is that today, we have capacity to produce over 1 billion VCSEL arrays annually. So we have a tremendous amount of capacity that we could deploy.

And so to the extent that we saw a broad-based adoption and rapid ramp of VCSEL technology, I don't think we -- there's probably anybody else in the industry that's better positioned to source or to supply that demand than Coherent is. And so we feel really well positioned on gallium arsenide VCSEL technology and/or capacity. And then on indium phosphide, obviously, we're ramping that very, very quickly. And so I think on both, we're on a good path.

Operator

Our last question today comes from Ananda Baruah with Loop Capital.

Ananda Baruah

Congrats on the progress here. And I guess, Jim, let me ask you a capacity -- supply -- capacity supply question from the other direction. Yes, there's also been a conversation just in the financial community about at an industry level excess InP EML capacity at some point in calendar '28. You're also -- now you're also talking about sort of contract agreements going sort of well into that time frame. So I would love to get your opinion on sort of how you view the potential for an oversupply situation at the industry level going through '28.

James Anderson

Yes. I think through '27 and into '28, I don't see that as a risk, given that -- given not just the demand that pluggable transceivers is creating, but also the demand that integrated optics is going to create for lasers as well as there's quite a bit of indium phosphide, that's also used in some of the DCI applications, too. So when I look across all of that, I don't see an oversupply risk through calendar '28.

I also think for us because most of our indium phosphide gets used in our -- well, I guess today all of it, today, all of our indium phosphide gets used in our own products. And we don't even produce enough today for -- to meet all the demand of our own products. For us, I don't -- that's not a primary concern for us. If I think if we were -- if there were more supply externally, that may affect us if we were supplying indium phosphide lasers externally, but we're not. And I think our plan is to -- over time, start to use more indium phosphide -- more of our internal indium phosphide production in our -- to change to increase the mix towards internal over time. So yes, I guess the short answer to your question is we don't see any big oversupply situation into '28.

Operator

This concludes our question-and-answer session. I would like to turn the conference back over for any closing remarks.

James Anderson

Thanks, operator, and thanks, everybody, for joining us on the call today. So we certainly enter the remainder of the year with great momentum, exceptional customer demand and as we were just talking about rapidly expanding capacity. I just want to say thanks to -- especially to all of our employees, our customers, our partners and, of course, our shareholders for the continued support. And we look forward to updating you again next quarter. Thank you.

Operator

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.