Original | Translation |
---|---|
Credo Technology Group Holding Ltd (NASDAQ:CRDO) Q1 2026 Earnings Call September 3, 2025 5:00 PM EDT Company Participants Daniel O'Neil - Vice President of Corporate Development & Investor Relations William Brennan - President, CEO & Chairman Daniel Fleming - Chief Financial Officer Conference Call Participants Karl Ackerman - BNP Paribas Exane, Research Division Vivek Arya - BofA Securities, Research Division Thomas O'Malley - Barclays Bank PLC, Research Division Joshua Buchalter - TD Cowen, Research Division Tore Svanberg - Stifel, Nicolaus & Company, Incorporated, Research Division Christopher Rolland - Susquehanna Financial Group, LLLP, Research Division Vijay Rakesh - Mizuho Securities USA LLC, Research Division Quinn Bolton - Needham & Company, LLC, Research Division Sujeeva De Silva - ROTH Capital Partners, LLC, Research Division Presentation Operator Ladies and gentlemen, thank you for standing by. | ν¬λ λ ν
ν¬λλ‘μ§ κ·Έλ£Ή νλ© λ¦¬λ―Έν°λ (NASDAQ:CRDO) 2026λ
λ 1λΆκΈ° μ€μ λ°ν 컨νΌλ°μ€ μ½ 2025λ
9μ 3μΌ μ€ν 5μ (λλΆνμ€μ) νμ¬ μ°Έμμ λ€λμ μ€λ - κΈ°μ κ°λ° λ° ν¬μμκ΄κ³ λ΄λΉ λΆμ¬μ₯ μ리μ λΈλ λ - μ¬μ₯, CEO κ²Έ νμ₯ λ€λμ νλ λ° - μ΅κ³ μ¬λ¬΄μ± μμ 컨νΌλ°μ€ μ½ μ°Έμμ μΉΌ μ μ»€λ§ - BNP νλ¦¬λ° μμΈμΈ, 리μμΉ λΆλ¬Έ λΉλ²‘ μλ¦¬μΌ - λ± ν¬μ€λΈμλ©λ¦¬μΉ΄ μ¦κΆ, 리μμΉ λΆλ¬Έ ν λ§μ€ μ€λ§λ¦¬ - λ°ν΄λ μ΄μ¦ μν PLC, 리μμΉ λΆλ¬Έ μ‘°μμ λΆμ°°ν° - TD μ½μ¬, 리μμΉ λΆλ¬Έ ν λ μ€λ°λ²κ·Έ - μ€ν°ν , λμ½λΌμ°μ€ μ€ μ»΄νΌλ, μΈμ½νΌλ μ΄ν°λ, 리μμΉ λΆλ¬Έ ν¬λ¦¬μ€ν νΌ λ‘€λλ - μμ€μΌνλ νμ΄λΈμ κ·Έλ£Ή, LLLP, 리μμΉ λΆλ¬Έ λΉμ μ΄ λΌμΌμ - λ―Έμ¦νΈ μ¦κΆ USA LLC, 리μμΉ λΆλ¬Έ νΈ λ³Όν΄ - λλ€ μ€ μ»΄νΌλ, LLC, 리μμΉ λΆλ¬Έ μμ§λ° λ μ€λ° - λ‘μ€ μΊνΌνΈ ννΈλμ€, LLC, 리μμΉ λΆλ¬Έ λ°ν μ€νΌλ μ΄ν° μ μ¬ μλ μ¬λ¬λΆ, λκΈ°ν΄ μ£Όμ μ κ°μ¬ν©λλ€. |
[Operator Instructions] I would now like to turn the conference over to Dan O'Neil. Please go ahead, sir. Daniel O'Neil Vice President of Corporate Development & Investor Relations Good afternoon. Thank you for joining our earnings call for the first quarter of fiscal 2026. Today, I'm joined by Bill Brennan, Credos' Chief Executive Officer; and Dan Fleming, Credo's Chief Financial Officer. During this call, we will make certain forward-looking statements. These forward-looking statements are subject to risks and uncertainties discussed in detail in our documents filed with the SEC, which can be found in the Investor Relations section of the company's website. | [μ΄μμ μλ΄] μ΄μ νμλ₯Ό λ μ€λμκ² λκΈ°κ² μ΅λλ€. λ§μν΄ μ£Όμμμ€. λ€λμ μ€λ κΈ°μ κ°λ° λ° ν¬μμκ΄κ³ λ΄λΉ λΆμ¬μ₯ μλ νμΈμ. 2026 νκ³μ°λ 1λΆκΈ° μ€μ λ°ν 컨νΌλ°μ€ μ½μ μ°Έμ¬ν΄ μ£Όμ μ κ°μ¬ν©λλ€. μ€λ μ μ ν¨κ» ν¬λ λμ μ΅κ³ κ²½μμμΈ λΉ λΈλ λκ³Ό ν¬λ λμ μ΅κ³ μ¬λ¬΄μ± μμμΈ λ νλ λ°μ΄ μ°Έμνμ΅λλ€. μ΄λ² ν΅νμμ μ ν¬λ νΉμ λ―ΈλμμΈ‘μ± μ§μ μ ν μμ μ λλ€. μ΄λ¬ν λ―ΈλμμΈ‘μ± μ§μ μ SECμ μ μΆλ μ ν¬ λ¬Έμμ μμΈν λ Όμλ μνκ³Ό λΆνμ€μ±μ μν₯μ λ°μΌλ©°, ν΄λΉ λ¬Έμλ νμ¬ μΉμ¬μ΄νΈμ ν¬μμκ΄κ³ μΉμ μμ νμΈνμ€ μ μμ΅λλ€. |
It is not possible for the company's management to predict all risks nor can the company assess the impact of all factors on its business or the extent to which any factor or combination of factors may cause actual results to differ materially from those contained in any forward-looking statement. Given these risks, uncertainties and assumptions, the forward-looking events discussed during this call may not occur, and actual results could differ materially and adversely from those anticipated or implied. | νμ¬ κ²½μμ§μ΄ λͺ¨λ 리μ€ν¬λ₯Ό μμΈ‘νλ κ²μ λΆκ°λ₯νλ©°, νμ¬λ λͺ¨λ μμΈμ΄ μ¬μ μ λ―ΈμΉλ μν₯μ΄λ μ΄λ€ μμΈ λλ μμΈλ€μ μ‘°ν©μ΄ μ€μ κ²°κ³Όλ‘ νμ¬κΈ λ―ΈλμμΈ‘μ§μ μ ν¬ν¨λ λ΄μ©κ³Ό μ€λνκ² λ€λ₯΄κ² λ§λ€ μ μλ μ λλ₯Ό νκ°ν μ μμ΅λλ€. μ΄λ¬ν 리μ€ν¬, λΆνμ€μ± λ° κ°μ μ κ³ λ €ν λ, λ³Έ 컨νΌλ°μ€ μ½μμ λ Όμλ λ―ΈλμμΈ‘ μ¬κ±΄λ€μ λ°μνμ§ μμ μ μμΌλ©°, μ€μ κ²°κ³Όλ μμλκ±°λ μμλ κ²κ³Ό μ€λνκ³ λΆμ μ μΌλ‘ λ€λ₯Ό μ μμ΅λλ€. |
The company undertakes no obligation to publicly update forward-looking statements for any reason after the date of this call to conform these statements to actual results or to changes in the company's expectations, except as required by law. Also, during this call, we will refer to certain non-GAAP financial measures, which we consider to be important measures of the company's performance. These non-GAAP financial measures are provided in addition to and not as a substitute for or superior to financial performance prepared in accordance with U.S. GAAP. | νμ¬λ λ²λ₯ μμ μꡬνλ κ²½μ°λ₯Ό μ μΈνκ³ λ, λ³Έ 컨νΌλ°μ€μ½ μΌμ μ΄ν μ€μ κ²°κ³Όλ νμ¬μ κΈ°λμΉ λ³νμ λ§μΆ° λ―Έλμ λ§μ§μ μ 곡κ°μ μΌλ‘ μ λ°μ΄νΈν μ무λ₯Ό μ§μ§ μμ΅λλ€. λν λ³Έ 컨νΌλ°μ€μ½μμλ νμ¬ μ±κ³Όμ μ€μν μ§νλ‘ κ°μ£Όλλ νΉμ λΉGAAP μ¬λ¬΄μ§νλ€μ μΈκΈν μμ μ λλ€. μ΄λ¬ν λΉGAAP μ¬λ¬΄μ§νλ€μ λ―Έκ΅ GAAPμ λ°λΌ μμ±λ μ¬λ¬΄μ±κ³Όλ₯Ό λ체νκ±°λ μ΄λ³΄λ€ μ°μν κ²μ΄ μλλΌ μ΄μ μΆκ°λ‘ μ 곡λλ κ²μ λλ€. |
A discussion of why we use non-GAAP financial measures and reconciliations between our GAAP and non-GAAP financial measures is available in the earnings release we issued today, which can be accessed using the Investor Relations portion of our website. I will now turn the call over to our CEO. Bill? William Brennan President, CEO & Chairman Thanks, Dan, and thanks, everyone, for joining our earnings call for the first quarter of fiscal '26. I'll begin with an overview of our first quarter results and then discuss our forward outlook. After my comments, our Chief Financial Officer, Dan Fleming, will provide financial details for Q1 and our guidance for the second quarter. | λΉGAAP μ¬λ¬΄μ§νλ₯Ό μ¬μ©νλ μ΄μ μ GAAP λ° λΉGAAP μ¬λ¬΄μ§ν κ°μ μ‘°μ νμ λν λ
Όμλ μ€λ λ°νν μ€μ λ°νμλ£μμ νμΈνμ€ μ μμΌλ©°, λΉμ¬ μΉμ¬μ΄νΈμ ν¬μμ κ΄κ³ μΉμ
μ ν΅ν΄ μ κ·Όνμ€ μ μμ΅λλ€. μ΄μ CEOμκ² λ°μΈκΆμ λκΈ°κ² μ΅λλ€. λΉ? μ리μ λΈλ λ μ¬μ₯, CEO κ²Έ νμ₯ κ°μ¬ν©λλ€, λ. κ·Έλ¦¬κ³ 26νκ³μ°λ 1λΆκΈ° μ€μ λ°ν 컨νΌλ°μ€ μ½μ μ°Έμ¬ν΄ μ£Όμ λͺ¨λ λΆλ€κ» κ°μ¬λ립λλ€. λ¨Όμ 1λΆκΈ° μ€μ κ°μλ₯Ό λ§μλλ¦° ν ν₯ν μ λ§μ λν΄ λ Όμνκ² μ΅λλ€. μ λ°μΈ νμλ μ΅κ³ μ¬λ¬΄μ± μμμΈ λ νλ λ°μ΄ 1λΆκΈ° μ¬λ¬΄ μΈλΆμ¬νκ³Ό 2λΆκΈ° κ°μ΄λμ€λ₯Ό μ 곡ν μμ μ λλ€. |
In the first quarter, we delivered revenue of $223 million, an increase of 31% sequentially and an increase of 274% year-over-year. Our non-GAAP gross margin was 67.6%, and we achieved nearly $100 million of non-GAAP net income. Demand for Credo's reliable and power-efficient high-speed connectivity solutions continues to ramp as hyperscalers and data center operators accelerate investments in AI-driven infrastructure. We provide state-of-the-art solutions for the most demanding connectivity needs, supporting data rates up to 1.6 terabits per second across a range of industry protocols. | 1λΆκΈ°μ μ ν¬λ 2μ΅ 2,300λ§ λ¬λ¬μ λ§€μΆμ λ¬μ±νμΌλ©°, μ΄λ μ λΆκΈ° λλΉ 31% μ¦κ°, μ λ λκΈ° λλΉ 274% μ¦κ°ν μμΉμ λλ€. λΉGAAP λ§€μΆμ΄μ΄μ΅λ₯ μ 67.6%λ₯Ό κΈ°λ‘νμΌλ©°, λΉGAAP μμ΄μ΅μ κ±°μ 1μ΅ λ¬λ¬μ λ¬νμ΅λλ€. νμ΄νΌμ€μΌμΌλ¬λ€κ³Ό λ°μ΄ν°μΌν° μ΄μμ 체λ€μ΄ AI κΈ°λ° μΈνλΌμ λν ν¬μλ₯Ό κ°μνν¨μ λ°λΌ, Credoμ μμ μ μ΄κ³ μ λ ₯ ν¨μ¨μ μΈ κ³ μ μ°κ²° μ루μ μ λν μμκ° μ§μμ μΌλ‘ μ¦κ°νκ³ μμ΅λλ€. μ ν¬λ κ°μ₯ κΉλ€λ‘μ΄ μ°κ²°μ± μꡬμ¬νμ μν μ΅μ²¨λ¨ μ루μ μ μ 곡νλ©°, λ€μν μ κ³ νλ‘ν μ½μμ μ΄λΉ μ΅λ 1.6ν λΌλΉνΈμ λ°μ΄ν° μ μ‘λ₯ μ μ§μν©λλ€. |
Credo's growth has been fueled by strategic partnerships with hyperscalers and key customers built on our ability to tackle their most complex connectivity challenges. We achieved this by delivering optimized solutions that span the design, development, qualification and production across our entire product line. Our 3-tiered innovation framework comprising of purpose-built SerDes technology, advanced integrated circuit design and system-level development approach integrates seamlessly with our pilot software and firmware platform, empowering customers to streamline system development and achieve peak performance, yield and reliability. | ν¬λ λμ μ±μ₯μ νμ΄νΌμ€μΌμΌλ¬λ€κ³Ό μ£Όμ κ³ κ°λ€κ³Όμ μ λ΅μ ννΈλμμ μν΄ κ²¬μΈλμ΄ μμΌλ©°, μ΄λ κ·Έλ€μ κ°μ₯ 볡μ‘ν μ°κ²°μ± κ³Όμ λ₯Ό ν΄κ²°ν μ μλ λΉμ¬μ μλμ λ°νμΌλ‘ ꡬμΆλμμ΅λλ€. λΉμ¬λ μ 체 μ ν λΌμΈμ κ±Έμ³ μ€κ³, κ°λ°, κ²μ¦ λ° μμ°μ μμ°λ₯΄λ μ΅μ νλ μ루μ μ μ 곡ν¨μΌλ‘μ¨ μ΄λ₯Ό λ¬μ±νμ΅λλ€. λͺ©μ λ³ λ§μΆ€ν SerDes κΈ°μ , κ³ κΈ μ§μ νλ‘ μ€κ³, κ·Έλ¦¬κ³ μμ€ν λ 벨 κ°λ° μ κ·Όλ²μΌλ‘ ꡬμ±λ λΉμ¬μ 3λ¨κ³ νμ νλ μμν¬λ νμΌλΏ μννΈμ¨μ΄ λ° νμ¨μ΄ νλ«νΌκ³Ό μλ²½νκ² ν΅ν©λμ΄, κ³ κ°λ€μ΄ μμ€ν κ°λ°μ κ°μννκ³ μ΅κ³ μ μ±λ₯, μμ¨ λ° μ λ’°μ±μ λ¬μ±ν μ μλλ‘ μ§μν©λλ€. |
Our innovative system-level approach has driven our leadership in pioneering the active electrical cable or AEC market. And looking ahead, we are applying this proven strategy to pursue additional system-level opportunities driven by the demand for better reliability, energy efficiency and performance. We expect our approach will lead to continued diversification in terms of customers, protocols and applications. We look forward to announcements over the next several months, including at upcoming trade shows. I'll now discuss our business in more detail. First, regarding AECs, our AEC product line continued its robust growth, driven by an increasingly diverse customer base. | μ ν¬μ νμ μ μΈ μμ€ν λ 벨 μ κ·Όλ²μ λ₯λν μ κΈ° μΌμ΄λΈ(AEC) μμ₯μ κ°μ²νλ λ° μμ΄ μ ν¬μ 리λμμ 견μΈν΄ μμ΅λλ€. κ·Έλ¦¬κ³ μμΌλ‘ μ ν¬λ μ΄ κ²μ¦λ μ λ΅μ μ μ©νμ¬ λ λμ μ λ’°μ±, μλμ§ ν¨μ¨μ± λ° μ±λ₯μ λν μμμ μν΄ μ£Όλλλ μΆκ°μ μΈ μμ€ν λ 벨 κΈ°νλ€μ μΆκ΅¬ν κ²μ λλ€. μ ν¬μ μ κ·Όλ²μ΄ κ³ κ°, νλ‘ν μ½ λ° μ ν리μΌμ΄μ μΈ‘λ©΄μμ μ§μμ μΈ λ€κ°νλ‘ μ΄μ΄μ§ κ²μΌλ‘ κΈ°λν©λλ€. λ€κ°μ€λ μ μνλ₯Ό ν¬ν¨νμ¬ ν₯ν λͺ λ¬κ° λ°νλ λ΄μ©λ€μ κΈ°λν΄ μ£ΌμκΈ° λ°λλλ€. μ΄μ μ ν¬ μ¬μ μ λν΄ λ μμΈν λ§μλλ¦¬κ² μ΅λλ€. λ¨Όμ AECμ κ΄λ ¨νμ¬, μ ν¬ AEC μ ν λΌμΈμ μ μ λ λ€μν΄μ§λ κ³ κ° κΈ°λ°μ νμ μ΄ κ²¬κ³ ν μ±μ₯μ μ§μνμ΅λλ€. |
Three hyperscalers each contributed over 10% of our revenue, and we expect our customer diversification to continue to broaden over the upcoming quarters. Based on customer forecasts, we anticipate significant year-over-year growth. While shipment timing may lead to nonlinear growth patterns at a customer level, we see every data center partner scaling their deployments. We're making strong progress with new customers as well, highlighted by the first material revenue contribution from a fourth hyperscaler in Q1. We anticipate this revenue to grow throughout the fiscal year, further strengthening our market position. The adoption of AECs continues to gain traction across the industry. | 3κ°μ νμ΄νΌμ€μΌμΌλ¬κ° κ°κ° λΉμ¬ λ§€μΆμ 10% μ΄μμ κΈ°μ¬νμΌλ©°, ν₯ν λΆκΈ°μ κ±Έμ³ κ³ κ° λ€λ³νκ° μ§μμ μΌλ‘ νλλ κ²μΌλ‘ μμν©λλ€. κ³ κ° μ λ§μ λ°νμΌλ‘ μ λ λλΉ μλΉν μ±μ₯μ μμνκ³ μμ΅λλ€. μΆν μκΈ°λ‘ μΈν΄ κ³ κ° μμ€μμ λΉμ νμ μ±μ₯ ν¨ν΄μ΄ λνλ μ μμ§λ§, λͺ¨λ λ°μ΄ν°μΌν° ννΈλλ€μ΄ λ°°μΉ κ·λͺ¨λ₯Ό νλνκ³ μλ κ²μ νμΈνκ³ μμ΅λλ€. μ κ· κ³ κ°κ³Όλ κ°λ ₯ν μ§μ μ 보μ΄κ³ μμΌλ©°, μ΄λ 1λΆκΈ°μ λ€ λ²μ§Έ νμ΄νΌμ€μΌμΌλ¬λ‘λΆν° 첫 λ²μ§Έ μ€μ§μ μΈ λ§€μΆ κΈ°μ¬λ‘ λΆκ°λμμ΅λλ€. μ΄ λ§€μΆμ νκ³μ°λ μ λ°μ κ±Έμ³ μ±μ₯ν κ²μΌλ‘ μμνλ©°, μ΄λ₯Ό ν΅ν΄ λΉμ¬μ μμ₯ μ§μκ° λμ± κ°νλ κ²μ λλ€. AECμ μ±νμ μ κ³ μ λ°μ κ±Έμ³ μ§μμ μΌλ‘ 견μΈλ ₯μ μ»κ³ μμ΅λλ€. |
AECs offer an unequaled combination of reliability, signal integrity, power efficiency and system cost, all critical to building and scaling leading -edge AI clusters. We've seen AEC adoption at data rates of 50 gig and 100 gig per lane, and we see this continuing for 200 gig per lane, 1.6 terabit per second solutions as next-generation architectures ramp. We also see the trend towards GPU and cluster densification to continue to be a catalyst for an expanding AEC TAM. Over the past year, we've seen customer interest for AECs expand from intra-rack solutions to rack-to-rack solutions. | AECλ μ΅μ²¨λ¨ AI ν΄λ¬μ€ν°λ₯Ό ꡬμΆνκ³ νμ₯νλ λ° μ€μν μ λ’°μ±, μ νΈ λ¬΄κ²°μ±, μ λ ₯ ν¨μ¨μ± λ° μμ€ν λΉμ©μ νμν μ‘°ν©μ μ 곡ν©λλ€. μ°λ¦¬λ λ μΈλΉ 50κΈ°κ° λ° 100κΈ°κ° λ°μ΄ν° μλμμ AEC μ±νμ νμΈνμΌλ©°, μ°¨μΈλ μν€ν μ²κ° νμ°λ¨μ λ°λΌ λ μΈλΉ 200κΈ°κ°, μ΄λΉ 1.6ν λΌλΉνΈ μ루μ μμλ μ΄λ¬ν μΆμΈκ° μ§μλ κ²μΌλ‘ λ³΄κ³ μμ΅λλ€. λν GPUμ ν΄λ¬μ€ν° κ³ λ°λν μΆμΈκ° AEC TAM νμ₯μ μ΄λ§€μ μν μ κ³μν κ²μΌλ‘ μμν©λλ€. μ§λ 1λ λμ AECμ λν κ³ κ° κ΄μ¬μ΄ λ λ΄λΆ μ루μ μμ λ κ° μ루μ μΌλ‘ νλλλ κ²μ νμΈνμ΅λλ€. |
Advances in liquid cooling and power sourcing have driven a quadrupling of GPU density with one customer, which enabled them to architect their entire scale-out network with AECs up to 7 meters in length. Reliability and power efficiency led to choosing AECs over optical solutions as they are up to 1,000x more reliable and consume half the power. AECs virtually eliminate link fabs, which are intermittent losses of connection, boosting cluster reliability and productivity while reducing power consumption. Our system-level approach drives innovation, accelerates time to market and delivers a distinct competitive edge. | μ‘체 λκ° λ° μ μ κ³΅κΈ κΈ°μ μ λ°μ μΌλ‘ ν κ³ κ°μ¬μ GPU λ°λκ° 4λ°° μ¦κ°νμΌλ©°, μ΄λ₯Ό ν΅ν΄ μ΅λ 7λ―Έν° κΈΈμ΄μ AECλ₯Ό μ¬μ©νμ¬ μ 체 μ€μΌμΌμμ λ€νΈμν¬λ₯Ό μ€κ³ν μ μκ² λμμ΅λλ€. μ λ’°μ±κ³Ό μ λ ₯ ν¨μ¨μ±μΌλ‘ μΈν΄ κ΄ν μ루μ λμ AECλ₯Ό μ ννκ² λμλλ°, AECλ κ΄ν μ루μ λλΉ μ΅λ 1,000λ°° λ λμ μ λ’°μ±μ μ 곡νκ³ μ λ ₯ μλΉλ μ λ° μμ€μ λλ€. AECλ κ°νμ μΈ μ°κ²° μμ€μΈ λ§ν¬ μ₯μ λ₯Ό μ¬μ€μ μ κ±°νμ¬ ν΄λ¬μ€ν° μ λ’°μ±κ³Ό μμ°μ±μ ν₯μμν€λ λμμ μ λ ₯ μλΉλ₯Ό μ€μ λλ€. λΉμ¬μ μμ€ν λ 벨 μ κ·Ό λ°©μμ νμ μ μ£Όλνκ³ , μΆμ κΈ°κ°μ λ¨μΆνλ©°, λλ ·ν κ²½μ μ°μλ₯Ό μ 곡ν©λλ€. |
By owning and delivering the entire solution stack, including series IP, retimer ICs, system-level design, qualification and production, Credo has forged strong customer relationships and solidified its position as the leader in the AEC market. Moving forward, we'll continue driving innovation with PCIe AECs and other advanced products coming to market in the near future. Let me now turn to the optical market. During the first quarter, we sustained strong momentum in our optical business. positioning us is on track to achieve our goal of again doubling optical revenue in fiscal '26. | μ 체 μ루μ μ€νμ μμ νκ³ μ 곡ν¨μΌλ‘μ¨ - μλ¦¬μ¦ IP, 리νμ΄λ¨Έ IC, μμ€ν λ 벨 μ€κ³, κ²μ¦ λ° μμ°μ ν¬ν¨νμ¬ - Credoλ κ°λ ₯ν κ³ κ° κ΄κ³λ₯Ό ꡬμΆνκ³ AEC μμ₯μ 리λλ‘μμ μ μ§λ₯Ό νκ³ ν νμ΅λλ€. μμΌλ‘λ PCIe AECμ κ°κΉμ΄ λ―Έλμ μΆμλ κΈ°ν μ²¨λ¨ μ νλ€λ‘ μ§μμ μΈ νμ μ μΆμ§ν΄ λκ° κ²μ λλ€. μ΄μ κ΄ν μμ₯μΌλ‘ λμ΄κ°κ² μ΅λλ€. 1λΆκΈ° λμ λΉμ¬λ κ΄ν μ¬μ μμ κ°λ ₯ν λͺ¨λ©ν μ μ μ§νμΌλ©°, μ΄λ 26νκ³μ°λμ κ΄ν λ§€μΆμ λ€μ λ λ°°λ‘ λλ¦¬κ² λ€λ λͺ©ν λ¬μ± κΆ€λμ μ¬λΌμ°μμ μλ―Έν©λλ€. |
We're delivering cutting-edge DSP solutions to an expanding roster of optical module customers and their hyperscale end users. With our expertise in optical DSPs, we're meeting the rigorous connectivity demands of next-generation applications. Our collaborative approach with customers helps ensure solutions that enhance performance, scalability and power efficiency, further building on our position in the market. We provide customers with a leading-edge portfolio of innovative full DSP and linear receive optical or LRO solutions, supporting port speeds up to 1.6 terabits per second. For several years, the industry has debated the shift from copper to optical connectivity solutions. | μ ν¬λ νμ₯λκ³ μλ κ΄λͺ¨λ κ³ κ°μ¬λ€κ³Ό κ·Έλ€μ νμ΄νΌμ€μΌμΌ μ΅μ’ μ¬μ©μλ€μκ² μ΅μ²¨λ¨ DSP μ루μ μ μ 곡νκ³ μμ΅λλ€. κ΄ DSP λΆμΌμ μ λ¬Έμ±μ λ°νμΌλ‘ μ°¨μΈλ μ ν리μΌμ΄μ μ κΉλ€λ‘μ΄ μ°κ²°μ± μꡬμ¬νμ μΆ©μ‘±νκ³ μμ΅λλ€. κ³ κ°κ³Όμ νλ ₯μ μ κ·Ό λ°©μμ ν΅ν΄ μ±λ₯, νμ₯μ± λ° μ λ ₯ ν¨μ¨μ±μ ν₯μμν€λ μ루μ μ 보μ₯νλ©°, μμ₯μμμ μ μ§λ₯Ό λμ± κ³΅κ³ ν νκ³ μμ΅λλ€. μ ν¬λ κ³ κ°λ€μκ² μ΄λΉ 1.6ν λΌλΉνΈκΉμ§μ ν¬νΈ μλλ₯Ό μ§μνλ νμ μ μΈ ν DSP λ° μ ν μμ κ΄ν(LRO) μ루μ μΌλ‘ ꡬμ±λ μ λμ μΈ ν¬νΈν΄λ¦¬μ€λ₯Ό μ 곡ν©λλ€. μλ κ° μ κ³μμλ ꡬ리μμ κ΄ μ°κ²° μ루μ μΌλ‘μ μ νμ λν΄ λ Όμν΄ μμ΅λλ€. |
While consensus holds that copper will remain prevalent in the foreseeable future, Credo is strategically prioritizing optical solutions as a cornerstone of our product road map. Credo sees an expanding TAM for both copper and optical connectivity solutions, and we're excited by the opportunity to bring our system-level expertise to bear in the optical market, further diversifying our position. We look forward to discussing these innovations in the coming months. And now regarding our retimer business. In Q1 fiscal '26, our Ethernet retimer business achieved strong results. | κ΅¬λ¦¬κ° κ°κΉμ΄ λ―Έλμλ κ³μ μ£Όλ₯λ₯Ό μ΄λ£° κ²μ΄λΌλ μμ₯ 컨μΌμμ€κ° μμ§λ§, Credoλ κ΄ν μ루μ μ λΉμ¬ μ ν λ‘λλ§΅μ ν΅μ¬μΌλ‘ μ λ΅μ μΌλ‘ μ°μ μμλ₯Ό λκ³ μμ΅λλ€. Credoλ ꡬ리μ κ΄ν μ°κ²° μ루μ λͺ¨λμμ TAMμ΄ νλλκ³ μλ€κ³ λ³΄κ³ μμΌλ©°, κ΄ν μμ₯μμ λΉμ¬μ μμ€ν λ 벨 μ λ¬Έμ±μ λ°ννμ¬ λΉμ¬μ ν¬μ§μ μ λμ± λ€κ°νν μ μλ κΈ°νμ λν΄ κΈ°λνκ³ μμ΅λλ€. μμΌλ‘ λͺ λ¬ λμ μ΄λ¬ν νμ μ λν΄ λ Όμν μ μκΈ°λ₯Ό κΈ°λν©λλ€. μ΄μ λΉμ¬μ 리νμ΄λ¨Έ μ¬μ μ λν΄ λ§μλλ¦¬κ² μ΅λλ€. 26νκ³μ°λ 1λΆκΈ°μ λΉμ¬μ μ΄λλ· λ¦¬νμ΄λ¨Έ μ¬μ μ κ°λ ₯ν μ€μ μ λ¬μ±νμ΅λλ€. |
Our retimers are recognized for their exceptional performance and energy efficiency, featuring advanced MACsec encryption and gearbox functionality. The Ethernet retimer market encompasses traditional switching applications and the emerging category of AI appliances. Our recently launched PCIe retimer family is gaining significant traction, driven by robust customer engagements. Leveraging the expertise of our world-class SerDes design team, our PCIe solutions deliver an unparalleled combination of maximum reach and minimal latency, a rare achievement as these attributes are typically a trade-off. | λΉμ¬μ 리νμ΄λ¨Έλ λ°μ΄λ μ±λ₯κ³Ό μλμ§ ν¨μ¨μ±μΌλ‘ μΈμ λ°κ³ μμΌλ©°, κ³ κΈ MACsec μνΈν λ° κΈ°μ΄λ°μ€ κΈ°λ₯μ νΉμ§μΌλ‘ ν©λλ€. μ΄λλ· λ¦¬νμ΄λ¨Έ μμ₯μ κΈ°μ‘΄μ μ€μμΉ μ ν리μΌμ΄μ κ³Ό μλ‘κ² λΆμνλ AI μ΄νλΌμ΄μΈμ€ μΉ΄ν κ³ λ¦¬λ₯Ό ν¬κ΄ν©λλ€. μ΅κ·Ό μΆμν λΉμ¬μ PCIe 리νμ΄λ¨Έ μ νκ΅°μ νλ°ν κ³ κ° μ°Έμ¬μ νμ μ΄ μλΉν 견μΈλ ₯μ μ»κ³ μμ΅λλ€. μΈκ³μ μμ€μ SerDes μ€κ³νμ μ λ¬Έμ±μ νμ©νμ¬, λΉμ¬μ PCIe μ루μ μ μ΅λ λλ¬κ±°λ¦¬μ μ΅μ μ§μ°μκ°μ νμν μ‘°ν©μ μ 곡νλλ°, μ΄λ μΌλ°μ μΌλ‘ μ΄λ¬ν νΉμ±λ€μ΄ μμΆ©κ΄κ³μ μλ€λ μ μμ λ§€μ° λλ¬Έ μ±κ³Όμ λλ€. |
Our engineering team's innovative design approach has resulted in a truly differentiated solution. Our top-tier devices, coupled with our award-winning pilot debug and telemetry tools, empower customers to integrate our solutions with faster time to market and better reliability. We are on track to secure PCIe design wins in calendar '25 with production revenue expected in calendar '26. Our expansion into PCIe-based solutions for AI scale-up networks significantly broadens our TAM, and we are well positioned to capitalize on the industry shift to 200-gig per lane scale-up solutions in the future, ensuring sustained growth and competitiveness. | μ ν¬ μμ§λμ΄λ§ νμ νμ μ μΈ μ€κ³ μ κ·Όλ²μ μ§μ μΌλ‘ μ°¨λ³νλ μ루μ μ ꡬνν΄λμ΅λλ€. μ ν¬μ μ΅κ³ κΈ λλ°μ΄μ€μ μμ κ²½λ ₯μ λΉλλ νμΌλΏ λλ²κ·Έ λ° ν λ λ©νΈλ¦¬ ν΄μ κ²°ν©νμ¬, κ³ κ°λ€μ΄ λ λΉ λ₯Έ μΆμ μκ°κ³Ό ν₯μλ μ λ’°μ±μΌλ‘ μ ν¬ μ루μ μ ν΅ν©ν μ μλλ‘ μ§μνκ³ μμ΅λλ€. μ ν¬λ 2025λ μ€ PCIe μ€κ³ μμ£Όλ₯Ό ν보ν μμ μ΄λ©°, 2026λ μ€ μμ° λ§€μΆμ΄ κΈ°λλ©λλ€. AI μ€μΌμΌμ λ€νΈμν¬λ₯Ό μν PCIe κΈ°λ° μ루μ μΌλ‘μ νμ₯μ μ ν¬ TAMμ ν¬κ² νλμν€κ³ μμΌλ©°, ν₯ν λ μΈλΉ 200κΈ°κ° μ€μΌμΌμ μ루μ μΌλ‘μ μ κ³ μ νμ νμ©ν μ μλ μ 리ν μμΉμ μμ΄ μ§μμ μΈ μ±μ₯κ³Ό κ²½μλ ₯μ 보μ₯νκ³ μμ΅λλ€. |
To summarize, during the first quarter of fiscal '26, Credo reported its highest quarterly revenue and profitability to date, reflecting exceptional operational execution to meet customer demand. Over the past several quarters, Credo has achieved extraordinary growth fueled by the surging demand for AI infrastructure. Credo is poised for continued success with a clear path for growth through fiscal '26 and beyond. Over the medium and long term, we anticipate multiple waves of growth opportunities, fueled by the evolving scale-out and scale-up networks and next-generation training and inference architectures. | μμ½νμλ©΄, 26νκ³μ°λ 1λΆκΈ° λμ Credoλ κ³ κ° μμλ₯Ό μΆ©μ‘±νκΈ° μν νμν μ΄μ μ€νλ ₯μ λ°μνμ¬ λΆκΈ° κΈ°μ€ μ¬μ μ΅κ³ λ§€μΆκ³Ό μμ΅μ±μ κΈ°λ‘νμ΅λλ€. μ§λ λͺ λΆκΈ°μ κ±Έμ³ Credoλ AI μΈνλΌμ λν κΈμ¦νλ μμμ νμ μ΄ λλΌμ΄ μ±μ₯μ λ¬μ±νμ΅λλ€. Credoλ 26νκ³μ°λμ κ·Έ μ΄νκΉμ§ λͺ νν μ±μ₯ κ²½λ‘λ₯Ό ν΅ν΄ μ§μμ μΈ μ±κ³΅μ μν μ€λΉκ° λμ΄ μμ΅λλ€. μ€μ₯κΈ°μ μΌλ‘λ μ§ννλ μ€μΌμΌμμ λ° μ€μΌμΌμ λ€νΈμν¬μ μ°¨μΈλ νλ ¨ λ° μΆλ‘ μν€ν μ²μ νμ μ΄ μ¬λ¬ μ°¨λ‘μ μ±μ₯ κΈ°νλ₯Ό μμνκ³ μμ΅λλ€. |
We expect that each of these waves will drive growing demand for innovative connectivity solutions, spanning diverse physical mediums, distances and protocols. Credo occupies a unique position in the industry as one of the few companies worldwide capable of delivering cutting-edge SerDes technology at the advanced speeds we are currently qualifying. By combining this core SerDes differentiation central to our competitive edge with our integrated circuit expertise and system-level design approach, we deliver solutions optimized to meet each customer's needs. Credo has become a high-value partner in the unprecedented build-out of hyperscale infrastructure. | μ°λ¦¬λ μ΄λ¬ν κ°κ°μ λ¬Όκ²°μ΄ λ€μν 물리μ 맀체, 거리 λ° νλ‘ν μ½μ κ±ΈμΉ νμ μ μΈ μ°κ²° μ루μ μ λν μμ μ¦κ°λ₯Ό 견μΈν κ²μΌλ‘ μμν©λλ€. ν¬λ λλ νμ¬ κ²μ¦ μ€μΈ μ²¨λ¨ μλμμ μ΅μ²¨λ¨ SerDes κΈ°μ μ μ 곡ν μ μλ μ μΈκ³ λͺ μ λλ κΈ°μ μ€ νλλ‘μ μ κ³μμ λ νΉν μμΉλ₯Ό μ°¨μ§νκ³ μμ΅λλ€. μ°λ¦¬μ κ²½μ μ°μμ ν΅μ¬μΈ μ΄λ¬ν ν΅μ¬ SerDes μ°¨λ³ν κΈ°μ μ μ§μ νλ‘ μ λ¬Έμ± λ° μμ€ν μμ€μ μ€κ³ μ κ·Όλ²κ³Ό κ²°ν©ν¨μΌλ‘μ¨, μ°λ¦¬λ κ° κ³ κ°μ μꡬμ¬νμ μΆ©μ‘±νλλ‘ μ΅μ νλ μ루μ μ μ 곡ν©λλ€. ν¬λ λλ μ λ‘ μλ νμ΄νΌμ€μΌμΌ μΈνλΌ κ΅¬μΆμμ κ³ λΆκ°κ°μΉ ννΈλκ° λμμ΅λλ€. |
I believe our strong results today demonstrate our potential, and I'm increasingly optimistic about the opportunities that lie ahead. Our Chief Financial Officer, Dan Fleming, will now provide additional details on our financial results. Daniel Fleming Chief Financial Officer Thank you, Bill, and good afternoon. I will first review our Q1 results and then discuss our outlook for Q2 of fiscal year '26. In Q1, we reported revenue of $223.1 million, up 31% sequentially and up 274% year-over-year and well above the high end of our guidance range. Our product business generated $217.1 million of revenue in Q1, up 31% sequentially and up 279% year-over-year. | μ€λμ κ°λ ₯ν μ€μ μ΄ μ°λ¦¬μ μ μ¬λ ₯μ 보μ¬μ€λ€κ³ λ―ΏμΌλ©°, μμΌλ‘ λ€κ°μ¬ κΈ°νλ€μ λν΄ μ μ λ λκ΄μ μΌλ‘ μκ°νκ³ μμ΅λλ€. μ΄μ μ ν¬ μ΅κ³ μ¬λ¬΄μ±
μμμΈ λ νλ λ°μ΄ μ¬λ¬΄ μ€μ μ λν μΆκ°μ μΈ μΈλΆμ¬νμ μ 곡νκ² μ΅λλ€. λ€λμ νλ λ° μ΅κ³ μ¬λ¬΄μ± μμ λΉ, κ°μ¬ν©λλ€. κ·Έλ¦¬κ³ μλ νμΈμ. λ¨Όμ 1λΆκΈ° μ€μ μ κ²ν ν ν νκ³μ°λ 26λ 2λΆκΈ° μ λ§μ λν΄ λ Όμνκ² μ΅λλ€. 1λΆκΈ°μ μ ν¬λ 2μ΅ 2,310λ§ λ¬λ¬μ λ§€μΆμ κΈ°λ‘νμΌλ©°, μ΄λ μ λΆκΈ° λλΉ 31% μ¦κ°, μ λ λκΈ° λλΉ 274% μ¦κ°ν μμΉλ‘ κ°μ΄λμ€ λ²μμ μνμ μ ν¬κ² μννλ μ€μ μ λλ€. μ ν¬ μ ν μ¬μ λΆλ 1λΆκΈ°μ 2μ΅ 1,710λ§ λ¬λ¬μ λ§€μΆμ μ°½μΆνμΌλ©°, μ΄λ μ λΆκΈ° λλΉ 31% μ¦κ°, μ λ λκΈ° λλΉ 279% μ¦κ°ν μμΉμ λλ€. |
Notably, our AEC product line again grew healthy double digits sequentially to achieve new record revenue levels once again. Our top 3 end customers were each greater than 10% of revenue in Q1. As a reminder, customer mix will vary from quarter-to-quarter, and we continue to make progress in diversifying our customer base. We continue to expect that three to four customers will be greater than 10% of revenue in the coming quarters and fiscal year as hyperscale customers continue to ramp to more significant volumes and as we expect to begin to ramp two new hyperscale customers in fiscal year '26. | νΉν λΉμ¬μ AEC μ ν λΌμΈμ λΆκΈ° λλΉ κ±΄μ ν λ μλ¦Ώμ μ±μ₯μ λ€μ ν λ² κΈ°λ‘νλ©° λλ€μ μ κΈ°λ‘ λ§€μΆ μμ€μ λ¬μ±νμ΅λλ€. 1λΆκΈ° μμ 3λ κ³ κ°μ κ°κ° λ§€μΆμ 10% μ΄μμ μ°¨μ§νμ΅λλ€. λ§μλλ¦° λ°μ κ°μ΄ κ³ κ° κ΅¬μ±μ λΆκΈ°λ§λ€ λ³λλ μ μμΌλ©°, λΉμ¬λ κ³ κ° κΈ°λ° λ€κ°νμμ μ§μμ μΌλ‘ μ§μ μ μ΄λ£¨κ³ μμ΅λλ€. νμ΄νΌμ€μΌμΌ κ³ κ°λ€μ΄ λ³΄λ€ μλΉν κ·λͺ¨λ‘ κ³μ λ¨νμ νκ³ μκ³ , 2026 νκ³μ°λμ λ κ°μ μλ‘μ΄ νμ΄νΌμ€μΌμΌ κ³ κ°μ λ¨νμ μ μμν κ²μΌλ‘ μμν¨μ λ°λΌ, ν₯ν λΆκΈ°λ€κ³Ό νκ³μ°λμ 3~4κ° κ³ κ°μ΄ λ§€μΆμ 10% μ΄μμ μ°¨μ§ν κ²μΌλ‘ κ³μ μμνκ³ μμ΅λλ€. |
Our team delivered Q1 non-GAAP gross margin of 67.6%, above the high end of our guidance range and up 20 basis points sequentially. Our product non-GAAP gross margin was 66.7% in the quarter, up 12 basis points sequentially and up 517 basis points year-over-year, primarily due to increasing scale. Total non-GAAP operating expenses in the first quarter were $54.5 million, at the low end of our guidance range and up 5% sequentially. Our non-GAAP operating income was $96.2 million in Q1 compared to non-GAAP operating income of $62.5 million in Q4, up demonstrably due to the leverage attained by achieving 31% sequential top line growth, while OpEx growth was in the mid-single digits. | μ ν¬ νμ 1λΆκΈ° non-GAAP μ΄μ΄μ΅λ₯ 67.6%λ₯Ό λ¬μ±νμΌλ©°, μ΄λ κ°μ΄λμ€ λ²μμ μλ¨μ μννκ³ μ λΆκΈ° λλΉ 20bp μ¦κ°ν μμΉμ λλ€. 1λΆκΈ° μ ν non-GAAP μ΄μ΄μ΅λ₯ μ 66.7%λ‘ μ λΆκΈ° λλΉ 12bp μμΉ, μ λ λκΈ° λλΉ 517bp μμΉνμΌλ©°, μ΄λ μ£Όλ‘ κ·λͺ¨ νλμ κΈ°μΈν©λλ€. 1λΆκΈ° μ΄ non-GAAP μ΄μλΉμ©μ 5,450λ§ λ¬λ¬λ‘ κ°μ΄λμ€ λ²μμ νλ¨μ μμΉνμΌλ©° μ λΆκΈ° λλΉ 5% μ¦κ°νμ΅λλ€. 1λΆκΈ° non-GAAP μμ μ΄μ΅μ 9,620λ§ λ¬λ¬λ‘ 4λΆκΈ° non-GAAP μμ μ΄μ΅ 6,250λ§ λ¬λ¬μ λΉκ΅ν΄ νμ ν μ¦κ°νμ΅λλ€. μ΄λ λ§€μΆμ΄ μ λΆκΈ° λλΉ 31% μ±μ₯ν λ°λ©΄ μ΄μλΉμ© μ¦κ°μ¨μ ν μλ¦Ώμ μ€λ°μ λ¨Έλ¬Όλ¬ λ¬μ±ν λ λ²λ¦¬μ§ ν¨κ³Όμ κΈ°μΈν©λλ€. |
Our non-GAAP operating margin was 43.1% in the quarter compared to a non-GAAP operating margin of 36.8% in the prior quarter, a sequential increase of 635 basis points. Perhaps nowhere does our performance show more clearly than at our bottom line. Our non-GAAP net income was $98.3 million in the quarter, a record high and a 51% sequential increase compared to non-GAAP net income of $65.3 million in Q4. And our non-GAAP net margin was 44.1% in the quarter as we drove significant leverage in the business. Cash flow from operations in the first quarter was $54.2 million, down $3.7 million sequentially due largely to increases in working capital. | λΉλΆκΈ° non-GAAP μμ μ΄μ΅λ₯ μ 43.1%λ‘ μ λΆκΈ° non-GAAP μμ μ΄μ΅λ₯ 36.8% λλΉ 635bp μμ¦κ°λ₯Ό κΈ°λ‘νμ΅λλ€. μ°λ¦¬μ μ±κ³Όλ μμ΄μ΅μμ κ°μ₯ λͺ ννκ² λλ¬λ©λλ€. λΉλΆκΈ° non-GAAP μμ΄μ΅μ 9,830λ§ λ¬λ¬λ‘ μ¬μ μ΅κ³ μΉλ₯Ό κΈ°λ‘νμΌλ©°, 4λΆκΈ° non-GAAP μμ΄μ΅ 6,530λ§ λ¬λ¬ λλΉ 51% μμ¦κ°νμ΅λλ€. κ·Έλ¦¬κ³ λΉλΆκΈ° non-GAAP μμ΄μ΅λ₯ μ 44.1%λ‘ μ¬μ μμ μλΉν λ λ²λ¦¬μ§λ₯Ό λ¬μ±νμ΅λλ€. 1λΆκΈ° μμ νκΈνλ¦μ 5,420λ§ λ¬λ¬λ‘ μ£Όλ‘ μ΄μ μλ³Έ μ¦κ°λ‘ μΈν΄ μ λΆκΈ° λλΉ 370λ§ λ¬λ¬ κ°μνμ΅λλ€. |
CapEx was $2.8 million in the quarter, driven largely by purchases of production equipment. And free cash flow was $51.3 million, down slightly from $54.2 million from the fourth quarter. We ended the quarter with cash and equivalents of $479.6 million, an increase of $48.3 million from the fourth quarter. We remain well capitalized to continue investing in our growth opportunities while maintaining a substantial cash buffer. Our Q1 ending inventory was $116.7 million, up $26.6 million sequentially. Now turning to our guidance. We currently expect revenue in Q2 of fiscal '26 to be between $230 million and $240 million, up 5% sequentially at the midpoint. | μ΄λ² λΆκΈ° μλ³Έμ§μΆ(CapEx)μ 280λ§ λ¬λ¬λ‘, μ£Όλ‘ μμ° μ₯λΉ κ΅¬λ§€μ μν κ²μ΄μμ΅λλ€. μμ¬νκΈνλ¦(Free Cash Flow)μ 5,130λ§ λ¬λ¬λ‘ 4λΆκΈ°μ 5,420λ§ λ¬λ¬ λλΉ μν κ°μνμ΅λλ€. λΆκΈ° λ§ νκΈ λ° νκΈμ±μμ°μ 4μ΅ 7,960λ§ λ¬λ¬λ‘ 4λΆκΈ° λλΉ 4,830λ§ λ¬λ¬ μ¦κ°νμ΅λλ€. λΉμ¬λ μλΉν νκΈ μμΆ© μκΈμ μ μ§νλ©΄μ μ±μ₯ κΈ°νμ μ§μμ μΌλ‘ ν¬μν μ μλ μΆ©λΆν μλ³Έλ ₯μ 보μ νκ³ μμ΅λλ€. 1λΆκΈ° λ§ μ¬κ³ μμ°μ 1μ΅ 1,670λ§ λ¬λ¬λ‘ μ λΆκΈ° λλΉ 2,660λ§ λ¬λ¬ μ¦κ°νμ΅λλ€. μ΄μ κ°μ΄λμ€λ‘ λμ΄κ°κ² μ΅λλ€. νμ¬ νκ³μ°λ 26λ 2λΆκΈ° λ§€μΆμ 2μ΅ 3,000λ§ λ¬λ¬μμ 2μ΅ 4,000λ§ λ¬λ¬ μ¬μ΄λ‘ μμνλ©°, μ΄λ μ€κ°κ° κΈ°μ€μΌλ‘ μ λΆκΈ° λλΉ 5% μ¦κ°ν μμ€μ λλ€. |
We expect Q2 non-GAAP gross margin to be within a range of 64% to 66%. We expect Q2 non-GAAP operating expenses to be between $56 million and $58 million. We expect Q2 diluted weighted average share count to be approximately 190 million shares. These expectations are based on the current tariff regime, which remains fluid. As we move forward throughout fiscal year '26, we expect sequential revenue growth in the mid-single digits, leading to approximately 120% year-over-year growth. We expect each of our top 3 customers from Q1 to grow significantly year-over-year in fiscal year '26. | 2λΆκΈ° non-GAAP λ§€μΆμ΄μ΄μ΅λ₯ μ 64%μμ 66% λ²μ λ΄μ μμ κ²μΌλ‘ μμν©λλ€. 2λΆκΈ° non-GAAP μμ λΉμ©μ 5,600λ§ λ¬λ¬μμ 5,800λ§ λ¬λ¬ μ¬μ΄κ° λ κ²μΌλ‘ μμν©λλ€. 2λΆκΈ° ν¬μ κ°μ€νκ· μ£Όμμλ μ½ 1μ΅ 9μ²λ§ μ£Όκ° λ κ²μΌλ‘ μμν©λλ€. μ΄λ¬ν μ λ§μ μ¬μ ν μ λμ μΈ νμ¬ κ΄μΈ 체μ λ₯Ό κΈ°λ°μΌλ‘ ν κ²μ λλ€. 26νκ³μ°λ μ λ°μ κ±Έμ³ μ€κ° ν μλ¦Ώμμ μμ°¨μ λ§€μΆ μ±μ₯μ ν΅ν΄ μ λ λλΉ μ½ 120%μ μ±μ₯μ λ¬μ±ν κ²μΌλ‘ μμν©λλ€. 1λΆκΈ° μμ 3λ κ³ κ° κ°κ°μ΄ 26νκ³μ°λμ μ λ λλΉ ν¬κ² μ±μ₯ν κ²μΌλ‘ μμν©λλ€. |
We also expect revenue diversification to strengthen further with the fourth customer surpassing the 10% revenue threshold for the year. We expect non-GAAP operating expenses to increase year-over-year by less than 50% in fiscal year '26. As a result, we expect our non-GAAP net margin to be approximately 40% in the coming quarters and for fiscal year '26. And with that, I will open it up for questions | λν λ€ λ²μ§Έ κ³ κ°μ΄ μ°κ° λ§€μΆμ 10% μκ³μ μ λμ΄μλ©΄μ λ§€μΆ λ€κ°νκ° λμ± κ°νλ κ²μΌλ‘ μμν©λλ€. 26νκ³μ°λμλ λΉGAAP μμ λΉμ©μ΄ μ λ λλΉ 50% λ―Έλ§ μ¦κ°ν κ²μΌλ‘ μμν©λλ€. κ·Έ κ²°κ³Ό, ν₯ν λΆκΈ°λ€κ³Ό 26νκ³μ°λμ λΉGAAP μμ΄μ΅λ₯ μ΄ μ½ 40%κ° λ κ²μΌλ‘ μμν©λλ€. μ΄μ μ§μμλ΅ μκ°μ κ°λλ‘ νκ² μ΅λλ€. |
β’ μ€μ νμ΄λΌμ΄νΈ:
- 1λΆκΈ° λ§€μΆ 2μ΅2,310λ§ λ¬λ¬ (μ λΆκΈ° λλΉ 31% μ¦κ°, μ λ
λλΉ 274% μ¦κ°)
- λΉGAAP λ§€μΆμ΄μ΄μ΅λ₯ 67.6%
- λΉGAAP μμ΄μ΅ 9,830λ§ λ¬λ¬ λ¬μ±
β’ AI μΈνλΌ μμ μ¦κ°λ‘ μΈν μ±μ₯:
- 3κ°μ νμ΄νΌμ€μΌμΌλ¬κ° κ°κ° λ§€μΆμ 10% μ΄μ κΈ°μ¬
- AEC(Active Electrical Cable) μ νκ΅°μ κ°λ ₯ν μ±μ₯μΈ μ§μ
- 4λ²μ§Έ νμ΄νΌμ€μΌμΌλ¬λ‘λΆν° 첫 μ€μ§μ λ§€μΆ λ°μ
β’ FY26 μ λ§:
- 2λΆκΈ° λ§€μΆ κ°μ΄λμ€: 2μ΅3,000λ§-2μ΅4,000λ§ λ¬λ¬
- FY26 μ°κ° λ§€μΆ μ±μ₯λ₯ μ½ 120% μμ
- λΉGAAP μμ
λΉμ© μ λ
λλΉ 50% λ―Έλ§ μ¦κ° μμ
- λΉGAAP μμ΄μ΅λ₯ μ½ 40% μ μ§ μ λ§
β’ μ λ΅μ κ°μ :
- SerDes κΈ°μ , IC μ€κ³, μμ€ν
λ 벨 κ°λ°μ 3λ¨κ³ νμ νλ μμν¬ λ³΄μ
- κ΄ν λ° PCIe μμ₯μΌλ‘μ μ±κ³΅μ μΈ μ¬μ
νμ₯
- κ²¬κ³ ν μ¬λ¬΄μν (νκΈ λ° νκΈμ±μμ° 4μ΅7,960λ§ λ¬λ¬ 보μ )
Original | Translation |
---|---|
Operator: [Operator Instructions] And our first question comes from the line of Karl Ackerman with BNP Paribas. | **Operator:** [μ΄μμ μλ΄] 첫 λ²μ§Έ μ§λ¬Έμ BNP ν리λ°μ μΉΌ μ μ»€λ§ λκ»μ μ£Όμ ¨μ΅λλ€. |
Karl Ackerman: BNP Paribas Exane, Research Division It sounds like you are seeing at least one of the two new hyperscalers adopting AEC this year, pull in orders relative to your previous outlook. Could you discuss whether all of these new developments are adopting -- across hyperscalers are adopting 100-gig per lane AECs and whether you have several programs across each of these hyperscalers that demonstrate the -- not only the breadth, but also the depth of your engagements? And I have a follow-up. | **Karl Ackerman:** μ¬ν΄ λ κ°μ μλ‘μ΄ νμ΄νΌμ€μΌμΌλ¬ μ€ μ΅μ ν κ³³μμ AECλ₯Ό λμ νλ©΄μ μ΄μ μ λ§ λλΉ μ£Όλ¬Έμ΄ μλΉκ²¨μ§κ³ μλ κ²μΌλ‘ 보μ λλ€. νμ΄νΌμ€μΌμΌλ¬λ€μ μ΄λ¬ν μλ‘μ΄ κ°λ° νλ‘μ νΈλ€μ΄ λͺ¨λ 100κΈ°κ° per lane AECλ₯Ό μ±ννκ³ μλμ§, κ·Έλ¦¬κ³ κ° νμ΄νΌμ€μΌμΌλ¬μμ μ¬λ¬ νλ‘κ·Έλ¨μ μ§ννκ³ μμ΄ κ·νμ μ°Έμ¬ λ²μλΏλ§ μλλΌ κΉμ΄κΉμ§ 보μ¬μ£Όλμ§μ λν΄ λ§μν΄ μ£Όμ€ μ μλμ? μΆκ° μ§λ¬Έμ΄ μμ΅λλ€. |
William Brennan: President, CEO & Chairman Yes. Thanks, Karl. I'll confirm that the most recent ramps that we've seen have been at 100 gig per lane. And depending on the customer, we've got multiple programs in flight. Naturally, as we begin a relationship with a new customer, it always starts with a single platform. And then typically, what we've seen is it expands from there. But I would say that everything we've seen over the last three months has been encouraging in a sense that we did see a pull-in from the schedule that we talked about, and we really effectively delivered to the volumes that they wanted. | **William Brennan:** λ€, κ°μ¬ν©λλ€, Karl. μ΅κ·Όμ μ ν¬κ° νμΈν λ¨νμ
μ λ μΈλΉ 100κΈ°κ°λ‘ μ§νλκ³ μλ€κ³ λ§μλ릴 μ μμ΅λλ€. κ³ κ°μ λ°λΌ λ€μμ νλ‘κ·Έλ¨μ λμμ μ§ννκ³ μμ΅λλ€. λΉμ°ν μλ‘μ΄ κ³ κ°κ³Ό κ΄κ³λ₯Ό μμν λλ νμ λ¨μΌ νλ«νΌμΌλ‘ μμνκ² λ©λλ€. κ·Έλ¦¬κ³ μΌλ°μ μΌλ‘ μ ν¬κ° κ²½νν λ°λ‘λ κ·Έ μ΄νμ νμ₯λλ ν¨ν΄μ 보μ
λλ€. μ§λ 3κ°μ λμ μ ν¬κ° λ³Έ λͺ¨λ κ²λ€μ΄ κ³ λ¬΄μ μ΄μλ€κ³ λ§μλλ¦¬κ³ μΆμ΅λλ€. μ ν¬κ° λ Όμνλ μΌμ λ³΄λ€ μλΉκ²¨μ§ μμλ₯Ό μ€μ λ‘ νμΈνκ³ , κ³ κ°λ€μ΄ μνλ λ¬Όλμ ν¨κ³Όμ μΌλ‘ 곡κΈν μ μμμ΅λλ€. |
Karl Ackerman: BNP Paribas Exane, Research Division Helpful. When you think about scale-up interconnect opportunities, several of your customers are taking different approaches with some using PCIe-based solutions and some using scale-up Ethernet solutions. Having said that, do you believe AECs offer greater competitive advantages to scale-up Ethernet or PCIe-based applications such as UALink, and therefore, the market opportunity of one could be larger than the other? | **Karl Ackerman:** μ€μΌμΌμ μΈν°μ»€λ₯νΈ κΈ°νλ₯Ό μκ°ν΄λ³Ό λ, κ³ κ°λ€μ΄ μλ‘ λ€λ₯Έ μ κ·Ό λ°©μμ μ·¨νκ³ μμ΅λλ€. μΌλΆλ PCIe κΈ°λ° μ루μ μ μ¬μ©νκ³ , μΌλΆλ μ€μΌμΌμ μ΄λλ· μ루μ μ μ¬μ©νκ³ μμ£ . κ·Έλ λ€λ©΄ AECκ° μ€μΌμΌμ μ΄λλ·μ΄λ UALink κ°μ PCIe κΈ°λ° μ ν리μΌμ΄μ μ€ μ΄λ μͺ½μ λ ν° κ²½μ μ°μλ₯Ό μ 곡νλ€κ³ 보μλμ§, κ·Έλ¦¬κ³ λ°λΌμ μ΄λ μͺ½μ μμ₯ κΈ°νκ° λ ν΄ κ²μΌλ‘ 보μλμ§ κΆκΈν©λλ€. |
William Brennan: President, CEO & Chairman Yes. I would say that the near-term opportunity for us to scale up is really with the PCIe protocol as we see the market moving from PCIe Gen 5 to Gen 6. We do see that AECs will represent a really nice opportunity, both for intra-rack as well as rack-to-rack as scale-up goes row scale. The longer-term discussion is around the debate in the industry about the 200-gig per lane protocol. And of course, we've heard from Broadcom with scale up Ethernet. UAL is really being driven by AMD and others. Of course, NVIDIA has come forward with NVLink Fusion. And PCIe is talking about going to PCIe Gen 7 and beyond. For us, we really think the opportunity is one that we're agnostic towards because the products that we're delivering can support any one of the 200-gig per lane protocols. Each one of them uses the same identical IEEE 200-gig SerDes. And so I think we're going to be in good shape. So it's -- again, it's Layer 1 from the standpoint of an AEC. And so I think that for us, we'd just like to see the market go faster sooner because the scale-up opportunity represents a significant increase in TAM really over the next two to five years. | **William Brennan:** λ€, μ ν¬κ° λ¨κΈ°μ μΌλ‘ κ·λͺ¨λ₯Ό νμ₯ν μ μλ κΈ°νλ μμ₯μ΄ PCIe Gen 5μμ Gen 6λ‘ μ΄λνλ©΄μ PCIe νλ‘ν μ½μ ν΅ν΄μλΌκ³ λ§μλ릴 μ μμ΅λλ€. AECλ€μ΄ μ€μΌμΌμ
μ΄ λ‘μ° μ€μΌμΌλ‘ νμ₯λ¨μ λ°λΌ μΈνΈλΌ λκ³Ό λ ν¬ λ λͺ¨λμμ μ λ§ μ’μ κΈ°νλ₯Ό μ 곡ν κ²μΌλ‘ λ³΄κ³ μμ΅λλ€. μ₯κΈ°μ μΈ λ Όμλ μ κ³μμ 200κΈ°κ° νΌ λ μΈ νλ‘ν μ½μ λν λ Όμμ μ€μ¬μΌλ‘ μ΄λ£¨μ΄μ§κ³ μμ΅λλ€. λ¬Όλ‘ λΈλ‘λμ»΄μμ μ€μΌμΌμ μ΄λλ·μ λν΄ λ€μκ³ , UALμ μ€μ λ‘ AMDμ λ€λ₯Έ μ 체λ€μ΄ μ£Όλνκ³ μμ΅λλ€. λΉμ°ν μλΉλμλ NVLink Fusionμ λ°ννκ³ μ. κ·Έλ¦¬κ³ PCIeλ PCIe Gen 7κ³Ό κ·Έ μ΄νλ‘ λμκ°λ κ²μ λν΄ λ Όμνκ³ μμ΅λλ€. μ ν¬μκ²λ μ΄ κΈ°νκ° μ λ§ μ€λ¦½μ μΈ μ μ₯μμ μ κ·Όν μ μλ κ²μ΄λΌκ³ μκ°ν©λλ€. μ ν¬κ° μ 곡νλ μ νλ€μ΄ 200κΈ°κ° per lane νλ‘ν μ½ μ€ μ΄λ€ κ²μ΄λ μ§μν μ μκΈ° λλ¬Έμ λλ€. κ°κ°μ νλ‘ν μ½λ€μ΄ λͺ¨λ λμΌν IEEE 200κΈ°κ° SerDesλ₯Ό μ¬μ©νκ±°λ μ. κ·Έλμ μ ν¬λ μ’μ ν¬μ§μ μ μλ€κ³ μκ°ν©λλ€. λ€μ λ§ν΄μ, AEC κ΄μ μμ 보면 Layer 1μ΄λΌλ κ²μ΄μ£ . μ ν¬λ‘μλ μμ₯μ΄ λ 빨리, λ μΌμ° μ±μ₯νκΈ°λ₯Ό λ°λ λΏμ λλ€. μ€μΌμΌμ κΈ°νκ° ν₯ν 2λ μμ 5λ λμ TAM(μ 체 μμ₯ κ·λͺ¨)μ μλΉν μ¦κ°λ₯Ό μλ―ΈνκΈ° λλ¬Έμ λλ€. |
Operator: And our next question comes from the line of Vivek Arya with Bank of America. | **Operator:** λ€μ μ§λ¬Έμ λ±
ν¬ μ€λΈ μλ©λ¦¬μΉ΄μ λΉλ²‘ μ리μλκ»μ μ£Όμ μ§λ¬Έμ
λλ€. κ·Έλ¦¬κ³ λ€μ μ§λ¬Έμ λ± ν¬ μ€λΈ μλ©λ¦¬μΉ΄μ λΉλ²‘ μ리μ μ λ리μ€νΈλ‘λΆν° λμμ΅λλ€. |
Vivek Arya: BofA Securities, Research Division First set of questions is on the AEC market. If you could quantify how large each of your 10% customers were if they were the same as you had in the prior quarter? And then Bill, if I zoom out, the market for you is now run rating closer to $1 billion or so. How large do you think this market is over time? And do you think this market is cannibalizing traditional copper? Or do you think at some point, it is even cannibalizing optical transceivers? | **Vivek Arya:** AEC μμ₯μ λν 첫 λ²μ§Έ μ§λ¬Έμ λλ€. 10% κ³ κ°λ€μ΄ μ΄μ λΆκΈ°μ λμΌνλ€λ©΄ κ°κ°μ κ·λͺ¨κ° μ΄λ μ λμλμ§ μμΉνν΄ μ£Όμ€ μ μλμ? κ·Έλ¦¬κ³ Bill, ν° κ·Έλ¦Όμμ 보면 νμ¬ μ¬λ¬λΆμ μμ₯μ΄ μ½ 10μ΅ λ¬λ¬ μ λμ λ°λ μ΄νΈ(run rate)λ‘ μ΄μλκ³ μλλ°μ. μκ°μ΄ μ§λλ©΄μ μ΄ μμ₯μ΄ μΌλ§λ μ»€μ§ κ²μΌλ‘ 보μλμ? κ·Έλ¦¬κ³ μ΄ μμ₯μ΄ κΈ°μ‘΄ ꡬ리μ μ μ μνκ³ μλ€κ³ μκ°νμλμ? μλλ©΄ μ΄λ μμ μμλ κ΄ νΈλμλ²(optical transceivers)κΉμ§λ μ μν κ²μΌλ‘ 보μλμ? |
Chief Financial Officer: Vivek, this is Dan. Let me address the 10% customer question that you had. So as we mentioned in our prepared remarks, we had three 10% customers in Q1. They were the same three customers that were 10% customers in Q4. The mix was a little bit different, though. Our largest customer was 35% of revenue. Second largest was 33% and the third largest was 20%. So we're quite pleased with that kind of the customer diversity that we demonstrated within Q1. But having said that, we expect continued diversification, as Bill highlighted, throughout fiscal '26. We do see two additional hyperscalers ramping, one of which, which was our fourth hyperscaler, we mentioned should reach to be a 10% customer for the full year of fiscal '26. And then the last thing I'll mention on customer our largest customer for fiscal '25 is the largest driver of our growth in fiscal '26 as we stand right now and look forward. So that's an important factor to bear in mind as well as you look at how our year will progress. | **Chief Financial Officer:** λΉλ²‘, λμ
λλ€. 10% κ³ κ°μ λν μ§λ¬Έμ λ΅λ³λλ¦¬κ² μ΅λλ€. μ€λΉλ λ°μΈμμ μΈκΈνλ―μ΄, 1λΆκΈ°μ 10% κ³ κ°μ΄ 3κ³³ μμμ΅λλ€. 4λΆκΈ°μ λμΌν 3κ³³μ κ³ κ°μ΄μμ£ . λ€λ§ ꡬμ±λΉλ μ‘°κΈ λ¬λμ΅λλ€. μ΅λ κ³ κ°μ΄ λ§€μΆμ 35%, λ λ²μ§Έκ° 33%, μΈ λ²μ§Έκ° 20%μμ΅λλ€. 1λΆκΈ°μ 보μ¬μ€ μ΄λ° κ³ κ° λ€μμ±μ λν΄ λ§€μ° λ§μ‘±νκ³ μμ΅λλ€. νμ§λ§ λΉμ΄ κ°μ‘°νλ―μ΄, 2026 νκ³μ°λ μ λ°μ κ±Έμ³ μ§μμ μΈ λ€κ°νλ₯Ό κΈ°λνκ³ μμ΅λλ€. 2κ³³μ μΆκ° νμ΄νΌμ€μΌμΌλ¬κ° μ¦κ°νκ³ μλ κ²μ λ³΄κ³ μμΌλ©°, κ·Έ μ€ νλμΈ λ€ λ²μ§Έ νμ΄νΌμ€μΌμΌλ¬λ 2026 νκ³μ°λ μ 체 κΈ°μ€μΌλ‘ 10% κ³ κ°μ λλ¬ν κ²μΌλ‘ μμνλ€κ³ μΈκΈνμ΅λλ€. κ·Έλ¦¬κ³ κ³ κ° κ΄λ ¨ν΄μ λ§μ§λ§μΌλ‘ λ§μλ릴 μ μ, 25νκ³μ°λ μ΅λ κ³ κ°μ΄ νμ¬ μμ μμ 26νκ³μ°λ μ±μ₯μ κ°μ₯ ν° λλ ₯μ΄ λ κ²μΌλ‘ μ λ§λλ€λ κ²μ λλ€. λ°λΌμ μ¬ν΄ μ°λ¦¬μ μ¬μ μ§ν μν©μ λ³΄μ€ λ μ΄ μ μ μΌλμ λμλ κ²μ΄ μ€μν©λλ€. |
William Brennan: President, CEO & Chairman Right. And I'll take a shot at answering your question as it relates to the overall TAM. So with all of our customers, we've got additional opportunities. We're not seeing AECs being adopted across all the possible applications. And we've talked about this before that historically, we've talked about it being an intra-rack kind of de facto solution as we replace passive copper cables. And the catalyst being speed and also functionality. And so we've seen a good transition as the market has gone to 50 gig per lane, and now we're seeing a really good traction as it relates to 100 gig per lane. I still think we're in the early innings from a market sizing standpoint, even for just breaking out the intra-rack. Again, to reiterate, we see a huge opportunity from NIC to TOR applications as well as switch rack applications. And that is for front end and really emphasized much more strongly in back end, both in scale out and scale up. What we've seen recently that we talked about in the prepared comments was that we're seeing a move from intra-rack to rack-to-rack. And traditionally, intra-rack has been passive copper cables or DACs. And we've seen rack-to-rack being more optical solutions because of the distance. But I think with the emphasis on reliability as it relates to clusters that we see customers really considering using AECs. It's really driven by reliability. I mentioned on the prepared comments that we're up to 1,000x more reliable, effectively reducing length laps and having the uptime of the cluster being much more. Again, reiterating that if you have got a single link flap in, say, 10,000 or 100,000 or 1 million GPU cluster, it brings the entire cluster down because there's no redundancy from that NIC to tour connection. And so we're actually seeing the TAM expanding. And I think for the first time in history that you're seeing copper replacing optical connections. So we're quite bullish on the market generally. | **William Brennan:** μ 체 TAMκ³Ό κ΄λ ¨λ μ§λ¬Έμ λ΅λ³λλ¦¬κ² μ΅λλ€. λͺ¨λ κ³ κ°μ¬μμ μΆκ° κΈ°νλ€μ΄ μμ΅λλ€. AECκ° κ°λ₯ν λͺ¨λ μ ν리μΌμ΄μ
μμ μ±νλκ³ μμ§λ μκ±°λ μ. μ΄μ μλ λ§μλλ Έμ§λ§, μμ¬μ μΌλ‘ ν¨μλΈ κ΅¬λ¦¬ μΌμ΄λΈμ λ체νλ©΄μ μΈνΈλΌ λ(intra-rack) λΆμΌμ μ¬μ€μ νμ€ μ루μ
μΌλ‘ μ리μ‘κ³ μλ€κ³ λ΄
λλ€. μ¬κΈ°μ μ΄λ§€ μν μ νλ κ²μ΄ μλμ κΈ°λ₯μ±μ΄μ£ . μμ₯μ΄ λ μΈλΉ 50κΈ°κ°λ‘ μ νλλ©΄μ μ’μ μ±κ³Όλ₯Ό λ΄€κ³ , μ§κΈμ λ μΈλΉ 100κΈ°κ° κ΄λ ¨ν΄μ μ λ§ μ’μ 견μΈλ ₯μ 보μ΄κ³ μμ΅λλ€. μμ₯ κ·λͺ¨ κ΄μ μμ 보면, μΈνΈλΌ λ λΆμΌλ§ λκ³ λ΄λ μμ§ μ΄κΈ° λ¨κ³λΌκ³ μκ°ν©λλ€. λ€μ νλ² κ°μ‘°νμλ©΄, NICμμ TOR μ ν리μΌμ΄μ μΌλ‘, κ·Έλ¦¬κ³ μ€μμΉ λ μ ν리μΌμ΄μ μΌλ‘μ μ νμμ μμ²λ κΈ°νλ₯Ό λ³΄κ³ μμ΅λλ€. μ΄λ νλ‘ νΈμλμ λ°±μλ λͺ¨λμ ν΄λΉλλ©°, νΉν λ°±μλμμλ μ€μΌμΌ μμκ³Ό μ€μΌμΌ μ λͺ¨λμμ ν¨μ¬ λ κ°μ‘°λκ³ μμ΅λλ€. μ€λΉλ λ°μΈμμ μΈκΈνλ―μ΄, μ΅κ·Ό μ°λ¦¬κ° λͺ©κ²©νκ³ μλ κ²μ μΈνΈλΌ λ(intra-rack)μμ λ ν¬ λ(rack-to-rack)μΌλ‘μ μ΄λμ λλ€. μ ν΅μ μΌλ‘ μΈνΈλΌ λμ ν¨μλΈ κ΅¬λ¦¬ μΌμ΄λΈμ΄λ DACλ₯Ό μ¬μ©ν΄μμ΅λλ€. κ·Έλ¦¬κ³ λ ν¬ λμ κ²½μ° κ±°λ¦¬ λλ¬Έμ λ λ§μ κ΄ν μ루μ μ μ¬μ©νλ κ²μ λ΄μμ΅λλ€. νμ§λ§ ν΄λ¬μ€ν°μ κ΄λ ¨λ μμ μ±μ λν κ°μ‘°λ‘ μΈν΄ κ³ κ°λ€μ΄ μ€μ λ‘ AEC μ¬μ©μ κ³ λ €νκ³ μλ κ²μ λ³΄κ³ μμ΅λλ€. μ΄λ μ λ§λ‘ μμ μ±μ μν΄ μ£Όλλκ³ μμ΅λλ€. μ€λΉλ λ°νμμ λ§μλλ¦° λ°μ κ°μ΄, μ ν¬λ μ΅λ 1,000λ°° λ λμ μ λ’°μ±μ λ¬μ±νμΌλ©°, μ΄λ₯Ό ν΅ν΄ μ°κ²° μ€λ¨μ ν¨κ³Όμ μΌλ‘ μ€μ΄κ³ ν΄λ¬μ€ν°μ κ°λ μκ°μ ν¬κ² ν₯μμμΌ°μ΅λλ€. λ€μ ν λ² κ°μ‘°νμλ©΄, μλ₯Ό λ€μ΄ 1λ§ κ°, 10λ§ κ°, λλ 100λ§ κ°μ GPU ν΄λ¬μ€ν°μμ λ¨μΌ λ§ν¬ μ₯μ κ° λ°μνλ©΄, ν΄λΉ NICμμ μ°κ²°κΉμ§μ μ€λ³΅μ±μ΄ μκΈ° λλ¬Έμ μ 체 ν΄λ¬μ€ν°κ° λ€μ΄λ©λλ€. κ·Έλμ μ€μ λ‘ TAM(μ 체 μμ₯ κ·λͺ¨)μ΄ νλλκ³ μλ κ²μ λͺ©κ²©νκ³ μμ΅λλ€. κ·Έλ¦¬κ³ μμ¬μ μ²μμΌλ‘ ꡬ리μ μ΄ κ΄ν μ°κ²°μ λ체νλ νμμ λ³΄κ³ μλ€κ³ μκ°ν©λλ€. λ°λΌμ μ ν¬λ μ λ°μ μΌλ‘ μμ₯μ λν΄ λ§€μ° λκ΄μ μ λλ€. |
Vivek Arya: BofA Securities, Research Division Got it. And for my follow-up, Bill, I actually have a question for both you and Dan, which is -- what is the next big thing for Credo? Is it optical DSP? If yes, how large can it be for you? And I ask that because it's great to see that you have this tremendous operating leverage. You're guiding every quarter, sales are up $15 million, $20 million and OpEx hardly goes up. So are you investing enough in diversifying beyond AECs and right? So just what is the next big thing for Credo? And are you investing adequately in that? | **Vivek Arya:** μκ² μ΅λλ€. νμ μ§λ¬ΈμΌλ‘ λΉκ³Ό λ λ λΆ λͺ¨λμκ² λλ¦¬κ³ μΆμ μ§λ¬Έμ΄ μλλ°μ. ν¬λ λμ λ€μ ν° μ±μ₯ λλ ₯μ 무μμΈκ°μ? κ΄ν DSPμΈκ°μ? λ§μ½ κ·Έλ λ€λ©΄, ν¬λ λμκ² μΌλ§λ ν° κΈ°νκ° λ μ μμκΉμ? μ΄λ° μ§λ¬Έμ λ리λ μ΄μ λ νμ¬ λ³΄μ¬μ£Όκ³ κ³μ μμ²λ μ΄μ λ λ²λ¦¬μ§κ° μΈμμ μ΄κΈ° λλ¬Έμ λλ€. λ§€ λΆκΈ° κ°μ΄λμ€μμ λ§€μΆμ 1,500λ§ λ¬λ¬, 2,000λ§ λ¬λ¬μ© μ¦κ°νλλ° μ΄μλΉμ©μ κ±°μ λμ§ μκ³ μκ±°λ μ. κ·Έλμ AECλ₯Ό λμ΄μ λ€κ°νμ μΆ©λΆν ν¬μνκ³ κ³μ μ§ κΆκΈν©λλ€. ν¬λ λμ λ€μ ν° μ±μ₯ λλ ₯μ΄ λ¬΄μμ΄κ³ , κ·Έ λΆμΌμ μ μ ν ν¬μνκ³ κ³μ κ°μ? |
William Brennan: President, CEO & Chairman Yes. So I think that we've talked about this in the past, and I'll speak again about the opportunity generally that we're seeing that's being driven by AI. Really, for our company, it's all about connectivity bottlenecks, and we see bottlenecks everywhere. We see that at a system level, the opportunity is abundant because really what's being driven in the market is the need for better reliability, better performance and better power efficiency. And so there's a lot to come as far as the AEC market is concerned, but we also see really big opportunities as we move from an Ethernet protocol to additional protocols in the scale-up network. We believe that there's system-level opportunities in optical. And we're not going to go into too many details here, but you'll see more in the upcoming months as we announce new products. And then even bottlenecks associated with the connection between GPUs and memory. We see opportunities there that can be quite large. Traditionally, we've seen that training has really been driving the market, but we see a really big upside in inference as well. And so again, we see multiple opportunities that will ultimately drive several pillars that we talk about as we maybe fast forward to two to three years from now and we look backwards. | **William Brennan:** λ€, μ΄μ μλ μ΄μ λν΄ λ§μλλ¦° μ μ΄ μλλ°, AIλ‘ μΈν΄ λνλκ³ μλ κΈ°νμ λν΄ μ λ°μ μΌλ‘ λ€μ λ§μλλ¦¬κ² μ΅λλ€. μ ν¬ νμ¬μκ²λ κ²°κ΅ μ°κ²°μ± λ³λͺ©νμμ΄ ν΅μ¬μ΄λ©°, μ ν¬λ λͺ¨λ κ³³μμ λ³λͺ©νμμ λ³΄κ³ μμ΅λλ€. μμ€ν
μ°¨μμμ 보면 κΈ°νκ° νλΆνλ°, μ΄λ μμ₯μμ λ λμ μ λ’°μ±, λ λμ μ±λ₯, κ·Έλ¦¬κ³ λ λμ μ λ ₯ ν¨μ¨μ±μ λν μκ΅¬κ° μ¦κ°νκ³ μκΈ° λλ¬Έμ
λλ€. AEC μμ₯κ³Ό κ΄λ ¨ν΄μλ μμΌλ‘ λ§μ κ²λ€μ΄ κΈ°λλμ§λ§, μ€μΌμΌμ λ€νΈμν¬μμ μ΄λλ· νλ‘ν μ½μμ μΆκ° νλ‘ν μ½λ‘ μ΄λνλ©΄μ μ λ§ ν° κΈ°νλ€λ λ³΄κ³ μμ΅λλ€. κ΄ν λΆμΌμμλ μμ€ν μ°¨μμ κΈ°νκ° μλ€κ³ λ―Ώκ³ μμ΅λλ€. μ¬κΈ°μ λ무 μμΈν λ΄μ©κΉμ§λ λ§μλλ¦¬μ§ μκ² μ§λ§, μμΌλ‘ λͺ λ¬ μμ μλ‘μ΄ μ νλ€μ λ°ννλ©΄μ λ λ§μ κ²λ€μ λ³΄μ€ μ μμ κ²λλ€. κ·Έλ¦¬κ³ GPUμ λ©λͺ¨λ¦¬ κ°μ μ°κ²°κ³Ό κ΄λ ¨λ λ³λͺ© νμλ€λ μμ΅λλ€. μ΄ λΆλΆμμ μλΉν ν° κΈ°νλ€μ λ³΄κ³ μμ΅λλ€. μ ν΅μ μΌλ‘λ νΈλ μ΄λμ΄ μμ₯μ μ£Όλν΄μμ§λ§, μΆλ‘ (inference) λΆμΌμμλ μ λ§ ν° μμΉ μ¬λ ₯μ λ³΄κ³ μμ΅λλ€. λ°λΌμ λ€μ λ§μλ리μ§λ§, 2-3λ νλ₯Ό λ΄λ€λ³΄κ³ λ€μ λμλ΄€μ λ μ°λ¦¬κ° μ΄μΌκΈ°νλ μ¬λ¬ ν΅μ¬ μΆλ€μ κΆκ·Ήμ μΌλ‘ 견μΈν λ€μν κΈ°νλ€μ λ³΄κ³ μμ΅λλ€. |
Chief Financial Officer: The one thing I'll add to that, Vivek, just from are we investing enough? We are continuing to manage our operating expenses to support these upcoming waves of growth that Bill's talked about. But we're at this point in time right now where our top line is growing at quite a fast clip. I will mention from just from a pure R&D standpoint this year, more than half of our OpEx in R&D is tied to optical projects and things that are in these forthcoming pillars of our business. | **Chief Financial Officer:** λΉλ²‘, ν κ°μ§ λ§λΆμ΄μλ©΄, μ°λ¦¬κ° μΆ©λΆν ν¬μνκ³ μλκ°μ λν΄μλ λΉμ΄ μΈκΈν λ€κ°μ€λ μ±μ₯μ λ¬Όκ²°λ€μ λ·λ°μΉ¨νκΈ° μν΄ μ΄μλΉμ©μ μ§μμ μΌλ‘ κ΄λ¦¬νκ³ μμ΅λλ€. νμ§λ§ νμ¬ μμ μμ μ°λ¦¬μ λ§€μΆμ΄ μλΉν λΉ λ₯Έ μλλ‘ μ±μ₯νκ³ μλ μν©μ λλ€. μμν R&D κ΄μ μμ λ§μλ리면, μ¬ν΄ μ°λ¦¬ R&D μ΄μλΉμ©μ μ λ° μ΄μμ΄ κ΄ν νλ‘μ νΈλ€κ³Ό μ°λ¦¬ μ¬μ μ ν₯ν ν΅μ¬ μΆμ΄ λ λΆμΌλ€κ³Ό μ°κ²°λμ΄ μμ΅λλ€. |
Operator: And our next question comes from the line of Tom O'Malley with Barclays. | **Operator:** λ€μ μ§λ¬Έμ λ°ν΄λ μ΄μ¦μ ν° μ€λ§λ¦¬λκ»μ μ£Όμ
¨μ΅λλ€. κ·Έλ¦¬κ³ λ€μ μ§λ¬Έμ λ°ν΄λ μ΄μ¦μ ν° μ€λ§λ¦¬λ‘λΆν° μμ΅λλ€. |
Thomas O'Malley: Barclays Bank PLC, Research Division I wanted to follow up on the optical DSP business. We're getting on the brink of a more material 1.6T transition. Could you talk about what your feelings are about your competitiveness at that node? And maybe give us a little bit of insight into the 3-nanometer platform, when you'll be in volume and how you think you're positioned versus the 800G generation? | **Thomas O'Malley:** κ΄ν DSP μ¬μ μ λν΄ νμ μ§λ¬Έμ λλ¦¬κ³ μΆμ΅λλ€. μ΄μ λμ± λ³Έκ²©μ μΈ 1.6T μ νμ λ¬Έν±μ μ μλλ°μ. ν΄λΉ λ Έλμμμ κ²½μλ ₯μ λν΄ μ΄λ»κ² μκ°νμλμ§ λ§μν΄ μ£Όμκ² μ΅λκΉ? κ·Έλ¦¬κ³ 3λλ Έλ―Έν° νλ«νΌμ λν΄μλ μ‘°κΈ λ μμΈν μλ €μ£Όμλ©΄ μ’κ² μ΅λλ€. μΈμ μ―€ μμ°μ λ€μ΄κ° μμ μ΄κ³ , 800G μΈλ λλΉ μ΄λ€ ν¬μ§μ μ μλ€κ³ 보μλμ§μ? |
William Brennan: President, CEO & Chairman Yes. Regarding our 200-gig per lane solution for the optical DSP, we'll have both a full DSP as well as LRO, the linear receive opticals. And so we moved to 3-nanometer straightaway due to power. We felt like the initial transceivers that were introduced to the market being well over 20 watts, something that was going to take off in high volume. And so we researched pretty extensively in 5-nanometer, and we've made the decision to really jump forward to 3, which has been -- it's a different strategy than we've pursued in the past, and it's really been driven by the need for power efficiency. So we feel like we're in great shape from the standpoint of the product that we're going to bring to production in the upcoming months. We think that the market for 1.6T transceivers is going to take time to develop. It's quite natural that every generation that we see, it takes a little bit more time than some of the forecasters talk about. But we feel great about the product that we've developed and the fact that we're going to bring it to production shortly. | **William Brennan:** λ€, κ΄ν DSPμ© 200κΈ°κ° λ μΈλΉ μ루μ
κ³Ό κ΄λ ¨ν΄μ λ§μλ리면, ν DSPμ LRO, μ¦ μ ν μμ κ΄ν μ루μ
μ λͺ¨λ μ 곡ν μμ μ
λλ€. κ·Έλ¦¬κ³ μ λ ₯ ν¨μ¨μ± λλ¬Έμ 3λλ
Έλ―Έν°λ‘ λ°λ‘ μ΄λνκΈ°λ‘ νμ΅λλ€. μμ₯μ μ²μ μΆμλ νΈλμλ²λ€μ΄ 20μνΈλ₯Ό ν¨μ¬ λμ΄μλ μμ€μ΄μλλ°, μ΄λ° μ νμ΄ λλμΌλ‘ μ±νλκΈ°λ μ΄λ €μΈ κ²μ΄λΌκ³ νλ¨νμ΅λλ€. κ·Έλμ 5λλ Έλ―Έν°μ λν΄ μλΉν κ΄λ²μνκ² μ°κ΅¬νμ§λ§, κ²°κ΅ 3λλ Έλ―Έν°λ‘ λ°λ‘ λμ½νκΈ°λ‘ κ²°μ νμ΅λλ€. μ΄λ κ³Όκ±°μ μ°λ¦¬κ° μΆμ§νλ μ λ΅κ³Όλ λ€λ₯Έ μ κ·Όλ²μ΄λ©°, μ λ ₯ ν¨μ¨μ±μ λν νμμ μν΄ μΆμ§λ κ²μ λλ€. μμΌλ‘ λͺ λ¬ λ΄μ μμ°μ λ€μ΄κ° μ ν κ΄μ μμ 보면 μ°λ¦¬κ° λ§€μ° μ’μ μμΉμ μλ€κ³ μκ°ν©λλ€. 1.6T νΈλμλ² μμ₯μ λ°μ νλ λ° μκ°μ΄ 걸릴 κ²μΌλ‘ λ³΄κ³ μμ΅λλ€. μ°λ¦¬κ° μ§μΌλ³Έ λͺ¨λ μΈλμμ μΌλΆ μμΈ‘κ°λ€μ΄ λ§νλ κ²λ³΄λ€ μ‘°κΈ λ μκ°μ΄ 걸리λ κ²μ λ§€μ° μμ°μ€λ¬μ΄ νμμ λλ€. νμ§λ§ μ°λ¦¬κ° κ°λ°ν μ νμ λν΄μλ λ§€μ° λ§μ‘±νκ³ μμΌλ©°, κ³§ μμ°μ λ€μ΄κ°λ€λ μ μμ ν° μμ κ°μ κ°μ§κ³ μμ΅λλ€. |
Thomas O'Malley: Barclays Bank PLC, Research Division Helpful. And then you spent a lot of time in the preamble talking about PCIe retimers. Today, that's the large GPU manufacturer and then also one large hyperscaler who's really started off their efforts there. When you talk about the opportunity in the future, are you thinking that, that is concentrated amongst those two customers? Or do you see PCIe proliferation amongst other hyperscalers as well? | **Thomas O'Malley:** λμμ΄ λμμ΅λλ€. κ·Έλ¦¬κ³ μλμμ PCIe 리νμ΄λ¨Έ(retimer)μ λν΄ λ§μ μκ°μ ν μ ν΄μ λ§μν΄ μ£Όμ ¨λλ°μ. νμ¬λ λν GPU μ μ‘°μ 체 ν κ³³κ³Ό κ·Έκ³³μμ 본격μ μΌλ‘ λ Έλ ₯μ μμν λν νμ΄νΌμ€μΌμΌλ¬ ν κ³³μ΄ μμ΅λλ€. ν₯ν κΈ°νμ λν΄ λ§μνμ€ λ, μ΄ λ κ³ κ°μ¬μ μ§μ€λ κ²μΌλ‘ λ³΄κ³ κ³μ 건κ°μ? μλλ©΄ λ€λ₯Έ νμ΄νΌμ€μΌμΌλ¬λ€ μ¬μ΄μμλ PCIe νμ°μ λ³΄κ³ κ³μ 건κ°μ? |
William Brennan: President, CEO & Chairman Yes, we definitely see it being broader than the two that you mentioned. We think that, that is going to be the protocol that most people pursue, most hyperscalers pursue in the near term. So we think the market opportunity is significant. And as things develop over time, I think that's when you're going to see -- as the market moves from PCIe Gen 6 into whatever is next, that's when you're going to see different hyperscalers taking an approach that's pursuing different protocols. But again, for us, it's really something that we're going to be able to serve regardless of the protocol based on the fact that, again, it's the same 200-gig per lane SerDes for every single one of these protocols. | **William Brennan:** λ€, λ§μνμ λ κ°μ§λ³΄λ€ ν¨μ¬ λ κ΄λ²μν κ²μΌλ‘ νμ€ν λ³΄κ³ μμ΅λλ€. λλΆλΆμ μ¬λλ€, νΉν λλΆλΆμ νμ΄νΌμ€μΌμΌλ¬λ€μ΄ λ¨κΈ°μ μΌλ‘ μΆκ΅¬ν νλ‘ν μ½μ΄ λ κ²μ΄λΌκ³ μκ°ν©λλ€. λ°λΌμ μμ₯ κΈ°νκ° μλΉν κ²μΌλ‘ λ΄ λλ€. κ·Έλ¦¬κ³ μκ°μ΄ μ§λλ©΄μ μν©μ΄ λ°μ νλ©΄μ - μμ₯μ΄ PCIe Gen 6μμ κ·Έ λ€μ μΈλλ‘ μ΄λν λ, κ·Έλ λ€μν νμ΄νΌμ€μΌμΌλ¬λ€μ΄ μλ‘ λ€λ₯Έ νλ‘ν μ½μ μΆκ΅¬νλ μ κ·Ό λ°©μμ μ·¨νλ κ²μ λ³΄κ² λ κ²μ λλ€. νμ§λ§ λ€μ λ§μλ리면, μ ν¬μκ²λ νλ‘ν μ½μ κ΄κ³μμ΄ μλΉμ€λ₯Ό μ 곡ν μ μλ κ²μ λλ€. μ΄ λͺ¨λ νλ‘ν μ½μ λμΌν λ μΈλΉ 200κΈ°κ° SerDesκ° μ¬μ©λκΈ° λλ¬Έμ λλ€. |
Operator: And our next question comes from the line of Joshua Buchalter with TD Cowen. | **Operator:** λ€μ μ§λ¬Έμ TD Cowenμ Joshua Buchalterλ‘λΆν° λ°μμ΅λλ€. |
Joshua Buchalter: TD Cowen, Research Division On the results. Maybe a follow-up to Tom's first one there. I think you've demonstrated an LRO at 800 gig with your WV850 platform. But it sounds like a lot of the industry research and maybe you guys think 1.6T is more of the catalyst for LRO. Maybe you could speak to the near to medium term, like how we should expect the adoption curve to ramp for the LRO solutions. | **Joshua Buchalter:** κ²°κ³Όμ λν΄μμ. ν°μ 첫 λ²μ§Έ μ§λ¬Έμ λν νμ μ§λ¬Έμ λλ€. WV850 νλ«νΌμΌλ‘ 800κΈ°κ°μμ LROλ₯Ό μμ°νμ κ²μΌλ‘ μκ³ μμ΅λλ€. νμ§λ§ μ κ³ μ°κ΅¬μ μλ§λ κ·νλ€λ 1.6Tκ° LROμ λ ν° μ΄λ§€μ κ° λ κ²μ΄λΌκ³ μκ°νμλ κ² κ°μ΅λλ€. λ¨κΈ°μμ μ€κΈ°μ μΌλ‘ LRO μ루μ μ λμ 곑μ μ΄ μ΄λ»κ² μ¦κ°ν κ²μΌλ‘ μμν΄μΌ νλμ§ λ§μν΄ μ£Όμ€ μ μλμ? |
William Brennan: President, CEO & Chairman Yes. We definitely see that every hyperscaler is kind of an independent market. The industry doesn't move one direction, and that's been for multiple generations now. So we've seen a lot of interest with 800-gig LRO, and I think there's good things coming. For 1.6T, again, power is becoming more and more important. And we've proven that we're able to deliver just rock-solid transceivers with our partners from a bit error rate standpoint, maintaining everything that's needed from an industry standard perspective as compared to LPO and offering significant power savings. So we think that we'll see good success with 800, but we do feel like 1.6T, it will become more and more popular given the need for power efficiency. | **William Brennan:** λ€, νμ€ν λͺ¨λ νμ΄νΌμ€μΌμΌλ¬κ° λ
립μ μΈ μμ₯μ΄λΌκ³ λ³Ό μ μμ΅λλ€. μ
κ³κ° ν λ°©ν₯μΌλ‘λ§ μμ§μ΄μ§ μλλ€λ κ²μ μ΄λ―Έ μ¬λ¬ μΈλμ κ±Έμ³ νμΈλ μ¬μ€μ
λλ€. 800κΈ°κ° LROμ λν λ§μ κ΄μ¬μ λ³΄κ³ μμΌλ©°, μ’μ κ²°κ³Όκ° λμ¬ κ²μΌλ‘ μκ°ν©λλ€. 1.6Tμ κ²½μ°, μ λ ₯ ν¨μ¨μ±μ΄ μ μ λ μ€μν΄μ§κ³ μμ΅λλ€. μ ν¬λ ννΈλλ€κ³Ό ν¨κ» λΉνΈ μ€λ₯μ¨ κ΄μ μμ λ§€μ° μμ μ μΈ νΈλμλ²λ₯Ό μ 곡ν μ μλ€λ κ²μ μ μ¦νμ΅λλ€. μ κ³ νμ€ κ΄μ μμ νμν λͺ¨λ μꡬμ¬νμ μ μ§νλ©΄μλ LPO λλΉ μλΉν μ λ ₯ μ μ½μ μ 곡νκ³ μμ΅λλ€. λ°λΌμ 800κΈ°κ°μμ μ’μ μ±κ³Όλ₯Ό κ±°λ κ²μΌλ‘ 보μ§λ§, μ λ ₯ ν¨μ¨μ±μ λν νμκ° μ»€μ§λ©΄μ 1.6Tκ° μ μ λ μΈκΈ°λ₯Ό μ»κ² λ κ²μ΄λΌκ³ μκ°ν©λλ€. |
Joshua Buchalter: TD Cowen, Research Division Got it. I just had a question. In the prepared remarks, you mentioned I think something about your customers not necessarily ramping linearly. Your guidance is obviously very strong and somewhat linear off of the strong 1Q. Maybe you could elaborate on those comments and bigger picture, how you feel about the visibility into the ramps as your expanding customer base scales? | **Joshua Buchalter:** μ€λΉλ λ°μΈμμ κ³ κ°λ€μ΄ λ°λμ μ νμ μΌλ‘ λ¨νμ νμ§ μλλ€κ³ λ§μνμ κ² κ°μλ°μ. κ°μ΄λμ€λ νμ€ν λ§€μ° κ°λ ₯νκ³ κ°ν 1λΆκΈ°λ₯Ό κΈ°λ°μΌλ‘ μ΄λ μ λ μ νμ μ λλ€. κ·Έ λ°μΈμ λν΄ μ’ λ μμΈν μ€λͺ ν΄ μ£Όμ€ μ μμκΉμ? κ·Έλ¦¬κ³ λ ν° κ΄μ μμ, νμ₯λλ κ³ κ° κΈ°λ°μ΄ κ·λͺ¨λ₯Ό ν€μλκ°λ©΄μ λ¨νμ μ λν κ°μμ±μ μ΄λ»κ² λ³΄κ³ κ³μ μ§μ? |
William Brennan: President, CEO & Chairman Yes. We've been pretty consistent in talking about -- at a given customer, we don't necessarily see a linear up and to the right ramp. That's just not how things work as customers prepare for next-generation ramps. Typically, we see a surge as they are preparing to start a new generation of deployment. And then we see it balancing out and then ultimately, it trends upward over time until we get to the next product transition. And so nothing new here. It's just if we start tracking quarter-to-quarter and at one point, there was a lot of concern about us having too much concentration. And in the next quarter, we saw a really good balance between multiple customers and some people were concerned about, hey, it's kind of a decrease from the concentration you had. And long term, I can just tell you that we've been consistent in the messaging that we're going to see diversification across customers, and we're going to see diversification across mediums, copper and optical. We're going to see diversification across protocols as well. So I think we're right on track with showing better stability as it relates to having multiple customers in flight. And so we feel good about the fact that we've kind of moved forward from a few quarters ago where there was a concern. | **William Brennan:** λ€, μ ν¬κ° μΌκ΄λκ² λ§μλλ €μ¨ λ΄μ©μΈλ°μ - νΉμ κ³ κ°μμ λ°λμ μ νμ μΌλ‘ μ°μν₯νλ μ¦κ°μΈλ₯Ό 보λ κ²μ μλλλ€. κ³ κ°λ€μ΄ μ°¨μΈλ μ¦μ€μ μ€λΉν λλ κ·Έλ° μμΌλ‘ μ§νλμ§ μκ±°λ μ. μΌλ°μ μΌλ‘ μλ‘μ΄ μΈλμ λ°°μΉλ₯Ό μμν μ€λΉλ₯Ό ν λ κΈμ¦νλ λͺ¨μ΅μ λ³΄κ² λ©λλ€. κ·Έ λ€μμλ κ· νμ λ§μΆ°κ°λ€κ° κΆκ·Ήμ μΌλ‘λ λ€μ μ ν μ νκΈ°κΉμ§ μκ°μ΄ μ§λλ©΄μ μμΉ μΆμΈλ₯Ό 보μ΄κ² λ©λλ€. μ¬κΈ°μ μλ‘μ΄ κ²μ μμ΅λλ€. λ€λ§ λΆκΈ°λ³λ‘ μΆμ νκΈ° μμνλ©΄μ, νλλ μ ν¬κ° λ무 μ§μ€λκ° λλ€λ μ°λ €κ° λ§μμ΅λλ€. κ·Έλ°λ° λ€μ λΆκΈ°μλ μ¬λ¬ κ³ κ°λ€ κ°μ μ λ§ μ’μ κ· νμ 보μκ³ , μΌλΆμμλ μ΄μ μ κ°μ‘λ μ§μ€λμμ κ°μν κ² μλλλ μ°λ €λ₯Ό ννκΈ°λ νμ΅λλ€. μ₯κΈ°μ μΌλ‘ λ§μλ리면, μ ν¬λ κ³ κ° λ€λ³νλ₯Ό μ΄λ£¨κ³ , ꡬ리μ κ³Ό κ΄ν λ± λ§€μ²΄ μ λ°μ κ±ΈμΉ λ€λ³νλ₯Ό μ€νν κ²μ΄λΌλ λ©μμ§λ₯Ό μΌκ΄λκ² μ λ¬ν΄μμ΅λλ€. νλ‘ν μ½ μΈ‘λ©΄μμλ λ€λ³νλ₯Ό μ΄λ£° μμ μ λλ€. λ°λΌμ μ¬λ¬ κ³ κ°λ€κ³Ό λμμ μ§νλλ νλ‘μ νΈλ₯Ό ν΅ν΄ λ λμ μμ μ±μ 보μ¬μ£Όλ κΆ€λμ μ¬λΌμλ€κ³ μκ°ν©λλ€. λͺ λΆκΈ° μ μ°λ €κ° μμλ μν©μμ λ²μ΄λ μμΌλ‘ λμκ°λ€λ μ μμ κΈμ μ μΌλ‘ νκ°νκ³ μμ΅λλ€. |
Operator: And our next question comes from the line of Tore Svanberg with Stifel. | **Operator:** λ€μ μ§λ¬Έμ Stifelμ Tore Svanbergλκ»μ μ£Όμ
¨μ΅λλ€. λ€, μλ νμΈμ. λͺ κ°μ§ μ§λ¬Έμ΄ μμ΅λλ€. λ¨Όμ μλμ°¨ λΆλ¬Έμ λν΄ λ§μλ리μλ©΄, μ΄λ² λΆκΈ° μ€μ μ΄ μ λ§ μΈμμ μ΄μμ΅λλ€. νΉν μ κΈ°μ°¨ κ΄λ ¨ λ§€μΆ μ±μ₯μ΄ λμ λμλλ°μ. ν₯ν λͺ λΆκΈ° λμ μ΄λ° μ±μ₯μΈκ° μ§μλ κ²μΌλ‘ 보μλμ§, κ·Έλ¦¬κ³ μλμ°¨ μμ₯μμμ μ λ°μ μΈ μ λ§μ μ΄λ»κ² λ³΄κ³ κ³μ μ§ κΆκΈν©λλ€. |
Tore Svanberg: Stifel, Nicolaus & Company, Incorporated, Research Division Congratulations on the record results. Bill, I was hoping to get just a little bit more color on the use cases for the AEC business. I think Credo has a pretty unique position, especially in row scale. So just hoping to get a little bit more color on the size or where we are with row scale networking. I assume it's still early days, but any color you could share with us would be great. | **Tore Svanberg:** κΈ°λ‘μ μΈ μ€μ μ μΆνλ립λλ€. Bill, AEC μ¬μ μ μ¬μ© μ¬λ‘μ λν΄ μ’ λ μμΈν μ€λͺ μ λ€μ μ μμκΉμ? Credoκ° νΉν λ μ€μΌμΌ(rack scale) λΆμΌμμ μλΉν λ νΉν μμΉμ μλ€κ³ μκ°νλλ°μ. κ·Έλμ λ μ€μΌμΌ λ€νΈμνΉμ κ·λͺ¨λ νμ¬ μν©μ λν΄ μ’ λ μμΈν μκ³ μΆμ΅λλ€. μμ§ μ΄κΈ° λ¨κ³λΌκ³ μκ°νμ§λ§, 곡μ ν΄ μ£Όμ€ μ μλ λ΄μ©μ΄ μλ€λ©΄ μ’κ² μ΅λλ€. |
William Brennan: President, CEO & Chairman Yes. So I think that as we talk about the market opportunity, the big shift, I think, in the TAM has been with the recognition that with the densification of the rows, that the opportunity for AEC is expanding. And the drivers, again, are really clear. It's reliability and it's power efficiency. And third is probably system cost, and that's opposed to optical solutions. And so I could make an argument that the intra-rack market is really still in the early stages, although now you've seen the solution really take off with four of the hyperscalers. We've talked about how we fully expect that all of the hyperscalers will be adopting AECs at some point in the future. And so when we talk about intra-rack, we really talk about server racks, whether those are AI appliances or whether they're general compute. And so NIC to TOR is a big market. And then we've also talked about switch racks. And in the market, there's a trend towards disaggregation of chassis. That's been a trend that's been in place for many years. We've seen that part of the market not take off as fast as NIC to TOR, and so it represents a big growth opportunity. When we talk about rack-to-rack, this is really the expanding part of the TAM really in both markets from the standpoint of AI back-end networks. as well as switch racks. Switch racks are starting to go to multiple rack architectures. And so as we talk about the near-term opportunity for rack-to-rack, it's really represented by the scale-out network. But long term, we see that the scale-up network also represents a really large growth in TAM, given the fact that we expect the volumes for scale-up to be potentially in order of magnitude larger than the scale-up connections. So I think on several fronts, we can make the argument that we're still in the early stages. And I don't think there's any doubt in the market about if you can use copper, you will use copper given the advantages for reliability, power and cost. | **William Brennan:** μμ₯ κΈ°νμ λν΄ λ§μλ리면, TAMμμ κ°μ₯ ν° λ³νλ λ λ΄ λ°λκ° λμμ§λ©΄μ AECμ κΈ°νκ° νλλκ³ μλ€λ μΈμμ΄λΌκ³ μκ°ν©λλ€. κ·Έλ¦¬κ³ λλ ₯μ μ λ§ λͺ
νν©λλ€. λ°λ‘ μ λ’°μ±κ³Ό μ λ ₯ ν¨μ¨μ±μ
λλ€. μΈ λ²μ§Έλ μλ§λ μμ€ν
λΉμ©μΌ ν
λ°, μ΄λ κ΄ν μ루μ
κ³Ό λλΉλλ λΆλΆμ
λλ€. λ°λΌμ μΈνΈλΌ λ(intra-rack) μμ₯μ μ¬μ ν μ΄κΈ° λ¨κ³μ μλ€κ³ λ³Ό μ μμ§λ§, μ΄μ 4κ°μ νμ΄νΌμ€μΌμΌλ¬μμ μ루μ μ΄ λ³Έκ²©μ μΌλ‘ λμ λκΈ° μμνμ΅λλ€. μ ν¬λ λͺ¨λ νμ΄νΌμ€μΌμΌλ¬λ€μ΄ ν₯ν μ΄λ μμ μλ AECλ₯Ό μ±νν κ²μΌλ‘ μΆ©λΆν μμνλ€κ³ λ§μλλ¦° λ° μμ΅λλ€. μΈνΈλΌ λ(intra-rack)μ λν΄ λ§μλ리면, μ€μ λ‘λ μλ² λμ μλ―Ένλλ°, μ΄λ AI μ΄νλΌμ΄μΈμ€λ μΌλ° μ»΄ν¨ν μ΄λ μκ΄μμ΅λλ€. κ·Έλμ NIC to TORμ΄ ν° μμ₯μ λλ€. λν μ€μμΉ λμ λν΄μλ λ§μλλ¦° λ° μμ΅λλ€. μμ₯μμλ μμ λΆλ¦¬(disaggregation) νΈλ λκ° μλλ°, μ΄λ μλ κ° μ§μλμ΄μ¨ μΆμΈμ λλ€. ν΄λΉ μμ₯ λΆλ¬Έμ΄ NIC to TORλ§νΌ λΉ λ₯΄κ² μ±μ₯νμ§ μλ κ²μ 보μκΈ° λλ¬Έμ, μ΄λ ν° μ±μ₯ κΈ°νλ₯Ό μλ―Έν©λλ€. λ ν¬ λ(rack-to-rack)μ λν΄ λ§μλ리면, μ΄λ μ€μ λ‘ TAM(μ 체 μμ₯ κ·λͺ¨)μμ νμ₯λκ³ μλ λΆλΆμΌλ‘, AI λ°±μλ λ€νΈμν¬μ μ€μμΉ λ μμͺ½ μμ₯ λͺ¨λμμ κ·Έλ μ΅λλ€. μ€μμΉ λλ€μ΄ λ©ν° λ μν€ν μ²λ‘ μ νλκΈ° μμνκ³ μμ΅λλ€. λ κ° μ°κ²°(rack-to-rack)μ λ¨κΈ° κΈ°νλ₯Ό λ§μλ리면, μ΄λ μ€μ λ‘ μ€μΌμΌμμ λ€νΈμν¬(scale-out network)λ‘ λνλ©λλ€. νμ§λ§ μ₯κΈ°μ μΌλ‘λ μ€μΌμΌμ λ€νΈμν¬(scale-up network)λ TAMμμ μ λ§ ν° μ±μ₯μ λνλΌ κ²μΌλ‘ λ³΄κ³ μμ΅λλ€. μ€μΌμΌμ μ λ³Όλ₯¨μ΄ μ€μΌμΌμ μ°κ²°λ³΄λ€ μ μ¬μ μΌλ‘ ν μλ¦Ώμ λ ν΄ κ²μΌλ‘ μμνκΈ° λλ¬Έμ λλ€. λ°λΌμ μ¬λ¬ μΈ‘λ©΄μμ μ°λ¦¬λ μ¬μ ν μ΄κΈ° λ¨κ³μ μλ€κ³ λ§ν μ μλ€κ³ μκ°ν©λλ€. κ·Έλ¦¬κ³ κ΅¬λ¦¬λ₯Ό μ¬μ©ν μ μλ€λ©΄ μ λ’°μ±, μ λ ₯, λΉμ© λ©΄μμμ μ₯μ μ κ³ λ €ν λ ꡬ리λ₯Ό μ¬μ©ν κ²μ΄λΌλ μ μ λν΄μλ μμ₯μμ μμ¬μ μ¬μ§κ° μλ€κ³ μκ°ν©λλ€. |
Tore Svanberg: Stifel, Nicolaus & Company, Incorporated, Research Division That's really helpful. And I had a follow-up question on the customer concentration and drivers for fiscal '26. So I think Dan mentioned that your largest customer will still be a significant growth driver in fiscal '26. But that particular customer, I think, was sort of down the last 2 quarters. So is that customer sort of just pausing to ramp? Or are the use cases going to expand with that particular customer in fiscal '26? | **Tore Svanberg:** μ λ§ λμμ΄ λ©λλ€. κ³ κ° μ§μ€λμ 2026 νκ³μ°λ μ±μ₯ λλ ₯μ λν νμ μ§λ¬Έμ΄ μμ΅λλ€. Danμ΄ μ΅λ κ³ κ°μ΄ 2026 νκ³μ°λμλ μ¬μ ν μ€μν μ±μ₯ λλ ₯μ΄ λ κ²μ΄λΌκ³ μΈκΈν κ² κ°μλ°μ. νμ§λ§ κ·Έ νΉμ κ³ κ°μ μ§λ 2λΆκΈ° λμ λ§€μΆμ΄ κ°μνλ κ² κ°μ΅λλ€. κ·Έ κ³ κ°μ΄ λ¨μν νμ₯μ μν΄ μ μ λ©μΆ κ²μΈμ§, μλλ©΄ 2026 νκ³μ°λμ κ·Έ νΉμ κ³ κ°κ³Όμ μ¬μ© μ¬λ‘κ° νλλ μμ μΈμ§ κΆκΈν©λλ€. |
Chief Financial Officer: I think maybe the answer to both of those questions is probably yes. We -- it's -- their history, if you look back over time at how they've purchased inventory from us, it's been not consistently up and to the right, as Bill noted. So -- and that's kind of what we believe to be kind of an ordinary purchasing pattern for potentially any of these hyperscalers. But I did just for clarity, Tore, I wanted to make sure that it's understood by everyone that our largest customer from fiscal '25, we expect to be the largest customer in our fiscal '26 by the year-end and by far, actually. So it won't be a close second in all likelihood as we have forecast today from all of these customers. | **Chief Financial Officer:** λ μ§λ¬Έμ λν λ΅λ³μ μλ§λ λͺ¨λ 'μ'μΌ κ² κ°μ΅λλ€. μ΄λ€μ κ³Όκ±° μ΄λ ₯μ 보면, μκ°μ΄ μ§λλ©΄μ μ°λ¦¬λ‘λΆν° μ¬κ³ λ₯Ό ꡬ맀νλ ν¨ν΄μ΄ λΉμ΄ μΈκΈνλ―μ΄ μ§μμ μΌλ‘ μ°μν₯νμ§λ μμμ΅λλ€. κ·Έλ¦¬κ³ μ΄κ²μ΄ μ΄λ¬ν νμ΄νΌμ€μΌμΌλ¬λ€μ μΌλ°μ μΈ κ΅¬λ§€ ν¨ν΄μ΄λΌκ³ μ ν¬λ μκ°ν©λλ€. νμ§λ§ ν λ , λͺ νν νκΈ° μν΄ λͺ¨λ λΆλ€μ΄ μ΄ν΄νμ ¨μΌλ©΄ νλ μ μ΄ μμ΅λλ€. 25νκ³μ°λ μ΅λ κ³ κ°μ΄ 26νκ³μ°λ λ§κΉμ§λ λ¨μ°μ½ μ΅λ κ³ κ°μ΄ λ κ²μΌλ‘ μμνλ€λ μ μ λλ€. νμ¬ λͺ¨λ κ³ κ°λ€λ‘λΆν°μ μμμΉλ₯Ό 보면, μλ§λ κ·Όμν μ°¨μ΄μ 2μλ μμ κ² κ°μ΅λλ€. |
Operator: And our next question comes from the line of Christopher Rolland with Susquehanna. | **Operator:** λ€μ μ§λ¬Έμ μμ€μΌνλμ ν¬λ¦¬μ€ν νΌ λ‘€λλλκ»μ μ£Όμ μ§λ¬Έμ λλ€. |
Christopher Rolland: Susquehanna Financial Group, LLLP, Research Division On the results and this will probably be for Bill. So I guess, first of all, for 1.6T, I believe there are some supply constraints that appear to be emerging for 1.6T optical and lasers in particular. So I have two questions here. The first one is, do you think this affects your optical DSP outlook? And then secondly, do you think that these optical constraints could actually accelerate your AEC business? | **Christopher Rolland:** μ€μ μ λν μ§λ¬Έμ΄κ³ μλ§ λΉμκ² λ리λ μ§λ¬ΈμΌ κ² κ°μ΅λλ€. μ°μ 1.6Tμ κ²½μ°, 1.6T κ΄ν λΆνκ³Ό νΉν λ μ΄μ μμ μΌλΆ κ³΅κΈ μ μ½μ΄ λνλκ³ μλ κ²μΌλ‘ 보μ λλ€. μ¬κΈ°μ λ κ°μ§ μ§λ¬Έμ λλ¦¬κ² μ΅λλ€. 첫 λ²μ§Έλ, μ΄κ²μ΄ κ·νμ κ΄ν DSP μ λ§μ μν₯μ λ―ΈμΉ κ²μΌλ‘ μκ°νμλμ§μ? κ·Έλ¦¬κ³ λ λ²μ§Έλ‘, μ΄λ¬ν κ΄ν λΆν μ μ½μ΄ μ€μ λ‘ κ·νμ AEC μ¬μ μ κ°μνν μ μλ€κ³ 보μλμ§μ? |
William Brennan: President, CEO & Chairman Yes. So I think, generally speaking, for our optical business, I think we're seeing great results. And this is really -- we began our big ramp with 400 gig or 50-gig per lane solutions. We're seeing recent success with 800 gig, and we expect to be able to grow for a long period of time with that 100-gig per lane market. I think we've been pretty consistent saying that the transition in lane speed over time typically takes a lot longer, and it's really driven by each independent customer strategy. So we see big opportunity for 100 gig per lane. We see the same thing that you're seeing from a 1.6T transceiver perspective that it's going to take more time than the industry analysts have predicted. And so we don't see any slowing of the growth that we're expecting based on that delay. It will be quite good if there's an acceleration. But generally speaking, it doesn't really impact our bullishness as it relates to the optical space. | **William Brennan:** λ€, μΌλ°μ μΌλ‘ λ§μλ리면 κ΄ν μ¬μ
μμ νλ₯ν κ²°κ³Όλ₯Ό λ³΄κ³ μμ΅λλ€. μ΄λ μ€μ λ‘ 400κΈ°κ° λλ λ μΈλΉ 50κΈ°κ° μ루μ
μΌλ‘ λκ·λͺ¨ λ¨νμ
μ μμν κ²μμ λΉλ‘―λ©λλ€. μ΅κ·Ό 800κΈ°κ°μμ μ±κ³΅μ κ±°λκ³ μμΌλ©°, λ μΈλΉ 100κΈ°κ° μμ₯μμ μ₯κΈ°κ° μ±μ₯ν μ μμ κ²μΌλ‘ κΈ°λν©λλ€. μκ°μ΄ μ§λλ©΄μ λ μΈ μλμ μ νμ μΌλ°μ μΌλ‘ ν¨μ¬ μ€λ 걸리며, μ΄λ μ€μ λ‘ κ° κ³ κ°μ λ 립μ μΈ μ λ΅μ μν΄ μ’μ°λλ€κ³ μ§μμ μΌλ‘ λ§μλλ € μμ΅λλ€. λ°λΌμ λ μΈλΉ 100κΈ°κ°μμ ν° κΈ°νλ₯Ό λ³΄κ³ μμ΅λλ€. 1.6T νΈλμλ² κ΄μ μμ 보μλ κ²κ³Ό λμΌνκ², μ κ³ λΆμκ°λ€μ΄ μμΈ‘ν κ²λ³΄λ€ λ λ§μ μκ°μ΄ 걸릴 κ²μ΄λΌκ³ λ΄ λλ€. κ·Έ μ§μ°μΌλ‘ μΈν΄ μ ν¬κ° μμνκ³ μλ μ±μ₯μΈκ° λνλλ κ²μ μ ν 보μ΄μ§ μμ΅λλ€. λ§μ½ κ°μνκ° μλ€λ©΄ μλΉν μ’κ² μ§λ§, μΌλ°μ μΌλ‘ λ§μλ리면 κ΄ν λΆμΌμ λν μ ν¬μ λκ΄μ μΈ μ λ§μλ μ€μ§μ μΌλ‘ μν₯μ μ£Όμ§ μμ΅λλ€. |
Christopher Rolland: Susquehanna Financial Group, LLLP, Research Division And might it accelerate your AEC if there are supply constraints? | **Christopher Rolland:** κ³΅κΈ μ μ½μ΄ μμ κ²½μ° AECκ° κ°μνλ μ μμκΉμ? |
William Brennan: President, CEO & Chairman I think that the AEC TAM is going to stand alone. And people are going to choose to use AECs based on whether they can or not. I think for sure, what we're seeing consistently of our customers can use an AEC, they will use an AEC. And so I don't think supply constraints in optical are going to affect that at all. | **William Brennan:** AEC TAMμ λ 립μ μΌλ‘ μ‘΄μ¬ν κ²μ΄λΌκ³ μκ°ν©λλ€. μ¬λλ€μ AECλ₯Ό μ¬μ©ν μ μλμ§ μ¬λΆμ λ°λΌ μ ννκ² λ κ²μ λλ€. νμ€ν κ²μ κ³ κ°λ€μ΄ AECλ₯Ό μ¬μ©ν μ μλ€λ©΄ λ°λμ μ¬μ©νλ€λ μ μ μ§μμ μΌλ‘ νμΈνκ³ μλ€λ κ²μ λλ€. λ°λΌμ κ΄ν λΆλ¬Έμ κ³΅κΈ μ μ½μ΄ μ΄μ μ ν μν₯μ λ―ΈμΉμ§ μμ κ²μ΄λΌκ³ λ΄ λλ€. |
Christopher Rolland: Susquehanna Financial Group, LLLP, Research Division Excellent. And for a second question, Bill, I think there's some progress here on your fourth hyperscaler for sure. Any more color on when the fifth could be a 10% or? And then I know we're running out of hyperscalers, but any more color on any other engagements, including the big GPU guy. Is there any possibility of landing them for branded cables? Or do you think they have their own initiative here? | **Christopher Rolland:** λ€ λ²μ§Έ νμ΄νΌμ€μΌμΌλ¬μμ νμ€ν μ§μ μ΄ μλ κ² κ°μ΅λλ€. λ€μ― λ²μ§Έ νμ΄νΌμ€μΌμΌλ¬κ° μΈμ 10%μ λλ¬ν μ μμμ§μ λν΄ λ μμΈν λ§μν΄ μ£Όμ€ μ μλμ? κ·Έλ¦¬κ³ νμ΄νΌμ€μΌμΌλ¬λ€μ΄ μΌλ§ λ¨μ§ μμλ€λ κ²μ μκ³ μμ§λ§, λν GPU μ 체λ₯Ό ν¬ν¨ν λ€λ₯Έ μ°Έμ¬ μ 체λ€μ λν΄ λ μμΈν λ΄μ©μ μλ €μ£Όμ€ μ μλμ? λΈλλ μΌμ΄λΈ λΆμΌμμ κ·Έλ€μ ν보ν κ°λ₯μ±μ΄ μμκΉμ? μλλ©΄ κ·Έλ€μ΄ μ체μ μΈ κ³νμ κ°μ§κ³ μλ€κ³ μκ°νμλμ? |
William Brennan: President, CEO & Chairman Yes. So first, I'll talk about what's upcoming. And I think your question was maybe one step ahead of where we are. The first thing is we need to get that fifth one in production. We're through qualification. And I think that from a timing standpoint, no changes to what we've talked about in the past. We expect towards the end of this fiscal year that we'll see first traction. I've mentioned that all of these hyperscalers have the ability to be a 10% customer for us. It's a function of their deployment plans specifically as well as where in their network they're planning on using AECs. There's a lot of interesting things happening. We talked about the neo cloud market. We've talked about sovereign. And what we're seeing is that -- and prior to the comment, I'll say that finally, you talked about the big GPU guy. And what we're seeing even today is that we're connecting with many GPUs and including the big guy in the market. And so I think there's a recognition that the things we're doing are advantageous to the overall performance and power efficiency of AI clusters. And so we're connecting to plenty of NVIDIA GPUs now as we're in production. And we see opportunities that go beyond hyperscalers as CapEx globally expands with Neo clouds as well as sovereign opportunities. So I think that every time you see a new announcement of a gigawatt data center, you can rest assured that, that -- we view that as an opportunity as probably everybody else in the market views it. | **William Brennan:** λ€, λ¨Όμ μμΌλ‘ μμ λ κ²λ€μ λν΄ λ§μλλ¦¬κ² μ΅λλ€. μ§λ¬Έμ΄ μ ν¬κ° νμ¬ μλ λ¨κ³λ³΄λ€ ν λ° μμ μλ κ² κ°μ΅λλ€. κ°μ₯ λ¨Όμ ν΄μΌ ν μΌμ λ€μ― λ²μ§Έ κ³ κ°μ μμ° λ¨κ³μ μ§μ
μν€λ κ²μ
λλ€. μ격 μ¬μ¬λ ν΅κ³Όνμ΅λλ€. νμ΄λ° κ΄μ μμ 보면, κ³Όκ±°μ λ§μλλ¦° κ²κ³Ό λ³κ²½μ¬νμ μμ΅λλ€. μ΄λ² νκ³μ°λ λ§μ―€μ 첫 λ²μ§Έ 견μΈλ ₯μ λ³΄κ² λ κ²μΌλ‘ μμν©λλ€. μ΄λ―Έ μΈκΈνλ―μ΄ μ΄ λͺ¨λ νμ΄νΌμ€μΌμΌλ¬λ€μ μ ν¬μκ² 10% κ³ κ°μ΄ λ μ μλ λ₯λ ₯μ κ°μ§κ³ μμ΅λλ€. μ΄λ κ·Έλ€μ ꡬ체μ μΈ λ°°μΉ κ³νκ³Ό λ€νΈμν¬ λ΄μμ AECλ₯Ό μ΄λμ μ¬μ©ν κ³νμΈμ§μ λ°λΌ λ¬λΌμ§λλ€. ν₯λ―Έλ‘μ΄ μΌλ€μ΄ λ§μ΄ μΌμ΄λκ³ μμ΅λλ€. λ€μ€ ν΄λΌμ°λ μμ₯μ λν΄μλ μ΄μΌκΈ°νκ³ , μλ²λ¦°μ λν΄μλ λ Όμνμ΅λλ€. κ·Έλ¦¬κ³ μ°λ¦¬κ° λ³΄κ³ μλ κ²μ -- κ·Έ λ§μμ λ리기 μ μ, ν° GPU μ 체μ λν΄ λ§μνμ ¨λλ°μ. μ€λλ μλ μ°λ¦¬λ μμ₯μ κ·Έ ν° μ 체λ₯Ό ν¬ν¨ν΄μ λ§μ GPUλ€κ³Ό μ°κ²°νκ³ μμ΅λλ€. κ·Έλμ μ°λ¦¬κ° νκ³ μλ μΌλ€μ΄ AI ν΄λ¬μ€ν°μ μ λ°μ μΈ μ±λ₯κ³Ό μ λ ₯ ν¨μ¨μ±μ μ 리νλ€λ μΈμμ΄ μλ€κ³ μκ°ν©λλ€. νμ¬ μμ° λ¨κ³μμ μλ§μ NVIDIA GPUλ€κ³Ό μ°κ²°νκ³ μκ³ μ. κ·Έλ¦¬κ³ νμ΄νΌμ€μΌμΌλ¬λ₯Ό λμ΄μλ κΈ°νλ€μ λ³΄κ³ μμ΅λλ€. μ μΈκ³μ μΌλ‘ μλ³Έμ§μΆ(CapEx)μ΄ λ€μ€ ν΄λΌμ°λμ κ΅κ° μ£ΌκΆ κΈ°νλ€λ‘ νμ₯λλ©΄μ λ§μ΄μ£ . λ°λΌμ κΈ°κ°μνΈκΈ λ°μ΄ν°μΌν°μ μλ‘μ΄ λ°νλ₯Ό λ³Ό λλ§λ€, κ·Έκ²μ μ°λ¦¬κ° κΈ°νλ‘ λ³Έλ€κ³ νμ νμ λ λ©λλ€. μλ§ μμ₯μ λ€λ₯Έ λͺ¨λ μ 체λ€λ κ·Έλ κ² λ³΄κ² μ§λ§μ. |
Operator: And our next question comes from the line of Vijay Rakesh with Mizuho. | **Operator:** λ€μ μ§λ¬Έμ λ―Έμ¦νΈμ λΉμ μ΄ λΌμΌμλκ»μ μ£Όμ μ§λ¬Έμ λλ€. |
Vijay Rakesh: Mizuho Securities USA LLC, Research Division Dan, congratulations on a pretty good quarter and guide here as well. So on the fourth CSP, it looks like a pretty nice ramp there. It looks like it's growing a lot faster than your prior ramps. Just wondering, is that driven by multiple engagements, multiple cable wins you have within there? Or are you seeing a nice pickup on ASPs as you go to 100 gig per lane? | **Vijay Rakesh:** λ€ λ²μ§Έ CSPμμ κ½€ μ’μ μμΉμΈλ₯Ό 보μ΄κ³ μλ€μ. μ΄μ μμΉμΈλ€λ³΄λ€ ν¨μ¬ λΉ λ₯΄κ² μ±μ₯νκ³ μλ κ² κ°μ΅λλ€. μ΄κ²μ΄ μ¬λ¬ κ³μ½ 체결μ΄λ μ¬λ¬ μΌμ΄λΈ μμ£Όλ‘ μΈν κ²μΈμ§, μλλ©΄ λ μΈλΉ 100κΈ°κ°λ‘ μ΄λνλ©΄μ ASP(νκ· νλ§€ κ°κ²©)κ° μ’κ² μμΉνκ³ μλ κ²μΈμ§ κΆκΈν©λλ€. |
William Brennan: President, CEO & Chairman I apologize, but you broke up as you're asking the question. Do you think you could repeat that? | **William Brennan:** μ£μ‘ν©λλ€λ§ μ§λ¬Ένμλ λμ€μ μμ±μ΄ λμ΄μ‘μ΅λλ€. λ€μ ν λ² λ§μν΄ μ£Όμκ² μ΅λκΉ? |
Vijay Rakesh: Mizuho Securities USA LLC, Research Division Yes. The ramp with the fourth customer seems to be a pretty nice accelerated ramp. Are you seeing -- is that driven by a much higher content on 100 gig per lane or what is driving the pretty accelerated pickup there, I guess? | **Vijay Rakesh:** λ€, λ€ λ²μ§Έ κ³ κ°μ¬μμ λ¨νμ μ΄ μλΉν κ°μνλκ³ μλ κ² κ°μ΅λλ€. μ΄κ²μ΄ λ μΈλΉ 100κΈ°κ° νκ²½μμ ν¨μ¬ λμ μ½ν μΈ λΉμ€ λλ¬ΈμΈμ§, μλλ©΄ λ€λ₯Έ μμΈμ΄ μ΄λ° κ°μνλ μ±μ₯μ μ΄λκ³ μλ κ±΄μ§ κΆκΈν©λλ€. |
William Brennan: President, CEO & Chairman Yes. I think that the accelerated pickup was based on our kind of natural position to be a little more conservative with the way that we're guiding. Things came together very quickly with their qualification of this first program that's ramping. And I think you can get a feel for the scale of all of these deployments. I don't think there's any big change from the standpoint of ASPs or margins. I think it's right down the middle of the fairway from what we're seeing elsewhere. | **William Brennan:** λ€, κ°μνλ μ±μ₯μ μ ν¬κ° κ°μ΄λμ€λ₯Ό μ μν λ λ€μ 보μμ μΈ μ κ·Όμ μ·¨νλ μμ°μ€λ¬μ΄ ν¬μ§μ μ κΈ°λ°ν κ²μ΄λΌκ³ μκ°ν©λλ€. νμ¬ λ¨νμ μ€μΈ 첫 λ²μ§Έ νλ‘κ·Έλ¨μ μ격 μκ±΄μ΄ λ§€μ° λΉ λ₯΄κ² μΆ©μ‘±λλ©΄μ μν©μ΄ κΈμν μ§μ λμμ΅λλ€. μ΄λ¬ν λͺ¨λ λ°°μΉμ κ·λͺ¨κ° μ΄λ μ λμΈμ§ κ°μ μ‘μΌμ€ μ μμ κ²μ λλ€. ASPλ λ§μ§ μΈ‘λ©΄μμλ ν° λ³νκ° μλ€κ³ λ΄ λλ€. λ€λ₯Έ κ³³μμ λ³΄κ³ μλ κ²κ³Ό μ νν μ€κ° μ§μ μ μμΉνλ€κ³ μκ°ν©λλ€. |
Vijay Rakesh: Mizuho Securities USA LLC, Research Division Got it. And then on the margin guide, I think I know you mentioned some impact from -- is the read through that the impact is potentially more from the talent side? Or is there some product mix in there? | **Vijay Rakesh:** λ§μ§ κ°μ΄λμ€μ λν΄μλ, λ§μνμ μν₯μ΄ μ£Όλ‘ μΈμ¬ μΈ‘λ©΄μμ μ€λ κ²μΈμ§μ? μλλ©΄ μ ν λ―Ήμ€λ μ΄λ μ λ μν₯μ μ£Όλ 건κ°μ? |
Chief Financial Officer: Yes. It's -- let me just state it this way. We -- if you look at Q4 to Q1, we still or a bit of an impact of the benefit of increasing scale, kind of a very small 20 basis point sequential increase in gross margin. So we're at the stage or we're at the scale where any fluctuations are probably going to be caused by some product mix issues. So -- and all of these things, none of them will ever be linear. We've been very clear in stating that, although we've been very fortunate over the last few [indiscernible] So, I guess, the key message to take away is long term, we don't see any differences in what we stated in our gross margin model [indiscernible] even though we're are closed right now where that gross margin at or above. | **Chief Financial Officer:** λ€, μ΄λ κ² λ§μλλ¦¬κ² μ΅λλ€. 4λΆκΈ°μμ 1λΆκΈ°λ₯Ό 보μλ©΄, κ·λͺ¨ νλμ ννμ΄ μ¬μ ν μ½κ° μμκ³ , λ§€μ° μμ νμ΄μ§λ§ λ§€μΆμ΄μ΄μ΅λ₯ μ΄ μμ°¨μ μΌλ‘ 20λ² μ΄μμ€ν¬μΈνΈ μ¦κ°νμ΅λλ€. νμ¬ μ°λ¦¬κ° λλ¬ν κ·λͺ¨μμλ μμΌλ‘μ λ³λμ μλ§λ μ ν λ―Ήμ€ μ΄μλ‘ μΈν΄ λ°μν κ°λ₯μ±μ΄ λμ΅λλ€. κ·Έλ¦¬κ³ μ΄ λͺ¨λ κ²λ€μ μ λ μ νμ μ΄μ§ μμ κ²μ
λλ€. μ°λ¦¬λ μ΄ μ μ λ§€μ° λͺ
ννκ² λ°νμκ³ , μ§λ λͺ λ
κ° λ§€μ° μ΄μ΄ μ’μλ€κ³ μκ°ν©λλ€. ν΅μ¬ λ©μμ§λ μ₯κΈ°μ μΌλ‘ μ°λ¦¬κ° μ μν λ§€μΆμ΄μ΄μ΅λ₯ λͺ¨λΈμμ λ¬λΌμ§ κ²μ΄ μλ€λ μ μ λλ€. νμ¬ ν΄λΉ λ§€μΆμ΄μ΄μ΅λ₯ μμ€μ κ·Όμ νκ±°λ κ·Έ μ΄μμ μλλΌλ λ§μ λλ€. |
Operator: [Operator Instructions] And our next question comes from the line of Quinn Bolton with Needham & Company. | **Operator:** λ€μ μ§λ¬Έμ Needham & Companyμ Quinn BoltonλμΌλ‘λΆν° λ°μμ΅λλ€. |
Quinn Bolton: Needham & Company, LLC, Research Division I guess maybe Dan and Bill, the first question I had, that fourth hyperscaler, that first program, is that rack-to-rack? Or is that intra-rack? And do you see other opportunities at that hyperscaler ramping so that you ultimately will have both intra-rack and rack-to-rack programs at that customer? And then I've got a follow-up. | **Quinn Bolton:** 첫 λ²μ§Έ μ§λ¬Έμ λκ³Ό λΉμκ² λ리λ 건λ°μ, λ€ λ²μ§Έ νμ΄νΌμ€μΌμΌλ¬μ 첫 λ²μ§Έ νλ‘κ·Έλ¨μ΄ λ κ°(rack-to-rack) μ°κ²°μΈκ°μ, μλλ©΄ λ λ΄λΆ(intra-rack) μ°κ²°μΈκ°μ? κ·Έλ¦¬κ³ ν΄λΉ νμ΄νΌμ€μΌμΌλ¬μμ λ€λ₯Έ κΈ°νλ€μ΄ νλλμ΄ κΆκ·Ήμ μΌλ‘λ λ λ΄λΆμ λ κ° μ°κ²° νλ‘κ·Έλ¨μ λͺ¨λ ν보νκ² λ κ°λ₯μ±μ΄ μλμ§ κΆκΈν©λλ€. κ·Έλ¦¬κ³ μΆκ° μ§λ¬Έμ΄ μμ΅λλ€. |
William Brennan: President, CEO & Chairman I appreciate the question and the goal to understand more. I'm not going to be too specific with the exact application, but I will say it's not a single rack architecture. | **William Brennan:** μ§λ¬Έν΄ μ£Όμκ³ λ μμΈν μ΄ν΄νλ €κ³ ν΄μ£Όμ μ κ°μ¬ν©λλ€. μ νν μ ν리μΌμ΄μ μ λν΄μλ λ무 ꡬ체μ μΌλ‘ λ§μλ리μ§λ μκ² μ§λ§, λ¨μΌ λ μν€ν μ²λ μλλΌλ μ μ λ§μλ릴 μ μμ΅λλ€. |
Quinn Bolton: Needham & Company, LLC, Research Division Okay. And are there multiple programs at that customer you expect to ramp over time? | **Quinn Bolton:** λ€, ν΄λΉ κ³ κ°μ¬μμ μκ°μ΄ μ§λλ©΄μ λ¨νμ λ κ²μΌλ‘ μμνλ μ¬λ¬ νλ‘κ·Έλ¨λ€μ΄ μλμ? |
William Brennan: President, CEO & Chairman Yes, sure. That's -- we're right on track with where we would expect to be as we turn on a new relationship. Once there's a very good first experience, it typically leads to next-generation planning. And this is really where we do our best work is as people are looking at architecting the next-generation deployment, we typically get involved very early in architecting the racks even. That's one of the areas where we're able to show several different racking options as we talk about next generation. And so it's really natural from a customer standpoint that after we get the first program kicked off that we're almost invited in as a collaborative partner on the next-generation rack and row designs. | **William Brennan:** λ€, λ¬Όλ‘ μ
λλ€. μλ‘μ΄ κ΄κ³λ₯Ό μμν λ μμνλ λλ‘ μμ‘°λ‘κ² μ§νλκ³ μμ΅λλ€. 첫 λ²μ§Έ κ²½νμ΄ λ§€μ° μ’μΌλ©΄ μΌλ°μ μΌλ‘ μ°¨μΈλ κ³νμΌλ‘ μ΄μ΄μ§κ² λ©λλ€. μ΄κ²μ΄ λ°λ‘ μ ν¬κ° κ°μ₯ μνλ λΆμΌμΈλ°, κ³ κ°λ€μ΄ μ°¨μΈλ λ°°ν¬λ₯Ό μ€κ³ν λ μ ν¬λ λ³΄ν΅ λ μ€κ³ λ¨κ³λΆν° λ§€μ° μ΄κΈ°μ μ°Έμ¬νκ² λ©λλ€. μ°¨μΈλμ λν΄ λ Όμν λ μ¬λ¬ κ°μ§ λ€μν λνΉ μ΅μ μ 보μ¬λ릴 μ μλ μμ μ€ νλμ λλ€. λ°λΌμ κ³ κ° κ΄μ μμ 보면 첫 λ²μ§Έ νλ‘κ·Έλ¨μ μμν ν μ°¨μΈλ λκ³Ό λ‘μ° μ€κ³μμ νλ ₯ ννΈλλ‘ μ΄λλ°λ κ²μ΄ λ§€μ° μμ°μ€λ¬μ΄ μΌμ λλ€. |
Quinn Bolton: Needham & Company, LLC, Research Division Excellent. And then second question for me. Over the last, I think, month or so, you guys have announced patent lawsuit settlements, I believe, with Amphenol and Volex. I know you can't get into specifics of either of those agreements. But is there anything you can say about whether you've granted licenses that could generate royalties going forward? And just any thoughts you have now that perhaps you've settled lawsuits with two of the four cable vendors. How do you see competition going forward? Are you effectively enabling greater competition by granting those licenses or signing those patent settlement agreements with other suppliers now of AECs? | **Quinn Bolton:** νλ₯ν©λλ€. κ·Έλ¦¬κ³ λ λ²μ§Έ μ§λ¬Έμ
λλ€. μ§λ ν λ¬ μ λ λμ Amphenolκ³Ό Volexμ νΉν μμ‘ ν©μλ₯Ό λ°ννμ κ²μΌλ‘ μκ³ μμ΅λλ€. λ ν©μμ ꡬ체μ μΈ λ΄μ©μ λν΄μλ λ§μνμ€ μ μλ€λ κ²μ μκ³ μμ΅λλ€. νμ§λ§ ν₯ν λ‘μ΄ν°λ₯Ό μ°½μΆν μ μλ λΌμ΄μ μ€λ₯Ό λΆμ¬νλμ§μ λν΄ λ§μν΄ μ£Όμ€ μ μλ λΆλΆμ΄ μλμ? κ·Έλ¦¬κ³ 4κ° μΌμ΄λΈ λ²€λ μ€ 2κ³³κ³Ό μμ‘μ ν©μνμ μν©μμ, μμΌλ‘μ κ²½μ μν©μ μ΄λ»κ² λ³΄κ³ κ³μ μ§μ? λ€λ₯Έ AEC 곡κΈμ 체λ€κ³Ό λΌμ΄μ μ€λ₯Ό λΆμ¬νκ±°λ νΉν ν©μ κ³μ½μ 체결ν¨μΌλ‘μ¨ μ¬μ€μ λ ν° κ²½μμ κ°λ₯νκ² νκ³ μλ κ²μΈμ§μ λν μκ°λ λ£κ³ μΆμ΅λλ€. |
William Brennan: President, CEO & Chairman Sure. So I'll keep the comments somewhat high level. But I think that if we go back in time, this is a market that we pioneered. And along the way, we developed a lot of intellectual property. We think it's reasonable to have the expectation to have our IP be respected in the market. We're very happy that the agreements that we've announced have taken us to the point where we would expect to be. The -- and you'll see that we're pretty active in conversations with others as well. Really nothing in these announcements changes our sense of our place in the market. And when we think about our market share or forecast going forward. We've been really consistent in saying that the market for AECs is going to be very large, really over the next 5 to 10 years, and it's large enough for multiple winners. And when we get back to how we compete as a supplier to our partners, -- it's always driven by customer engagement. Our goal is to deliver the next-generation solutions first, achieve qualification first and ultimately helping our customers ramp successfully. This is what we believe is going to be the big difference maker long term in how market share is ultimately going to play out. So we're trying to do such a good job for our customers that we're difficult to compete with. And I think we've talked about our moat and we've talked about expanding our moat, and we feel quite good about owning all aspects of the product, and we think it's uniquely competitive. | **William Brennan:** λ€, μ’ λ λμ μ°¨μμμ λ§μλλ¦¬κ² μ΅λλ€. μκ°μ κ±°μ¬λ¬ μ¬λΌκ° 보면, μ΄ μμ₯μ μ ν¬κ° κ°μ²ν μμ₯μ λλ€. κ·Έ κ³Όμ μμ μ ν¬λ λ§μ μ§μ μ¬μ°κΆμ κ°λ°νμ΅λλ€. μμ₯μμ μ ν¬ IPκ° μ‘΄μ€λ°κΈ°λ₯Ό κΈ°λνλ κ²μ ν©λ¦¬μ μ΄λΌκ³ μκ°ν©λλ€. μ ν¬κ° λ°νν κ³μ½λ€μ΄ μ ν¬λ₯Ό κΈ°λνλ μ§μ κΉμ§ μ΄λμ΄ μ€ κ²μ λν΄ λ§€μ° λ§μ‘±μ€λ½κ² μκ°ν©λλ€. κ·Έλ¦¬κ³ λ³΄μλ€μνΌ μ ν¬λ λ€λ₯Έ μ 체λ€κ³Όλ μλΉν νλ°νκ² λνλ₯Ό λλκ³ μμ΅λλ€. μ΄λ² λ°νλ€λ‘ μΈν΄ μμ₯μμμ μ ν¬ μμΉμ λν μΈμμ΄ λ°λλ κ²μ μ ν μμ΅λλ€. μ ν¬ μμ₯ μ μ μ¨μ΄λ ν₯ν μ λ§μ μκ°ν λλ λ§μ°¬κ°μ§μ λλ€. μ ν¬λ AEC μμ₯μ΄ ν₯ν 5~10λ μ κ±Έμ³ λ§€μ° ν° κ·λͺ¨κ° λ κ²μ΄λΌκ³ μΌκ΄λκ² λ§μλλ €μκ³ , μ¬λ¬ μΉμκ° λμ¬ μ μμ λ§νΌ μΆ©λΆν ν° μμ₯μ λλ€. ννΈλλ€μκ² κ³΅κΈμ 체λ‘μ μ΄λ»κ² κ²½μνλλλ‘ λμκ°λ©΄, νμ κ³ κ° μ°Έμ¬κ° ν΅μ¬ λλ ₯μ λλ€. μ ν¬ λͺ©νλ μ°¨μΈλ μ루μ μ κ°μ₯ λ¨Όμ μ 곡νκ³ , μΈμ¦μ κ°μ₯ λ¨Όμ νλνλ©°, κΆκ·Ήμ μΌλ‘ κ³ κ°λ€μ μ±κ³΅μ μΈ λ¨νμ μ μ§μνλ κ²μ λλ€. μ΄κ²μ΄ μ₯κΈ°μ μΌλ‘ μμ₯ μ μ μ¨μ΄ μ΄λ»κ² κ²°μ λ μ§μ μμ΄ ν° μ°¨λ³ν μμκ° λ κ²μ΄λΌκ³ λ―Ώμ΅λλ€. μ ν¬λ κ³ κ°λ€μ μν΄ λ무λ νλ₯ν μΌμ ν΄μ κ²½μνκΈ° μ΄λ €μ΄ μ‘΄μ¬κ° λλ €κ³ ν©λλ€. μ ν¬κ° ν΄μμ λν΄ λ§μλλ Έκ³ ν΄μ νμ₯μ λν΄μλ λ Όμνλλ°, μ νμ λͺ¨λ μΈ‘λ©΄μ μμ νκ³ μλ€λ μ μμ μλΉν λ§μ‘±μ€λ½κ² μκ°νκ³ μμΌλ©°, μ΄κ²μ΄ λ νΉν κ²½μ μ°μλΌκ³ λ΄ λλ€. |
Operator: And our next question comes from the line of Suji Desilva with ROTH Capital. | **Operator:** λ€μ μ§λ¬Έμ ROTH Capitalμ Suji Desilvaλκ»μ μ£Όμ ¨μ΅λλ€. |
Sujeeva De Silva: ROTH Capital Partners, LLC, Research Division Congrats on the progress here. Trying to ask Quinn's question maybe a little different way. I'm just trying to -- the spirit of understanding how broadly inter-rack is being adopted. You have five customers you're working with. Do they all have programs that are going to be inter-rack cables along with 3-meter rack solutions? Or is it more of a niche with specific architectures? Just trying to get a feel for that dynamic though. | **Sujeeva De Silva:** μ§μ μ λν΄ μΆνλ립λλ€. Quinnμ μ§λ¬Έμ μ‘°κΈ λ€λ₯Έ λ°©μμΌλ‘ λ¬Όμ΄λ³΄λ €κ³ ν©λλ€. μΈν°λ(inter-rack) μ루μ μ΄ μΌλ§λ κ΄λ²μνκ² μ±νλκ³ μλμ§ μ΄ν΄νλ €κ³ ν©λλ€. νμ¬ 5κ° κ³ κ°μ¬μ νλ ₯νκ³ κ³μ λ°, μ΄λ€ λͺ¨λκ° 3λ―Έν° λ μ루μ κ³Ό ν¨κ» μΈν°λ μΌμ΄λΈμ ν¬ν¨νλ νλ‘κ·Έλ¨μ μ§ννκ³ μλμ? μλλ©΄ νΉμ μν€ν μ²μ κ΅νλ νμ μμ₯μΈκ°μ? κ·Έ μνκ΄κ³μ λν κ°μ μ‘κ³ μΆμ΅λλ€. |
William Brennan: President, CEO & Chairman Yes, I wouldn't say it's across the board. I would say that the first step typically is intra-rack, so 3-meter or less connections within the same rack. And this is just recent over the last 6 to 9 months that we've seen traction as our customers start to realize the opportunity to deliver much better cluster reliability and also secondarily better power. And so I would say that we're at the early stages still of having the market expand into rack-to-rack types of solutions. But I do think there's going to be an acceleration in the way that our customers view and use AECs. | **William Brennan:** λ€, μ λ°μ μΌλ‘ κ·Έλ° κ²μ μλλΌκ³ λ§μλλ¦¬κ² μ΅λλ€. μΌλ°μ μΌλ‘ 첫 λ²μ§Έ λ¨κ³λ μΈνΈλΌ λ(intra-rack), μ¦ λμΌν λ λ΄μμ 3λ―Έν° μ΄νμ μ°κ²°λΆν° μμλ©λλ€. κ·Έλ¦¬κ³ μ§λ 6κ°μμμ 9κ°μ μ¬μ΄μ κ³ κ°λ€μ΄ ν¨μ¬ λ λμ ν΄λ¬μ€ν° μμ μ±κ³Ό λΆμ°¨μ μΌλ‘λ λ λμ μ λ ₯ ν¨μ¨μ±μ μ 곡ν μ μλ κΈ°νλ₯Ό κΉ¨λ«κΈ° μμνλ©΄μ 견μΈλ ₯μ 보기 μμν κ²μ μ΅κ·Όμ μΌμ
λλ€. λ°λΌμ μμ₯μ΄ λ κ°(rack-to-rack) μ νμ μ루μ μΌλ‘ νμ₯λλ κ²μ μμ§ μ΄κΈ° λ¨κ³μ μλ€κ³ λ§μλλ¦¬κ² μ΅λλ€. νμ§λ§ κ³ κ°λ€μ΄ AECλ₯Ό 보λ κ΄μ κ³Ό νμ© λ°©μμμ κ°μνκ° μμ κ²μ΄λΌκ³ μκ°ν©λλ€. |
Sujeeva De Silva: ROTH Capital Partners, LLC, Research Division Okay. That's helpful color, Bill. And then you made a comment earlier about disaggregated architectures being slower to take up than you would have thought. Any color there? Any upcoming catalysts that would shift that transition of market inflection? And most importantly, would that -- what kind of credo addressable market multiplier might that create? | **Sujeeva De Silva:** λΆμ°ν μν€ν
μ²(disaggregated architectures)κ° μμλ³΄λ€ λμ
μ΄ λλ¦° κ²μ λν΄ μ’ λ μμΈν λ§μλλ¦¬κ² μ΅λλ€. μ΄ λΆλΆμμ μμ₯ μ νμ μ μλΉκΈΈ μ μλ μ΄λ§€μ κ° μλμ§, κ·Έλ¦¬κ³ κ°μ₯ μ€μν κ²μ μ΄κ²μ΄ μ°λ¦¬μ μ΄ μ£Όμ κ°λ₯ μμ₯(TAM)μ μ΄λ€ λ°°μ ν¨κ³Όλ₯Ό κ°μ Έμ¬ μ μλμ§μ λν μ§λ¬Έμ΄κ΅°μ. νμ¬ λΆμ°ν μν€ν μ² λμ μ΄ μ§μ°λλ μ£Όμ μ΄μ λ κΈ°μ‘΄ μΈνλΌμμμ μ ν λΉμ©κ³Ό 볡μ‘μ± λλ¬Έμ λλ€. νμ§λ§ λͺ κ°μ§ κΈμ μ μΈ μ νΈλ€μ΄ 보μ΄κ³ μμ΅λλ€. ν₯ν μ΄λ§€μ λ‘λ ν΄λΌμ°λ λ€μ΄ν°λΈ μ ν리μΌμ΄μ μ νμ°, μ£μ§ μ»΄ν¨ν μꡬμ¬ν μ¦κ°, κ·Έλ¦¬κ³ 5G λ€νΈμν¬ λ°°ν¬ κ°μνλ₯Ό λ€ μ μμ΅λλ€. νΉν λν ν΄λΌμ°λ μλΉμ€ μ 곡μ 체λ€μ΄ λΆμ°ν λͺ¨λΈλ‘μ μ νμ 본격ννλ©΄μ μμ₯ 견μΈλ ₯μ΄ μκΈ°κ³ μμ΅λλ€. TAM κ΄μ μμ 보면, λΆμ°ν μν€ν μ²κ° 본격 λμ λλ©΄ μ°λ¦¬ μμ₯ κΈ°νκ° 2-3λ°° νλλ κ²μΌλ‘ μμν©λλ€. κΈ°μ‘΄ μ€μμ§μ€μ λͺ¨λΈ λλΉ λ λ§μ μ μ μμ μ°λ¦¬ μ루μ μ΄ νμνκ² λκΈ° λλ¬Έμ λλ€. |
William Brennan: President, CEO & Chairman Sure. So we're seeing -- the first part of that market that's taking off is the switch racks that are dedicated to back-end networks for AI. When we thought about that market as we were really launching this product family a few years ago, it was really the leaf and spine within the front-end networks. And that part of the market has moved more slowly to next-generation speeds. So I think speed is going to be a catalyst for it to take off more broadly. But we are seeing the first uptake in 100-gig per lane back-end networks for the switching application specifically. | **William Brennan:** λ€, μ ν¬κ° λ³΄κ³ μλ κ²μ λ¨Όμ AIμ© λ°±μλ λ€νΈμν¬ μ μ© μ€μμΉ λ λΆλΆμ΄ μμ₯μμ μ±μ₯νκ³ μλ€λ μ μ
λλ€. λͺ λ
μ μ΄ μ νκ΅°μ μΆμν λΉμμλ νλ‘ νΈμλ λ€νΈμν¬ λ΄μ 리ν μ€ μ€νμΈ(leaf and spine) ꡬ쑰λ₯Ό μΌλμ λκ³ μμμ΅λλ€. κ·Έλ°λ° μμ₯μ μ΄ λΆλΆμ μ°¨μΈλ μλλ‘μ μ νμ΄ λ€μ λλκ² μ§νλκ³ μμ΅λλ€. λ°λΌμ μλκ° λ κ΄λ²μν μμ₯ μ±μ₯μ μ΄λ§€μ κ° λ κ²μ΄λΌκ³ μκ°ν©λλ€. νμ§λ§ νμ¬ μ€μμΉ μ ν리μΌμ΄μ μμ λ μΈλΉ 100κΈ°κ°(100-gig per lane) λ°±μλ λ€νΈμν¬μ 첫 λ²μ§Έ λμ μ΄ μμλκ³ μλ κ²μ νμΈνκ³ μμ΅λλ€. |
Operator: And we have no further questions at this time. Mr. Brennan, I will turn the call back over to you. | **Operator:** λ μ΄μ μ§λ¬Έμ΄ μμ΅λλ€. λΈλ λ λ, λ€μ λ§μν΄ μ£ΌμκΈ° λ°λλλ€. |
William Brennan: President, CEO & Chairman Well, thank you, everybody, for the questions, and I really appreciate the ongoing strong interest. So we'll speak quite soon. Thanks. | **William Brennan:** μ§λ¬Έν΄ μ£Όμ λͺ¨λ λΆλ€κ» κ°μ¬λ리며, μ§μμ μΈ κ΄μ¬μ μ λ§ κ°μ¬ν©λλ€. κ³§ λ€μ λ΅κ² μ΅λλ€. κ°μ¬ν©λλ€. |
Operator: Ladies and gentlemen, this concludes today's call. We thank you for your participation. You may now disconnect. | **Operator:** μ μ¬ μλ μ¬λ¬λΆ, μ€λ 컨νΌλ°μ€ μ½μ λ§μΉκ² μ΅λλ€. μ°Έμ¬ν΄ μ£Όμ μ κ°μ¬ν©λλ€. μ΄μ μ°κ²°μ μ’ λ£νμ λ λ©λλ€. |
Here's a summary of the key points from the earnings call in Korean:
β’ μ£Όμ μ€μ λ° μ λ§
- 3κ°μ 10% μ΄μ κ³ κ° ν보 (μ΅λ κ³ κ° 35%, 2μ 33%, 3μ 20%)
- 4λ²μ§Έ νμ΄νΌμ€μΌμΌλ¬ κ³ κ°μ΄ 2026 νκ³μ°λμ 10% κ³ κ°μΌλ‘ μ±μ₯ μμ
- νμ¬ μ΅λ κ³ κ°μ΄ 2026λ
μ±μ₯μ μ£Όλν κ²μΌλ‘ μ λ§
β’ μ¬μ
μ λ΅ λ° μμ₯ κΈ°ν
- AEC(Active Electrical Cable) μμ₯μ΄ λ λ΄λΆμμ λκ° μ°κ²°λ‘ νλ μ€
- μ λ’°μ±κ³Ό μ λ ₯ν¨μ¨μ± μΈ‘λ©΄μμ κ΄ν μ루μ
λλΉ κ²½μλ ₯ 보μ
- PCIe Gen 6 λ° 200Gbps νλ‘ν μ½ μμ₯μμ μ±μ₯ κΈ°ν νμΈ
β’ κΈ°μ λ° μ ν κ°λ°
- 3nm 곡μ κΈ°λ° 1.6T κ΄ν DSP κ°λ° μ§ν μ€
- μ λ ₯ ν¨μ¨μ± ν₯μμ μν΄ 5nmμμ 3nmλ‘ μ§μ μ ν κ²°μ
- λ€μν νλ‘ν μ½ μ§μ κ°λ₯ν μ ν ν¬νΈν΄λ¦¬μ€ ꡬμΆ
⒠리μ€ν¬ μμΈ
- κ³ κ°λ³ λΉμ νμ μ±μ₯ ν¨ν΄
- μ κ· κΈ°μ λμ
μκΈ°μ λΆνμ€μ±
- 곡κΈλ§ μ μ½ κ°λ₯μ±