การแถลงผลประกอบการไตรมาส 4 ปีงบประมาณ 2026 ของ Aeluma (ALMU): การดำเนินการเชิงพาณิชย์ AI DataCom และยอดเงินสดคงเหลือ 56 ล้านดอลลาร์

ประเด็นสำคัญ

• เอลูมา (Aeluma) รายงานรายได้ประจำไตรมาส 4 ปีงบประมาณ 2026 ที่ 582,000 ดอลลาร์ และรายได้ตลอดทั้งปีที่ 4.5 ล้านดอลลาร์ ซึ่งใกล้เคียงกับระดับสูงของช่วงที่เคยประเมินไว้ในการแถลงผลประกอบการไตรมาส 3 ทั้งนี้ รายได้ต่อปีค่อนข้างทรงตัวเมื่อเทียบกับ 4.7 ล้านดอลลาร์ในปีงบประมาณ 2025 และยังคงผูกอยู่กับสัญญารัฐบาลเป็นหลัก
• ผลขาดทุนสุทธิตามมาตรฐาน GAAP ในปีงบประมาณ 2026 ขยายตัวขึ้นเป็น 9.2 ล้านดอลลาร์ หรือ 0.52 ดอลลาร์ต่อหุ้น จาก 3.0 ล้านดอลลาร์ หรือ 0.23 ดอลลาร์ต่อหุ้นในปีงบประมาณ 2025 โดยฝ่ายบริหารระบุว่าสาเหตุหลักมาจากจำนวนพนักงานที่เพิ่มขึ้นและค่าใช้จ่ายในการดำเนินงานอื่น ๆ
• เงินสดและรายการเทียบเท่าเงินสด ณ สิ้นปีงบประมาณแตะที่ 56.0 ล้านดอลลาร์ เพิ่มขึ้นจาก 37.8 ล้านดอลลาร์ ณ สิ้นไตรมาสมีนาคม และ 15.7 ล้านดอลลาร์ในปีก่อนหน้า โดยเอลูมาไม่มีหนี้สิน
• AI DataCom เป็นเป้าหมายหลักในการดำเนินการเชิงพาณิชย์ในระยะสั้นของบริษัท โดยการพัฒนาเน้นไปที่เครื่องตรวจจับแสงความเร็วสูง LINX และเลเซอร์ควอนตัมดอท Quasar ที่ผลิตบนแผ่นฐานรองที่ไม่ใช่อินเดียมฟอสไฟด์
• สำหรับปีงบประมาณ 2027 เอลูมาคาดว่าจะรับรู้รายได้จากสัญญารัฐบาลที่บันทึกไว้แล้วประมาณ 2.3 ล้านดอลลาร์ และมีโอกาสทางธุรกิจอีก 2.0 ล้านดอลลาร์ที่อยู่ระหว่างการหารือ ทั้งนี้ บริษัทไม่ได้ให้ประมาณการรายได้อย่างเป็นทางการ
• ฝ่ายบริหารคาดการณ์รายจ่ายลงทุนในปีงบประมาณ 2027 ไว้ที่ประมาณ 10 ล้านดอลลาร์ถึง 12 ล้านดอลลาร์ โดยหลัก ๆ จะใช้สำหรับเตาปฏิกรณ์ MOCVD รุ่น AIXTRON G10 จำนวนสองเครื่อง รวมถึงศักยภาพในการผลิต การทดสอบ และการตรวจสอบความถูกต้องที่เกี่ยวข้อง

ผลประกอบการทางการเงินที่สำคัญ

ในระหว่างไตรมาส 4 เอลูมาได้ออกหุ้นจำนวน 830,484 หุ้นผ่านโครงการเสนอขายหุ้นตามราคาตลาด (At-The-Market Facility) ที่ราคาเฉลี่ย 24.87 ดอลลาร์ สร้างรายได้สุทธิ 20.1 ล้านดอลลาร์

ผลการดำเนินงานทางธุรกิจและการปฏิบัติการ

เอลูมากำลังทุ่มเททรัพยากรการพัฒนาเชิงพาณิชย์ไปที่การเชื่อมต่อศูนย์ข้อมูล AI ทั้งนี้ ฝ่ายบริหารได้อ้างอิงการคาดการณ์ของ Dell’Oro Group ที่ระบุว่ารายจ่ายลงทุนของศูนย์ข้อมูลอาจแตะระดับ 1.8 ล้านล้านดอลลาร์ในปี 2030 โดยประมาณ 10% ถึง 15% จะถูกจัดสรรให้กับเทคโนโลยีโฟโตนิกส์

เครื่องตรวจจับแสง LINX S-Series มุ่งเป้าไปที่อัตราการส่งข้อมูลตั้งแต่ไม่กี่กิกะบิตต่อวินาที ไปจนถึง 64 กิกะบิตต่อวินาที สำหรับการเชื่อมต่อ AI ระยะสั้นแบบ slow-and-wide โดยฝ่ายบริหารเชื่อว่าการใช้งานเหล่านี้อาจเป็นโอกาสในการผลิตปริมาณมากอย่างมีนัยสำคัญ เมื่อขยายการติดตั้งใช้งานตั้งแต่ปี 2028 ถึงปี 2030

LINX F-Series กำลังได้รับการพัฒนาสำหรับการเชื่อมต่อแบบขยายระบบ (Scale-out) สไตล์ fast-and-narrow โดยเริ่มต้นตั้งเป้าไว้ที่ 224 กิกะบิตต่อวินาทีต่อเลน และต่อมาจะเพิ่มเป็น 448 กิกะบิตต่อวินาทีต่อเลน ทั้งนี้ ฝ่ายบริหารกล่าวว่าการมีส่วนร่วมของลูกค้ามุ่งเน้นไปที่ทั้งการทดแทนชิ้นส่วนเดิมและการกำหนดคุณลักษณะเฉพาะสำหรับสถาปัตยกรรมใหม่

ตระกูลผลิตภัณฑ์ Quasar ใช้เทคโนโลยีเลเซอร์ควอนตัมดอท MOCVD โดยลูกค้ากำลังประเมินเลเซอร์สำหรับการทำงานที่อุณหภูมิสูง ความน่าเชื่อถือ และบรรจุภัณฑ์ที่ไม่ต้องใชุปกรณ์แยกสัญญาณ (Isolator-free) เอลูมาระบุว่าแนวทางของบริษัทมีเป้าหมายเพื่อรองรับการผลิตเชิงปริมาณมาก ขณะเดียวกันก็ลดการพึ่งพาแผ่นฐานรองอินเดียมฟอสไฟด์

บริษัทได้ลงนามในข้อตกลงร่วมกับ Sumitomo Chemical Advanced Technologies เพื่อเพิ่มกำลังการผลิตเวเฟอร์เอพิแท็กซี (Epitaxy wafer) โดยใช้อุปกรณ์ MOCVD ที่มีอยู่ นอกจากนี้ เอลูมากำลังจัดซื้อระบบ MOCVD รุ่น AIXTRON G10 อีกหลายเครื่อง และทำงานร่วมกับพันธมิตรโรงงานผลิตเซมิคอนดักเตอร์สารประกอบและซิลิคอนที่รองรับขนาดเวเฟอร์ตั้งแต่ 150 มิลลิเมตร ไปจนถึงใหญ่สุด 300 มิลลิเมตร

จำนวนพนักงานของเอลูมาเพิ่มขึ้นจาก 14 คน ณ วันที่ 30 มิถุนายน 2025 เป็นมากกว่า 30 คน ณ ช่วงเวลาการประชุม ฝ่ายบริหารมีแผนที่จะจ้างงานเพิ่มเติมเพื่อสนับสนุนการพัฒนาผลิตภัณฑ์ การดำเนินงานตามโครงการ และการขยายสู่เชิงพาณิชย์

แนวโน้มจากฝ่ายบริหาร

เอลูมาไม่ได้ออกประมาณการรายได้อย่างเป็นทางการสำหรับปีงบประมาณ 2027 เนื่องจากสัญญากหลายฉบับที่อยู่ระหว่างการรออนุมัติอาจส่งผลกระทบอย่างมีนัยสำคัญต่อผลลัพธ์

บริษัทคาดว่าจะรับรู้รายได้จากสัญญารัฐบาลที่บันทึกไว้แล้วในปัจจุบันประมาณ 2.3 ล้านดอลลาร์ในปีงบประมาณ 2027 นอกจากนี้ ยังอยู่ระหว่างการหารือเกี่ยวกับโอกาสเพิ่มเติมจากรัฐบาลมูลค่าสูงสุด 2.0 ล้านดอลลาร์ และข้อตกลงวิศวกรรมแบบไม่เกิดซ้ำ (Non-recurring engineering หรือ NRE) เชิงพาณิชย์มูลค่าหลายล้านดอลลาร์กับว่าที่ลูกค้า โดยการหารือเกี่ยวกับ NRE ในระยะสั้นส่วนใหญ่เกี่ยวข้องกับ AI DataCom แต่กรอบเวลาและการบรรลุข้อตกลงยังคงมีความไม่แน่นอน

เอลูมายังคงเจรจาเงื่อนไขขั้นสุดท้ายสำหรับเงินทุนที่มีศักยภาพสูงสุด 30 ล้านดอลลาร์จากสำนักงานวิจัยและพัฒนา CHIPS ของกระทรวงพาณิชย์สหรัฐฯ ทั้งนี้ เมื่อพิจารณาจากโครงสร้างปัจจุบัน ฝ่ายบริหารคาดว่าเงินสนับสนุนที่เสร็จสมบูรณ์ใดๆ น่าจะถูกบันทึกเป็นเงินลงทุนในส่วนของผู้ถือหุ้นจากรัฐบาลสหรัฐฯ มากกว่าที่จะเป็นรายได้

คาดว่ารายจ่ายลงทุนในปีงบประมาณ 2027 จะมีมูลค่ารวมประมาณ 10 ล้านดอลลาร์ถึง 12 ล้านดอลลาร์ โดยฝ่ายบริหารระบุว่าการจัดซื้อเตาปฏิกรณ์ MOCVD สองเครื่องที่วางแผนไว้นั้น ส่วนใหญ่เป็นการลงทุนเพียงครั้งเดียวที่มีวัตถุประสงค์เพื่อเร่งการพัฒนากระบวนการและการขยายการผลิตเชิงปริมาณ โดยไม่เปลี่ยนรูปแบบธุรกิจที่เน้นใช้สินทรัพย์น้อยของเอลูมา

ตามข้อมูลจากประธานเจ้าหน้าที่บริหารฝ่ายการเงิน (CFO) คาดว่าค่าใช้จ่ายในการดำเนินงานในไตรมาส 1 ปีงบประมาณจะไม่ปรับตัวเพิ่มขึ้นอย่างมีนัยสำคัญจากระดับที่สูงในไตรมาส 4 จากนั้นคาดว่าค่าใช้จ่ายจะค่อย ๆ เพิ่มขึ้นเมื่อบริษัทเพิ่มทรัพยากรด้านการวิจัยและพัฒนา (R&D) รวมถึงพนักงาน

ความเสี่ยงและประเด็นที่ต้องจับตา

• ความชัดเจนของรายได้ในปีงบประมาณ 2027 ยังมีจำกัด เนื่องจากโครงการเพิ่มเติมของรัฐบาลและข้อตกลง NRE เชิงพาณิชย์ยังคงอยู่ระหว่างการหารือ
• เงินทุน CHIPS ที่ได้รับการเสนอขึ้นอยู่กับการตรวจสอบสถานะและการเจรจาข้อตกลงขั้นสุดท้าย ซึ่งเงื่อนไขสุดท้ายและการปฏิบัติทางบัญชียังดำเนินไม่แล้วเสร็จ
• เอลูมากำลังพิจารณาคัดเลือกสัญญาวิจัยและพัฒนาของรัฐบาลอย่างพิถีพิถันยิ่งขึ้น ซึ่งอาจลดรายได้จากสัญญาในระยะสั้น ขณะที่ทรัพยากรถูกปรับเปลี่ยนไปสู่โครงการเชิงพาณิชย์
• การพัฒนาสู่เชิงพาณิชย์ขึ้นอยู่กับการรับรองคุณสมบัติโดยลูกค้า การดำเนินงานของพันธมิตรในห่วงโซ่อุปทาน และความสำเร็จในการขยายกำลังการผลิตเอพิแท็กซี การทดสอบ และการตรวจสอบความถูกต้อง
• ฝ่ายบริหารระบุว่าปัญหาการขาดแคลน ต้นทุนที่สูงขึ้น และข้อจำกัดในการขยายการผลิตของแผ่นฐานรองอินเดียมฟอสไฟด์เป็นข้อจำกัดในอุตสาหกรรม ดังนั้น กลยุทธ์ของเอลูมาจึงขึ้นอยู่กับการที่ลูกค้าปรับมาใช้แพลตฟอร์มแผ่นฐานรองทางเลือกของบริษัท
• จำนวนพนักงาน การใช้จ่ายด้าน R&D และการลงทุนด้านการผลิตที่สูงขึ้น ส่งผลให้ผลขาดทุนเพิ่มขึ้น และคาดว่าจะทำให้ค่าใช้จ่ายในการดำเนินงานยังคงอยู่ในแนวโน้มขาขึ้นหลังจากไตรมาส 1 ของปีงบประมาณ

ไฮไลต์การถาม-ตอบกับนักวิเคราะห์

ฝ่ายบริหารยืนยันว่า AI DataCom จะยังคงเป็นจุดเน้นหลักในการทำตลาดเชิงพาณิชย์ของเอลูมาไปจนถึงปีงบประมาณ 2027 แม้ว่าแพลตฟอร์มเทคโนโลยีดังกล่าวจะยังคงมีศักยภาพในการประยุกต์ใช้ในตลาดอื่น ๆ เช่น อุปกรณ์อิเล็กทรอนิกส์เคลื่อนที่ก็ตาม

เกี่ยวกับการเคลื่อนไหวของลูกค้า ฝ่ายบริหารระบุว่าการเจรจาสำหรับโครงการ NRE เชิงพาณิชย์เกือบทั้งหมดเกี่ยวข้องกับ AI DataCom โดยความสนใจดังกล่าวได้รับแรงหนุนจากข้อจำกัดด้านอุปทานในปัจจุบัน และความกังวลว่ากำลังการผลิตที่วางแผนไว้ในอุตสาหกรรมอาจยังไม่เพียงพอเมื่อความต้องการขยายตัวในช่วงปลายทศวรรษนี้

เอลูมากำลังจัดสรรทรัพยากรให้แก่ทั้งเครื่องตรวจจับแสงและเลเซอร์ควอนตัมดอท โดยไม่ได้ให้ความสำคัญกับเทคโนโลยีใดเทคโนโลยีหนึ่งก่อน ฝ่ายบริหารเห็นความต้องการเครื่องรับความเร็วสูง ชุดเครื่องตรวจจับแบบ slow-and-wide และเลเซอร์กำลังสูง ทั้งในอุปกรณ์รับส่งสัญญาณแบบเสียบถอดได้ (Pluggable transceivers), Near-packaged optics และ Co-packaged optics

ฝ่ายบริหารยังกล่าวอีกว่าการเปลี่ยนผ่านไปสู่ NPO และ CPO อาจสร้างโอกาสเนื่องจากสถาปัตยกรรมใหม่ ๆ อาจต้องการชิ้นส่วนที่ปรับแต่งเฉพาะ บรรจุภัณฑ์ขั้นสูง และการรวมในระดับเวเฟอร์กับ CMOS ที่ใกล้ชิดยิ่งขึ้น อย่างไรก็ตาม บริษัทคาดว่ารูปแบบ Pluggable, NPO และ CPO จะคงอยู่ร่วมกัน มากกว่าที่จะมีสถาปัตยกรรมใดสถาปัตยกรรมหนึ่งมาทดแทนสถาปัตยกรรมอื่นอย่างสมบูรณ์

บันทึกคำต่อคำฉบับเต็มของการแถลงผลประกอบการ

บทถอดเสียงฉบับเต็มของการประชุมทางโทรศัพท์ผลประกอบการ

คำชี้แจงจากฝ่ายบริหาร

Operator

Thank you. Good day, and thank you for standing by. Welcome to Aeluma's Fourth Quarter Fiscal 2026 Earnings Conference Call.

[Operator Instructions]

Please be advised that today's conference call is being recorded. At this time, I would like to turn the call over to Moira Conlon, Investor Relations for Aeluma. Please go ahead.

Moira Conlon

Good afternoon, and welcome to Aeluma's Fourth Quarter Fiscal 2026 Earnings Call. I am here today with Founder and CEO, Dr. Jonathan Klamkin, and CFO, Christopher Stewart. Discussions and responses to questions may include forward-looking statements, which are subject to various risks and uncertainties that could cause our actual results to differ materially from these statements. The risks and uncertainties are detailed in the earnings press release issued today, along with the reports filed with the U.S. Securities and Exchange Commission.

These reports, along with today's earnings release and fourth quarter presentation, which we will reference during this conference call, can be found under the investor section of our website. Aeluma assumes no obligation to update or revise any forward-looking statements to reflect events or circumstances that may arise after the date of this call. Throughout the discussion, the company will refer to non-GAAP financial measures, including EBITDA and adjusted EBITDA. A reconciliation of non-GAAP financial measures to the most directly comparable GAAP measures is included in our earnings press release and SEC filings. Now I will turn the call over to Aeluma's CEO, Jonathan Klamkin.

Jonathan Klamkin

Thank you, Moira, and thank you all for joining today's call. We are delighted to share a recap of our fiscal 2026 and to provide an outlook on Aeluma's path to commercialization.

The AI DataCom market opportunity has continued to become a priority for us. Demand for high-performance photonics is at an all-time high, with data center CapEx investments expected to reach $1.8 trillion in 2030, according to Dell'Oro Group, of which approximately 10% to 15% is for photonics. As illustrated in Slide 4, we explicitly built our technology platform for this scale. Our proprietary heterogeneous integration with large diameter non-indium phosphide substrates combines high performance materials with mass market microelectronics manufacturing.

To capitalize on the AI opportunity, in fiscal 2026, we prioritized resources to accelerate development and commercialization of our high-speed photodetectors and quantum dot lasers, which are applicable across a broad range of data center connectivity, including slow and wide short reach interconnects, fast and narrow interconnects, pluggable transceivers, near packaged optics or NPO, and co-packaged optics or CPO. Many traditional photonics components for data centers depend on indium phosphide substrates, which have become a bottleneck for suppliers and their customers. Aeluma's technologies are manufactured with non-indium phosphide substrates that may range in size from 150 millimeter or 6-inch to 200 millimeter and up to 300 millimeter.

Aeluma's photodetector and laser technologies have undergone significant foundational development, thanks in part to non-dilutive government funding. We have always been selective in bidding on government contracts that are synergistic with our commercial ambitions. An example is a Navy program that commenced in fiscal 2024 to develop high-speed photodetectors for short-reach links, a technology that is highly relevant to slow and wide data center interconnects for scale-up AI workloads. This Navy program has transitioned successfully into a second phase to focus on higher speed operation and transceiver integration.

We have also made considerable progress with customers evaluating our differentiated photonics platform, which eliminates the need for indium phosphide substrates, supports high-volume manufacturing, and provides a path to CMOS wafer-scale integration and packaging. These engagements have helped shape the development of our LINX S-Series photodetectors, which target data rates from a few gigabits per second to 64 gigabits per second. Based on customer discussions and market forecasts, we believe these applications represent a substantial volume opportunity as deployments scale in 2028 through 2030. To support fast and narrow 200 gigabits per second per lane and future 400 gigabits per second per lane for scale-out interconnects, we are developing our LINX F-Series photodetectors.

This represents another significant potential opportunity in the growing AI DataCom market. The non-dilutive government funding we secured played an important role in advancing our technology while also providing a potential long-term customer relationship with the U.S. Navy. As the AI DataCom market accelerates and demand grows over the next several years, we have prioritized our resources on the commercialization of our high-speed photodetectors to capture this opportunity.

Another example is funding from the [ Office of Secretary of War ] and from NASA to advance our MOCVD quantum dot laser technology. Today, our DataCom customers are evaluating our non-indium phosphide quantum dot lasers for high temperature operation, reliability, and isolator-free packaging. Our Quasar family of quantum dot lasers currently under development aims to address the demand for high-power lasers in scale-up and scale-out networks. We believe we are uniquely positioned as the first company to offer quantum dot lasers using MOCVD, an industry standard for high-throughput production. MOCVD, for example, is used exclusively for large volume VCSEL manufacturing for facial recognition in mobile phones.

Given the scale of investment and projects already underway for data center buildouts, we are more commercially focused than ever before. We stated that fiscal 2026 would establish the foundation for transitioning to commercialization. We delivered on this objective by expanding our team, establishing manufacturing supply chain partners, bolstering cash, and focusing on our go-to-market strategy. At June 30, 2025, Aeluma had 14 employees. As of today, we have more than 30 employees, and we plan to recruit additional talent to execute our strategic priorities.

In June, we announced the appointment of Dr. Brendan Moran as VP of Engineering, whose distinguished career at Lumileds included leading product strategy and development for mobile photonics components that generated $300 million in annual revenue, 8 consecutive design-ins for a high-profile mobile component, customer, and shipping of billions of photonics chips. Brendan is driving Aeluma's product development and commercialization efforts with an emphasis on photodetectors and lasers for AI DataCom. Along with Dr. Willy Rachmady, Aeluma's VP of Strategic Partnerships and Ecosystem, and myself, Brendan is helping advance our engagements with customers and define technology and product development roadmaps. We believe our product-focused strategy will benefit our business and our shareholders.

We also recently appointed Jason Taylor as Senior Director of Program and Project Management. Jason has nearly 3 decades of experience in program management at Intel, Kyocera, Lumileds, and other organizations. Jason is leading program management across Aeluma and is building out a program management system to ensure timely execution of priority programs, including high speed photodetectors and quantum dot lasers for AI DataCom. Such a system is key to delivering on our strategic priorities.

I am also proud to welcome Dr. Pramit Parikh as a strategic advisor. Pramit was VP and GM of the Gallium Nitride Business Division at Renesas following the acquisition of Transphorm Inc., premier in gallium nitride power semiconductors that Pramit co-founded and led from inception, its public company listing, and acquisition by Renesas. He also held leadership positions at Nitronex and Cree. His expertise spans several areas important to Aeluma, including high-performance semiconductors, AI data centers, manufacturing, IP strategy, and partnerships. This provides a strategic and operational perspective to Aeluma as we position the company for commercial growth. We are again adding more office and meeting space to support our growing team. And while we continue to be stewards of capital, the time for making investments in growth is now.

Moving through fiscal 2027 and beyond, we will be even more selective with government funding and more commercially focused. This may mean sacrificing near-term government contract revenue to focus resources toward achieving commercial revenue and growth opportunities, which is what matters in generating long-term shareholder value. In previous communications, we shared our intention to move away from early-stage government funding and instead focus on opportunities that accelerate manufacturing and commercialization. Consistent with that strategy, on July 29, we announced a letter of intent with the Department of Commerce, CHIPS R&D Office, for up to $30 million to accelerate development and commercialization of our scalable photonics for AI and advanced computing.

This is not for a specific [ Department of War ] or NASA or Department of Energy or other agency-driven application with technical metrics established by the government agency. For the Department of Commerce program, we proposed Aeluma's own vision. That is, to build the world's highest performance photonics with scalable manufacturing, leveraging domestic capabilities and advanced intellectual property to overcome supply chain constraints and to meet the demand for critically important AI and technology and advanced compute infrastructure investments. These themes have been central to Aeluma's strategy since our founding. The CHIPS initiative is aimed at supporting semiconductor innovation across integrated photonics, compute architectures, advanced packaging, substrates, materials, and memory for the AI and advanced compute supply chain. Strategic government investment could accelerate our development and commercialization efforts.

On scaling, I am happy to announce that we recently executed an important agreement with Sumitomo Chemical Advanced Technologies to strengthen our relationship, increase wafer production capacity, leveraging existing MOCVD tools that we can run our processes on, and provide a path for further increasing capacity in the future. We have been engaged with Sumitomo Chemical Advanced Technologies for several years, and this new agreement will support commercialization efforts and will complement Aeluma's in-house capability. The official focus of this effort is on high-speed photodetectors for AI DataCom. For future scaling, we are procuring multiple AIXTRON G10 MOCVD systems. These types of tools are already in use by major Tier 1 photonics component manufacturers for lasers and other photonics components for AI DataCom. Aeluma recently configured these tools to implement our proprietary non-indium phosphide substrate processes, and we are planning for installation.

While our precise strategy is confidential, we have communicated our multi-pronged approach to leverage different non-indium phosphide substrate types and sizes to pair technology with market demand in the most economical and strategic way. As we shared on our Q3 earnings call, we work with several supply chain partners. These include both compound semiconductor fabs and silicon fabs, whose capabilities range from 150 millimeter fabrication to 200 millimeter and some up to 300 millimeter. For many of our target markets, 150 millimeter wafers, either gallium arsenide or silicon, are appropriate. Aeluma's use of non-indium phosphide substrates at this size and partnership with Volume Microelectronics Foundries provides a path to scale and meet demand while overcoming supply chain constraints and winning on cost.

Aeluma produces starting wafers in-house, but also works with partners such as Sumitomo Chemical Advanced Technologies for scaling, and we have recently strengthened this relationship with a new agreement. Following epitaxy wafer production, our proprietary wafers are sent to foundries and other supply chain partners and then returned to Aeluma for test and validation. Like other photonics IDMs, or Integrated Device Manufacturers, Aeluma has developed several key compound semiconductor technologies to address different uses and functions, such as wavelength specifications and market applications. As illustrated in Slide 5, while our technology platform is broadly applicable, our current commercialization focus is AI DataCom.

Referring to Slide 6, we are leveraging our large diameter substrate platform to commercialize our LINX high speed photodetectors and our Quasar quantum dot laser technologies. LINX is designed to address both scale-up and scale-out AI interconnects. Aeluma's non-indium phosphide platform is attractive to customers who need to overcome supply chain constraints and enable scaling with a roadmap for wafer scale integration with CMOS, which is key for the slow and wide scale-up AI interconnects. And to cover the playing field, the LINX F-Series photodetectors are aimed at fast and narrow 224 and then 448 gigabits per second scale-out AI interconnects. Our MOCVD quantum dot technology is driving Quasar product development.

Current laser technology for AI interconnects is based on indium phosphide. And as we have continually messaged, indium phosphide substrates are in short supply, are small, expensive, and fragile. And the indium phosphide manufacturing doesn't scale to large volumes. Investment in 6-inch indium phosphide are being made, but this will take time to qualify. And costs of 6-inch indium phosphide substrates are increasing. Also, indications are that 6-inch indium phosphide is not sufficient for the AI infrastructure build-out. Our LINX effort is focused on 6-inch gallium arsenide manufacturing initially, but will also leverage larger diameter silicon manufacturing in the future to enable further scaling and direct integration with silicon photonics. Our MOCVD approach is key to supporting this roadmap.

Looking forward to fiscal 2027 and beyond, our enthusiasm continues to grow. AI DataCom opportunity is driving our near-term product-focused commercialization efforts. In addition to prospective capital from the CHIPS program that is under definitive agreement negotiations, we are negotiating several multimillion-dollar NRE agreements with commercial customers. NRE programs with top-tier manufacturers are an effective way to accelerate qualification and commercialization. We are ramping activities with our supply chain partners, expanding MOCVD manufacturing capabilities, and will continue to build out our team to support commercialization. It is highly motivating to see Aeluma's vision come to life as we execute our go-to-market strategy and the demand for high-performance photonics grows at unprecedented rates. Now I will turn the call over to our CFO, Chris Stewart, to discuss the financials.

Christopher Stewart

Thanks, Jonathan. Over the last year, the opportunity for Aeluma, particularly in the AI data center interconnect market, has really come into focus. Projections for the size of the opportunity continue to increase dramatically. We hear from customers and partners that our technology is well positioned to support an industry that is currently struggling to scale up to meet the demand. This year, we made significant progress towards commercialization, and I am confident in our team's ability to execute our strategy and capitalize on the opportunity.

Now I will review our fiscal fourth quarter and full year 2026 financial results. Revenue in Q4 was $582,000, resulting in total revenue of $4.5 million for the year, near the high end of the range we provided on our Q3 call. Revenue primarily reflects our government contracts, was approximately flat to the $4.7 million we reported for fiscal 2025. GAAP net loss for the fourth quarter is $4 million, or $0.22 per share. The full fiscal year net loss was $9.2 million, $0.52 per share, compared to a net loss of $3 million, $0.23 per share in fiscal '25. Excluding stock-based compensation, adjusted net loss for the fourth quarter, $2.7 million, or $0.15 per share. Adjusted net loss for the full fiscal year $4.6 million, or $0.26 per share, compared to an adjusted net loss of $116,000, or $0.01 per share, in fiscal 2025.

Adjusted EBITDA was negative $2.9 million for the quarter and negative $5.2 million for the year, compared to negative $911,000 last quarter, positive adjusted EBITDA of $186,000 fiscal 2025. The greater net loss and lower adjusted EBITDA are primarily driven by increased headcount and other operating expenses.

We closed the fourth quarter with a strong balance sheet, including $56 million in cash and cash equivalents, compared to $37.8 million at the close of the March quarter, $15.7 million at June 30, 2025. We continue to have no debt on our balance sheet. During the fourth quarter, we issued 830,484 shares under our ATM facility at an average price of $24.87, netting proceeds of $20.1 million. As you may recall, in March, we established a $50 million ATM facility to provide flexibility to raise capital believe it is in the best interest of our shareholders. We believe that maintaining a strong balance sheet is critical to executing against the significant growth opportunity we see ahead, reinforcing our credibility with customers and partners. At the same time, we remain committed to operating in a disciplined manner.

Now, I will provide some color on our outlook for fiscal 2027. Over the past several years, government R&D contracts have been a great source of non-dilutive funding for development and relationship building. While we will continue to execute on R&D programs that are aligned with our commercial direction, our focus in fiscal 2027 is on commercialization. We expect to recognize approximately $2.3 million currently booked government contract revenue fiscal 2027 with the potential for an additional $2 million of opportunities under discussion. As Jonathan mentioned, we are also in discussions with several prospective marquee customers regarding commercial NRE programs that may bring additional revenue in fiscal 2027. We are continuing to work through diligence and drafting of the definitive agreements related to our letter of intent with the Department of Commerce, CHIPS R&D Office, for up to $30 million in funding.

While the terms are not yet final, our current understanding is that if finalized, this award is likely to be accounted for as an equity investment in Aeluma by the U.S. government and not as revenue of our previous government R&D contracts. We are proud to have been selected for this high-profile program centered around the development and commercialization of our products for AI. We have always viewed the U.S. government as a strategic partner, and an investment in Aeluma would help accelerate our commercialization efforts.

With respect to manufacturing, we are focused on expanding epitaxy wafer production capacity, Sumitomo Chemical Advanced Technologies, qualifying our processes with additional supply chain partners, and expanding our capabilities through the purchase of the latest generation of MOCVD reactors, test and validation tools. As such, we expect to make meaningful investments in fiscal 2027, anticipating capital expenditures of approximately $10 to $12 million. These investments are not a departure from our capital light model, but will help position us to ramp our proprietary process with supply chain partners.

Finally, we will continue investing in the talent, systems, and processes needed to support customer and product development programs, and to scale the business, move to towards commercialization. We made investments in talent in fiscal 2026. Going forward, we will continue to build the organizational foundation required to execute our strategy. With our strong balance sheet and potential CHIPS funding, we believe we are well positioned to make prudent investments to capitalize the tremendous opportunity we see ahead to create long-term value for our shareholders. I will turn the call back over to Jonathan, his closing remarks, before we open the call to your questions.

Jonathan Klamkin

Thank you, Chris. To summarize, we are pleased with the progress made in fiscal 2026 to establish the foundation to transition to commercialization. We continue to execute our strategic priorities and are committed to our go-to-market strategy with a near-term focus on our LINX high-speed photodetectors and Quasar quantum dot laser technologies for the AI DataCom market. As always, I want to thank our incredible team for their hard work and dedication, and I want to thank all of you for your unwavering support and enthusiasm. You can now open the call for questions. Thank you.

Operator

[Operator Instructions]

The first question will come from [ Daniel Yermackin ] with [ Freedom Broker ]. Please go ahead.

ช่วงถาม-ตอบ

Unknown Analyst

Hey, guys, congrats with big steps to the commercialization. Just a quick question. Last quarter you mentioned, the shortfall of fiscal 2026 revenue was entirely tied based with the delayed government programs, so does it mean that the revenues going to carry on on fiscal 2027?

Christopher Stewart

Well, the revenue, as we talked about in the first remarks, we are not providing official guidance this year, mainly because there is a handful of contracts that could really change the range quite a bit. We continue to deliver on the projects that we signed last year, and there are some programs that are continuing on from prior years. As we said, that book revenue, including the new contracts that we signed this year, is about $2.3 million. And then clearly there is a number of things that could bring upside from there, but at this point, you know, we are prepared to give more visibility than that.

Unknown Analyst

Okay, thanks. Comparing 2 end markets, the AI and data center with the mobile and electronics market, where is your near-term focus now? Is it on photodetectors of data centers or do you focus more on mobile and market?

Jonathan Klamkin

We remain active in other markets and, for example, mobile and have some indications of the size of opportunities and timeline of those opportunities. But mobile, for example, is a market with a more concentrated customer base. AI DataCom as a market is really at the forefront for us right now and for the entire industry. So this is the focus of our near-term commercialization efforts. But I think we should always remember that given the nature of our technology, anything we do to advance commercialization for 1 market will benefit others. It is just that today the focus, today and into fiscal 2027 is AI DataCom.

Unknown Analyst

Okay, thanks. That makes sense. Yeah, that's it from my side. Thank you.

Operator

The next question will come from Richard Shannon with Craig-Hallum. Please go ahead.

Richard Shannon

Well, great. Thanks, Jonathan and Christopher. Let me ask a couple questions here. Jonathan, I wanted to ask about your commentary about in some stage of negotiation with multiple contracts with NRE. I wanted to get a little bit more detail here. If I caught the language right now, I probably missed some of it, but I want to be clear here. This sounds like this is all entirely within the DataCom market in some manner. Is that accurate or are there other engagements in other market areas as well?

Jonathan Klamkin

I would say it is almost entirely in the AI data market in the near term. That is correct.

Richard Shannon

Okay. Kind of what I thought, but I just want to make sure. So can you maybe talk about how long these discussions have been in place here and any sense of urgency you are seeing? You know, the themes that you have been talking about for last conference calls really revolve around supporting supply chain issues and we have seen other companies in the broader space that we cover pretty well. That suggests there is that their urgency is only increasing here. So I would imagine this is getting pretty, you know, accelerating and going pretty quickly and a lot of interest here. So maybe you can help us understand, you know, the breadth and depth and the stage of some of these discussions, if you would, please.

Jonathan Klamkin

Yes, I would say in the AI DataCom market, that is correct. A lot of this is centered around supply chain constraints and how large the market opportunity is. And it seems like every time, forecasts are put out, they revisit and increase the forecasts, especially in that sort of 2028 through 2030 and even beyond timeframe. And what we are seeing is that our customers and some of these discussions we are having to initiate NRE programs come from the fact that there are investments being made to sort of fill the gap in the but investments being made are not going to be sufficient for that market a few years out.

So without sharing significant detail, you can imagine what this relates to. I spoke about sort of different substrate sizes that we are leveraging now, but how Aeluma's technology could take us to other substrate sizes in the future. So there is interest in our technology to fill the gap now, but also to fill the gap that is expected for years to come, just because of how significant this build-out is.

Richard Shannon

Okay, that is helpful. Looking at the DataCom opportunity or another angle here, your prepared remarks and your slides here, you talk about both photodetectors and quantum dot lasers here. I want to get a sense of the balance between those 2, even in terms of, you know, the number of engagements focused on one or the other and the opportunity sitting in front of you, wise over time. If you could just kind of qualitatively describe that, that would be great, Jonathan.

Jonathan Klamkin

Yes, I mean, they are both significant. I cannot quantify precisely on the call, but they are both very significant. And I think we have shared information in the past that typically people think about the lasers, everyone is hearing about the lack of indium phosphide substrates and the need for more and more lasers in the AI DataCom market. And there is plenty of discussions about the type of modulator that might be selected or types of modulators that might be selected as the industry starts to roll out 400G per lane beyond the 100 and 200G per lane, but sort of the forgotten technology is the photodetector and there is 2 angles there.

Photodetectors need to keep up with those modulation speeds, the full photodetector receiver, and there is sort of this emerging market for slow and wide transceiver components. And those require arrays of emitters, micro LEDs, micro pixels, and then arrays of photodetectors that might have different specifications or characteristics than photodetectors that you can buy on the shelf. And that is what we refer to in the slide, the LINX S-Series for sort of the slow and wide and the LINX F-Series for fast and narrow. This market seems to be very significant in size in the coming years. And the laser market, I think most of us know, is so significant.

Some of the suppliers of lasers or AI DataCom would be selling many more lasers than they are selling today if they had more indium phosphide substrates and if they had more FAB capacity. So I will just close this thought on, you know, we are still a relatively small company, but these 2 technologies are so important that we are not going to choose to develop just 1 of them. They are so important, and we are putting resources to both. So I would not say one outweighs the other, at least not today. Okay.

Richard Shannon

Okay, that is great characterization. My last question, I will jump back into the queue here, is probably more for Chris here, but just want to get a sense from 2 financial statement line items here, OpEx and CapEx here. So on OpEx here, notably above what we had estimated, but clearly you have been hiring here and a lot of activity going on. Maybe if you could just give us a little bit just characterize how we should expect to see that OpEx in the first quarter this year and then going forward then the CapEx here. Is this kind of a one-time, this $10 to $12 million, is that kind of a one-time dynamic? Or could we see subsequent years of, you know, similar level of CapEx? Thank you.

Christopher Stewart

Yep, thanks, Richard. Yes, so a couple things on that. We have been growing headcount through the year. But if you look, really, OpEx did not increase all that much from Q1 to Q3. And then in Q4, the combination of full-quarter full quarter salaries for folks that we were hiring through Q3 and early Q4, as well as some kind of normal kind of year-end accruals and things caused a little bit of an unusual bump in Q4. So I would not expect a big increase in Q1, if any. And then going forward, like we said, we are going to continue to invest both in the R&D side and in headcount, so it will drift up from there.

The big CapEx this year is the 2 MOCVD reactors that we talked about in the call, and, you know, I do not really want to guide CapEx beyond that, but like I said, it is not a change to our capital light model. So, yes, we are increasing these, buying these 2 tools really to drive and facilitate the transition to scale and help our supply chain partners scale faster and give us a little bit more involvement in the process development. So you have to think about those 2 tools in particular, more of a one-time thing this year and then then as we go forward, we will talk more about the CapEx later.

Richard Shannon

Thanks for all that clarity, guys. Thank you very much.

Operator

The next question will come from Suji De Silva with Roth Capital. Please go ahead.

Suji De Silva

Hi, Jonathan. Hi, Chris. Congratulations on the progress here. Thank you. Sounds good. Thanks. Yes, no problem. So the CHIPS funding. The $30 million that would phase in. I am wondering, is there any specific plan for that, or would that just be in the general pool of your spending? And, did I hear you right, Chris, because the accounting, is there an equity component the government will hold in Aeluma, or is that now how it will be transacted?

Christopher Stewart

So, again, you know, it is based on the LOI and what we are seeing from the other CHIPS Act deals. It is likely, and again, until everything is finalized, we are not sure. The main point there was it is not likely to be counted as revenue. It will look like an equity investment from the U.S. government. The key thing, you know, and I think the important point is, it is capital that will allow us to accelerate our commercialization efforts. It does support the, the high-speed photodetectors primarily, as Jonathan talked about. And so, you know, you could think of it in a way similar to the government program revenue we have had in the past, but it does not show up on the revenue line. It is still capital, advancing our commercialization efforts on AI data centers.

Suji De Silva

Okay, great. And then perhaps for Jonathan, I know there is just a supply-demand imbalance for lasers and so forth, but does the progression from traditional pluggables to NPO, CPO, does that create more demand for you? Does that have some specific opportunities for you to take advantage of and just can you characterize kind of the move to NPO in the marketplace and whether that is part of the conversations here.

Jonathan Klamkin

I would say yes, but across the board there is, you know, there is demand for pluggables, demand to transition to NPO and in some respects CPO, although I that will slowly roll out and it is not anticipated to sort of take over. These will all sort of sort of operate in unison inside data centers. And obviously the mix of data centers transceiver architectures that are used do depend on specific AI workloads. So this might be, data center or AI customer dependent. To your question, the transition creates additional opportunities for a number of reasons. One, when new formats or architectures are being adopted, that is usually a great time to intersect the market because there might be different needs.

They may not be using off-the-shelf components anymore and some customizations required. And so that is a great opportunity to adopt new technology that does not necessarily exist on the shelf. And we have opportunities in front of us to not only build components with our non-indium phosphide substrate technology to displace what is already in the market, in other words, meet the specifications of data sheets that already exist, but use our same platform to develop new products where data sheets do not necessarily exist. And we are working very closely with our customers to define those data sheets. We have our own internal spec sheets, but they are very customer driven when we build them.

So with photodetectors, there is sort of both opportunities. Building the real high speed photodetectors that are increasingly in high demand as the industry starts to adopt things like 400G per lane transceivers for 800G and 1.8. But then this slow and wide opportunity that does require some customer arrays of emitters and detectors on the receiver end and ideally direct CMOS integration. That is a very important aspect of our technology that our customers are attracted to that not only can we manufacture on bigger substrates that happen to be more available, we are on substrate sizes that are amenable to wafer scale integration and packaging. And so the way we are packaging, some of these detectors for these applications, especially the slow and wide.

It is not sort of the traditional chip-on-carrier or chip-on-submountain wire bond. It leverages more advanced packaging that you do not necessarily have access to when you are on the smaller substrates, like small indium phosphide substrates. And then to your question on lasers, the demand is very high across the board, and certainly for some of the folks that have been developing and are pushing to roll out CPO, they do need these very high power lasers. And there is opportunities to provide scale for those high power lasers that are going to be fed externally into, say, silicon photonic CPO transceivers, and there is some interest in maybe bringing the lasers into the package because there are new cooling technologies being adopted.

And if you bring the laser in the package, you might overcome some of the losses. So you may be able to operate the lasers at lower power and overcome some of the reliability challenges. So I would say for lasers, the demand is just, so high reliability is sort of a top priority for customers, as is sort of simplifying the supply chain, because it is more than just substrates and lasers. It is other aspects like the packaging. And the quantum dot lasers have the potential to eliminate the optical isolator in the package. So for a number of reasons, our customers are interested not only in the scale, of non-indium phosphide, they are very interested in the fact that quantum dot lasers offer these other attributes and that we build quantum dot lasers with MOCVD technology, which scales.

Suji De Silva

Thanks a lot for the question. No, Jonathan, it sounds like interesting conversations. Thanks for the color.

Operator

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

Jonathan Klamkin

Thank you for joining our call today. We look forward to connecting in the future. Have a great day.

Operator

The conference is now concluded. Thank you for participating. You may now disconnect.