News: LEDs
23 July 2026
Palomino ships first-generation micro-LED light engine prototypes to first hyperscaler customer
Artificial intelligence (AI) interconnect technology company Palomino Laboratories Inc of Goleta, CA, USA has shipped its first micro-LED light engine prototypes to its first hyperscaler customer. Furthermore, the customer completed its evaluation and provided Palomino with test measurement results confirming that the measured micro-LED bandwidth exceeded Palomino’s initial design targets.
The firm is investing in R&D to improve bandwidth, efficiency and power performance of its gallium nitride (GaN) micro-LED light engine technology. Palamino is on-track to sample its second-generation micro-LED technology with improved bandwidth performance to multiple customers by the end of fiscal third-quarter 2026.
With improved prototype bandwidth and power performance, Palamino plans to pursue one or more development contracts to advance its micro-LEDs and align performance with its customer’s initial AI interconnect requirements. It hence expects to receive its first development contract from a strategic AI customer in fiscal fourth-quarter 2026.
“These first samples, confirmed by a leading hyperscaler customer, exceeded our design targets and validated that our GaN micro-LED technology is progressing,” says co-founder Dr Steven Denbaars, a pioneer of GaN micro-LEDs. “Now we are accelerating our next generation technology on an aggressive schedule to drive a meaningful improvement in system performance as we move toward integration within a full optical interconnect solution,” he adds. “With our recently announced LOI [letter of intent] to acquire Vega Links Inc, we expect to migrate from delivering breakthrough micro-LED devices toward delivering differentiated optical links for AI applications.”
Palomino believes this work targets an increasingly urgent constraint in modern AI interconnects, namely bandwidth and efficiency, which is widely viewed as a system bottleneck as compute performance scales more quickly than interconnect bandwidth, according to a 2024 IEEE Micro Journal publication from University of California Berkley.
By leveraging differentiated micro-LEDs, micro-VCSELs (vertical-cavity surface-emitting lasers) and copper engine technologies, Palomino is focused on enabling emerging short-reach interconnect applications and delivering faster, more efficient data movement for AI data centers. According to 2025 and 2026 published market reports from LightCounting and company estimates, the total addressable market for the Palamino’s AI interconnect solutions is more than $60bn in 2031.
Palomino acquires AI interconnect firm Vega Links
On 16 July, Palomino announced that it has signed a binding letter of intent (LOI) to acquire Vega Links Inc. in an all-stock transaction, subject to customary closing conditions.
AI infrastructure is undergoing one of the most significant architectural transitions in decades, notes Palomino. The rapid growth of trillion-parameter AI models, accelerated computing and hyperscale AI factories is driving unprecedented demand for higher bandwidth, lower latency and dramatically improved power efficiency throughout the networking stack. Industry initiatives such as Ultra Ethernet, UALink, scale-up and scale-in architectures, co-packaged optics (CPO), near-packaged optics (NPO), and next-generation optical I/O are re-shaping how GPUs, CPUs, memory and networking systems communicate.
Palomino believes that the combination with Vega Links positions it to participate in this technology transition by offering complementary interconnect technologies across copper, micro-VCSEL and micro-LED-based optical interconnects. Rather than competing in a single connectivity technology, the combined company intends to address multiple layers of the AI interconnect ecosystem as customers optimize cost, power consumption, reach, bandwidth and latency.
Transaction highlights are cited as:
- transforming Palomino into a comprehensive AI interconnect company;
- expanding Palomino’s estimated addressable market by about 10x — from about $6bn to more than $60bn;
- adding AI systems architecture, networking and product development skillsets to Palomino’s engineering team;
- positioning Palomino to participate in emerging AI networking trends including UALink, scale-up and scale-in architectures, optical I/O, CPO and NPO, and
- following the closing of this transaction, Palomino expects to appoint Karthik Gopalakrishnan as chief technology officer (CTO) and Rajesh Radhamohan as chief product officer (CPO).
Picture: Co-founder & CEO Jeff Shealy (left) welcomes Vega Links CEO Karthik Gopalakrishnan (right) to Palomino Laboratories for the acquisition signing ceremony.
“AI infrastructure is evolving at an extraordinary pace, and interconnect technology has become one of the defining challenges for next-generation computing,” notes Palomino’s co-founder & CEO Jeffrey B. Shealy. “Customers are seeking the optimal combination of bandwidth, latency, power efficiency and cost across increasingly complex AI fabrics. We believe combining Vega Links’ systems architecture expertise with Palomino’s optical innovations creates a differentiated AI interconnect platform capable of addressing customer challenges across multiple technologies. Our vision is to build a leading AI infrastructure company that enables the next generation of hyperscale AI deployments while advancing our long-term objective of qualifying for a Nasdaq listing,” he adds.
“The future of AI networking will require close integration of system architecture, silicon and advanced interconnect technologies, such as GaN micro-LEDs,” says Karthik Gopalakrishnan. “By joining forces with Palomino, we believe we can accelerate innovation and deliver compelling solutions for hyperscale cloud providers, AI infrastructure companies and enterprise customers.”
Industry opportunity
Palamino believes that future AI clusters will rely on heterogeneous interconnect technologies rather than a single approach. Copper is expected to remain important for very-short-reach applications, while optical technologies, including micro-VCSEL and micro-LED-based solutions, are expected to play an increasingly important role as AI clusters continue to scale. Palamino believes that this broader technology portfolio positions it to pursue opportunities across multiple segments of the rapidly expanding AI networking market.
