Samsung Reportedly Aims for Silicio Photonics Mass Production in 2028

A recent report indicates that Samsung Electronics has set an ambitious goal: to begin mass production of silicio photonics components by 2028. This move, if confirmed, would position the Korean company at the forefront of developing a technology considered crucial for the future of data centers and, in particular, for the evolution of computationally intensive workloads such as those related to artificial intelligence and Large Language Models (LLMs).

Silicio photonics represents a paradigm shift from traditional electrical interconnects. Instead of transmitting data via electrical pulses over copper wires, this technology uses light (photons) to carry information through silicio waveguides. The result is higher transmission speed, significantly greater bandwidth, and substantially lower power consumption—factors that become increasingly critical as the demand for data processing grows exponentially.

The Role of Silicio Photonics in the AI Era

For AI and LLM workloads, the ability to move large volumes of data between GPUs, CPUs, and memory with minimal latency is fundamental. Current architectures, which rely on arrays of distributed accelerators, are often limited by the speed and bandwidth of electrical interconnects. Silicio photonics offers a solution to this bottleneck, enabling the creation of ultra-high-speed networks within servers and between nodes in a cluster.

This translates into greater efficiency for training complex models and for large-scale Inference. A system with more performant optical interconnects can handle larger batch sizes, reduce communication times between processing cores, and ultimately accelerate overall Throughput. For companies developing and deploying LLMs, this means being able to iterate on models faster, handle a greater number of requests, or support wider context windows with more agility.

Implications for On-Premise Infrastructure

The adoption of silicio photonics could have a profound impact on Deployment decisions, especially for Self-hosted and on-premise infrastructures. The ability to build AI clusters with faster and more efficient internal interconnects makes local Deployment even more competitive compared to cloud solutions, particularly for workloads requiring high data sovereignty or Air-gapped environments.

An on-premise infrastructure enhanced by silicio photonics could offer a more advantageous TCO in the long term, reducing operational costs related to energy consumption and improving overall data center efficiency. For CTOs, DevOps leads, and infrastructure architects evaluating Self-hosted vs. cloud alternatives for AI/LLM workloads, this technology represents an enabler for building high-performance local systems. AI-RADAR, through its analytical frameworks on /llm-onpremise, offers tools to evaluate the trade-offs between these different Deployment strategies.

Future Prospects and Adoption Challenges

Samsung's 2028 target underscores the growing importance of silicio photonics in the technology industry. However, mass production of this technology presents significant challenges, including the complexity of manufacturing processes, initial costs, and the need to integrate these new components with existing infrastructure. The industry is exploring various avenues to overcome these obstacles, with numerous players investing in research and development.

If these challenges are overcome, silicio photonics could become a standard for high-speed interconnects, not only in data centers but also in other applications requiring high bandwidth, such as high-performance networking and distributed computing systems. 2028 could mark a turning point for large-scale adoption, paving the way for a new generation of more powerful and efficient AI infrastructures.