News
The Center for Strategic and International Studies (CSIS) in the United States published a new article on the January 12th, 2024, suggesting that the new battleground in the US-China tech war could be silicon photonics technology. This technology aims to enhance transmission efficiency, reduce latency, and reshape the competition landscape between the US and China in semiconductors and AI.
According to TechNews’ report citing the author Matthew Reynolds’ notes in the article, unlike electronics, photonics uses photons instead of electrons to transmit information. When combined with electronic technology, photonics has the potential to create large-scale computing systems with higher bandwidth and energy efficiency, surpassing the physical limitations of traditional electronic chips.
However, the Chinese government has recently shown interest in photonics, seeing it as one way to bypass Western technological controls. Photonics technology is mentioned in China’s Outline of the 14th Five-Year Plan (2021-2025) for National Economic and Social Development and Vision 2035.
Yao Yang, the director of the National Development Institute at Beijing University, believes that US semiconductor restrictions are a “shooting themselves in the foot” because photonic chips will eventually make electronic chips obsolete.
He also sees this as an opportunity for China to overtake, asserting that China has the capability to take the lead in this emerging technology, as mentioned in his recent article.
However, Matthew Reynolds believes that it’s unlikely for photon chips to replace electronic chips, at least not in the near future. Photonics and electronics are more likely to coexist, forming a symbiotic relationship.
What is certain, though, is that silicon photonics technology holds the potential to become a breakthrough for China in advancing to the forefront of semiconductor manufacturing.
Reportedly, the most direct application of silicon photonics technology is in optical interconnects, replacing the copper wiring in circuits with photonics to speed the transmission of information between processors and/or memory, reducing the input/output bottlenecks currently plaguing AI computing.
In addition to optical interconnects, another application area for silicon photonics is in the emerging field of optical computing. Photon processors utilize light instead of electrons for computation. While their range of computational types is limited, they show significant promise in performing matrix multiplication operations, a crucial component, especially in large-scale language models, constituting over 90% of inference computations.
Chinese economist Chen Wenling from the China Center for International Economic Exchanges (CCIEE) stated in an article addressing the anti-American blockade that silicon photonics is the technology that China can use to overtake.
“China is preparing to build a photonic chip production line, which is expected to be completed in 2023, which means that China will be at the forefront of the world in terms of photonic chips, and even completely change the chip technology route. Photonic chips have many technical advantages. Its calculation speed is faster and its information capacity is larger, which will be more than 1,000 times higher than the current silicon-based chips.” Chen expressed.
Lightelligence, a U.S.-based optical computing company, previously received funding from the Chinese government and has recently launched the AI accelerator “Hummingbird.” Hummingbird utilizes optical interconnect components, connecting to chips manufactured by TSMC using 28-nanometer process.
Although this process may not be at the forefront of current technology, it aligns with China’s semiconductor manufacturing capabilities. Lightelligence even claims that its latency and efficiency metrics surpass those of competitors in certain AI tasks.
Additionally, Lightelligence has introduced the “Photonic Arithmetic Computing Engine” (PACE), an optical computing system. PACE integrates photonic and electronic components on a single chip and, in certain compute-intensive applications, boasts processing speeds 25-100 times faster than Nvidia’s high-end GPUs.
China’s SinTone Microelectronics is in the process of establishing a silicon photonics chip production line. Sui Jun, the president of SinTone Microelectronics, indicated that China has the capability to produce photon chips domestically because the manufacturing process does not require the use of extreme ultraviolet (EUV) lithography machines, which are subject to U.S. sanctions.
Simultaneously, a research team at Tsinghua University in China announced a breakthrough in overcoming the traditional physical limitations of chips, presenting a new computational framework that integrates optics and electronics. They successfully developed the world’s first all-simulated optoelectronic intelligent computing chip (ACCEL).
In terms of computational power for smart visual target recognition tasks, ACCEL exceeds current high-performance commercial chips by over 3,000 times. In the realms of smart visual target recognition tasks and computations for unmanned system scenarios, its energy efficiency surpasses existing high-performance chips by more than 4 million times.
While the commercialization timeline for ACCEL remains uncertain, researchers believe it holds the potential for applications in unmanned systems, industrial inspection, and AI large-scale models in the future.
Silicon Photonics Poised to Transform the US-China Tech War and AI Landscape
Matthew Reynolds believes that silicon photonics is the foundation and driving force behind advancements in optical interconnects and optical computing, reshaping the competitive landscape in the semiconductor and AI industries between the US and China.
While US export measures aim to sever China’s capabilities in advanced chip manufacturing, silicon photonics appears to be a new opportunity for China to take a different path.
However, Matthew Reynolds notes that despite the promotion of photon processor performance, its current applicability remains relatively narrow, contrasting sharply with the universality of electronic processors.
Additionally, the application of silicon photonics technology still faces numerous technical challenges, requiring software development in operating systems and applications to enhance performance in optical computing.
Therefore, achieving optical computing may still require several years, or even decades. Given the current pace of AI development, any delays could have serious consequences. Leading semiconductor companies in the United States and allied nations are also investing heavily in silicon photonics. It remains uncertain whether China can secure a leadership position.
Matthew Reynolds points out that regardless, new technologies and architectures are likely to redefine the components of advanced chips. They may weaken the impact of existing control measures or reshape the competitive landscape.
The US export controls may inadvertently stimulate China to allocate more resources to emerging technologies, positioning itself as a key player in the next generation of semiconductors, especially as Moore’s Law approaches its limits and demand for AI computing continues to grow.
Read more
(Photo credit: iStock)
News
According to a report from Taiwan’s Commercial Times, industry experts assert that Silicon Photonics (SiPh) is poised to revolutionize the cloud industry as communication transmission speeds surge beyond 1.6 Tbps. Utilizing Co-Packaged Optics (CPO) for integration, SiPh combines optical components and Application-Specific Integrated Circuit (ASIC) technology into a singular module, effectively mitigating power consumption challenges.
Moreover, the versatility of SiPh is highlighted by its applications in communication transmission, biomedical sensing, LiDAR, high-speed AI transmission, smart healthcare, and autonomous vehicles, showcasing significant potential. This expansive range of potential applications underscores the promising future of SiPh technology.
Major players in the semiconductor landscape, including TSMC, ASE, SunSin, and Accton, express bullish sentiments towards SiPh and CPO technologies.
However, current hurdles such as chip yields and standardization remain, awaiting resolution. The anticipated timeline for tangible contributions is expected to extend beyond 2025.
In the realm of photonic integration, TSMC takes the lead among Taiwanese manufacturers. The company’s Compact Universal Photonics Engine (COUPE) offers heterogeneous integration of Photonic ICs (PICs) and Electronic ICs (EICs), resulting in a 40% reduction in energy consumption and a considerable increase in customer adoption likelihood.
TSMC has reportedly invested in a 200-person R&D team, collaborating with international clients for joint development.
ASE is actively involved in the research and development of SiPh and CPO packaging technology. Leveraging the VIPack advanced packaging platform, the market anticipates a gradual uptick in related businesses in the latter half of 2024, with significant order momentum expected to surge in 2025.
Networking company Accton is channeling efforts into the photonic integration of various components for switches. On the other hand, SunSin, a System-in-Package (SiP) testing and packaging facility, is strategically positioning itself in CPO process technology.
Explore more:
(Image: ASE)
News
The Silicon Photonics topic is heating up as major companies race to address the data transfer speed between chips. Intel’s Silicon Photonics project has a leading advantage, while TSMC is collaborating with major customers Nvidia and Broadcom, investing 200 research and development personnel. They aim to complete the project in the second half of 2024, with production set to begin in 2025.
According to Taiwan’s Commercial Times, Luo Huaijia, the Executive Director of the Photonics Industry and Technology Development Association (PIDA) in Taiwan, stated that silicon photonics technology has always been a crucial focus in the field of photonics. Photonics products are evolving towards being compact, lightweight, energy-efficient, and power-saving.
Among Taiwan’s semiconductor fabs, TSMC stands out with its COUPE, which provides heterogeneous integration of photonic integrated circuits (PIC) and electronic integrated circuits (EIC), reducing energy consumption by 40%. TSMC is rumored to deploy a 200-person R&D team, collaborating with international major clients for joint development. Consequently, following the completion of its Hsinchu plant, TSMC invested NT$90 billion in constructing a new packaging plant in Tongluo, Miaoli, recognizing the significant demand and potential in heterogeneous integration.
Luo Huaijia pointed out that silicon photonics uses semiconductor technology to create a platform with optical properties, with the goal of integrating light and telecommunications signals. This involves packaging traditional optical components, including optical waveguides, light-emitting elements, and transceiver modules, together, thus also involving heterogeneous packaging.
As early as 2002, Intel publicly conducted research in the field of “Silicon Photonics,” but at that time, the data volume could be handled with copper wire transmission. Luo Huaijia believes that with the exponential increase in AI computing power, data processing will start in the gigabyte range, prompting companies to invest heavily in development.
Luo Huaijia analyzed that currently, GlobalFoundries is likely the first company to provide wafer foundry services for manufacturing optical fiber transceivers, using FD-SOI technology integration solutions. Intel also currently offers a 400Gb/s optical fiber transceiver solution. In addition to their own ASICs or FPGAs, this technology is applied to Switch ICs. Intel even plans to expand its silicon photonics solution into the automotive market, using it in Mobileye’s optical radar by 2025.
Explore More
(Photo credit: ITRI)
Insights
In recent years, with the rise of AI and 5G technologies leading to increasing computational demands, Silicon Photonics technology has once again become a focal point of discussion in the semiconductor industry.
TrendForce Perspective:
Since the development of the semiconductor industry, the industry’s trajectory has largely followed the development predicted by Gordon Moore – roughly doubling the number of transistors that can be accommodated on an integrated circuit approximately every two years. However, as chip sizes continue to shrink, chip architecture design is gradually being challenged. Semiconductor manufacturers, including TSMC, Samsung, and Intel, are striving to break through Moore’s Law as their goal. Others have publicly announced their focus on mature processes (the industry divides at 7nm, with 7nm and below considered advanced processes) and optimization of existing technologies.
However, even as manufacturers push the boundaries of Moore’s Law, leading to increased transistor density per unit area, signal loss issues inevitably arise during signal transmission since chips rely on electricity to transmit signals. Despite the increased transistor count, power consumption problems persist. Silicon Photonics technology, which replaces electrical signals with optical signals for high-speed data transmission, successfully overcomes this challenge, achieving higher bandwidth and faster data processing. With this approach, chips do not need to cram more transistors per unit area or pursue smaller nanometers and nodes. Instead, they can achieve higher integration and performance on existing processes, further advancing technology.
Currently, Silicon Photonics technology still faces various challenges, including alignment and coupling, thermal management, modulation and detection, expansion and integration, among others. Significant breakthroughs are unlikely in the short term, and major global manufacturers are still in the early development stages. In Taiwan, recent reports suggest that TSMC is actively venturing into Silicon Photonics technology. While TSMC has not officially confirmed this news, during the Silicon Photonics International Forum, a senior vice president from TSMC clearly stated, “If a good Silicon Photonics integration system can be provided, it can address the key issues of energy efficiency and AI computing power. This could be a Paradigm Shift, and we might be at the beginning of a new era.”
This suggests that TSMC is optimistic about the development of Silicon Photonics technology. Although Taiwanese companies have not formally announced their entry into the Silicon Photonics field, it is expected that with the explosive growth in demand for data transmission, storage, and computing driven by AI technology, Silicon Photonics will undoubtedly be a critical technology for future semiconductor development.
News
With the increasing demand for massive computing in fields such as AI, communication, and autonomous vehicles, the evolution of integrated circuits (ICs) has reached a physical limit under the premise of Moore’s Law. How can this limit be surpassed? The answer lies in the realm of optics. Currently, many domestic and international companies are actively embracing “Silicon Photonics” technology. When electronics meet photons, it not only addresses the signal transmission loss issue but is also considered a key technology that could usher in a new era, potentially revolutionizing the future world.
Integrated circuits (ICs) cram millions of transistors onto a single chip, performing various complex calculations. Silicon Photonics, on the other hand, represents integrated “light” paths, where light-conductive pathways are consolidated. In simple terms, it is a technology that converts “electronic signals” into “optical signals” on a silicon platform, facilitating the transmission of both electrical and optical signals.
As technology rapidly advances and computer processing speeds increase, communication between chips has become a critical factor in computing performance. For instance, when ChatGPT was first launched, there were issues with lag and interruptions during the question and answer process, which were related to data transmission problems. Therefore, as AI technology continues to evolve, maintaining computational speed is a crucial aspect of embracing the AI era.
Silicon Photonics has the potential to enhance the speed of optoelectronic transmission, addressing the signal loss and heat issues associated with copper wiring in current computer components. Consequently, semiconductor giants such as TSMC and Intel have already invested in related research and development efforts. In this context, we interviewed Dr. Fang Yen Hsiang, director of the Opto-Electronics Micro Device & System Application Division and Electronic and Optoelectronic System Research Laboratories at the Industrial Technology Research Institute (ITRI), to gain insights into this critical technology.
What Is the Relationship Between Silicon Photonics and Optical Transceivers?
An optical transceiver module comprises various components, including optical receivers, amplifiers, modulators, and more. In the past, these components were individually scattered on a PCB (printed circuit board). However, to reduce power consumption, increase data transmission speed, and minimize transmission loss and signal delay, these components have been integrated into a single silicon chip. Fang emphasizes that this integration is the core of Silicon Photonics.
Integrated Circuits’ Next Step: The Three Stages of Silicon Photonics
Silicon Photonics has been quietly developing for over 20 years. The traditional Silicon Photonics pluggable optical transceiver modules look very much like USB interfaces and connect to two optical fibers—one for incoming and one for outgoing light. However, the electrical transmission path in pluggable modules had a long distance before reaching the switch inside the server. This resulted in significant signal loss at high speeds. To minimize this loss, Silicon Photonics components have been moved closer to the server’s switch, shortening the electrical transmission path. Consequently, the original pluggable modules now only contain optical fibers.
This approach aligns with the actively developing “Co-Packaged Optics” (CPO) technology in the industry. The main idea is to assemble electronic integrated circuits (EIC) and photonic integrated circuits (PIC) onto the same substrate, creating a co-packaged board that integrates chips and modules. This co-packaging, known as CPO light engines (depicted in figure “d” below), replaces optical transceivers and brings optical engines closer to CPU/GPU chips (depicted in figure “d” as chips). This reduces transmission paths, minimizes transmission loss, and reduces signal delay.
According to ITRI, this technology reduces costs, increases data transmission by over 8 times, provides more than 30 times the computing power, and saves 50% in power consumption. However, the integration of chipsets is still a work in progress, and refining CPO technology will be the next important step in the development of Silicon Photonics.
Currently, Silicon Photonics primarily addresses the signal delay challenges of plug-in modules. As technology progresses, the next stage will involve solving the electrical signal transmission issues between CPUs and GPUs. Academics point out that chip-to-chip communication is primarily based on electrical signals. Therefore, the next step is to enable internal chip-to-chip communication between GPUs and CPUs using optical waveguides, converting all electrical signals into optical signals to accelerate AI computations and address the current computational bottleneck.
As technology advances even further, we will usher in the era of the “All-Optical Network” (AON). This means that all chip-to-chip communication will rely on optical signals, including random storage, transmission, switching, and processing, all of which will be transmitted as optical signals. Japan has already been actively implementing Silicon Photonics in preparation for the full transition to all-optical networks in this context.
Where Does Silicon Photonics Currently Face Technological Challenges?
Currently, Silicon Photonics faces several challenges related to component integration. First and foremost is the issue of communication. Dr. Fang Yen Hsiang provides an example: semiconductor manufacturers understand electronic processes, but because the performance of photonic components is sensitive to factors such as temperature and path length, and because linewidth and spacing have a significant impact on optical signal transmission, a communication platform is needed. This platform would provide design specifications, materials, parameters, and other information to facilitate communication between electronic and photonic manufacturers.
Furthermore, Silicon Photonics is currently being applied in niche markets, and various packaging processes and material standards are still being established. Most of the wafer foundries that provide Silicon Photonics chip fabrication belong to the realm of customized services and may not be suitable for use by other customers. The lack of a unified platform could hinder the development of Silicon Photonics technology.
In addition to the lack of a common platform, high manufacturing costs, integrated light sources, component performance, material compatibility, thermal effects, and reliability are also challenges in Silicon Photonics manufacturing processes. With ongoing technological progress and innovation, it is expected that these bottlenecks will be overcome in the coming years to a decade.
This article is from TechNews, a collaborative media partner of TrendForce.
(Photo credit: Google)