In recent discussions regarding the industrial semiconductor landscape, the advent of large AI models has ushered in notable growth opportunities as well as significant challengesCorporations across the sector find themselves at a crossroads where innovation must harmonize with energy efficiency, rising competition, and the increasing demand for embodied intelligence applicationsAs industry leaders converge during summits to share insights, strategies and approaches to leverage these opportunities while navigating complexities have become focal points of conversation.

At a recent industrial conference, Francesco Muggeri, Vice President of Power Discrete and Analog Products Device Marketing and Applications for STMicroelectronics in China, elaborated on the company's ambitious strategy dubbed "For China, In China." This initiative is shaping ST's commitment to establishing a comprehensive end-to-end supply chain within Chinese borders

The endeavor exemplifies the tech company's proactive adaptation to the evolving market landscape.

Building a solid foundation, ST has already inaugurated a back-end testing factory in ShenzhenIn June of this year, a joint venture was established with San’an Optoelectronics to build a silicon carbide (SiC) manufacturing facility in Chongqing, targeting applications within the automotive sectorRecently, further collaboration with Huahong Semiconductor saw the launch of 40nm node microcontrollers (MCUs) produced locally in ChinaMuggeri emphasized the rationale: “Our deep investment in the Chinese market is driven by observable competitionBy boosting our local production capabilities, we aim to enhance our service within this dynamic market.”

As a major player in the industrial semiconductor arena, ST has navigated broader market pressures with impressive resilience

Amid an overall decline in global industrial semiconductor revenues by 6.5% in 2023, ST surfaced with a remarkable 15.1% year-on-year growthSuch strides enabled an increase in their market share from 5.6% in 2022 to 6.9% in 2023, a testament to their strategic positioning.

The company’s choice to actively integrate advanced materials such as silicon carbide and gallium nitride within their product offerings aligns with current trends aimed at reducing energy consumption and carbon emissionsNotably, certain devices by ST are now capable of running large language models, facilitating advancements towards edge intelligence—a concept rapidly gaining traction in modern industrial applications.

In the past year, ST’s collaborative approach with San’an Optoelectronics to accelerate silicon carbide product development has garnered considerable attentionThis partnership is intricately tied to the explosive growth of the new energy market in China, where silicon carbide devices witness increasing demand.

Domenico Arrigo, a senior executive at ST, elaborated on the trajectory of compound semiconductor technology, underscoring a strategic focus on accelerating silicon carbide advancements due to its maturity and market relevance

Yet, he also highlighted ongoing efforts to enhance gallium nitride (GaN) solutions, indicating an overarching trend towards designing highly integrated systems.

“Comprehensive packaging solutions are critical in maximizing gallium nitride's performanceEffective and direct driving connections are key, with around 70% of market reliance on these integrated solutions.” Arrigo explained the significance of integration, underlining that its importance cannot be overstated in the gallium nitride market context.

While silicon carbide power devices have seen rising application in electric vehicles, cost remains a pivotal factorAlthough silicon carbide devices traditionally bear higher costs compared to their silicon counterparts—a challenge in early market adoption—companies like Tesla, which are leading the charge in electric vehicle efficiency, continue to drive silicon carbide trends forward.

Industry professionals assert that cost efficiency should not solely focus on the component costs

alefox

Instead, the broader narrative anticipates a comprehensive evaluation of the overall impact silicon carbide devices have on electric vehicle performance, which reveals a lower overall application cost compared to traditional silicon devicesViewed through an integrated lens, the potential of silicon carbide to enhance operational efficacy outweighs initial investment concerns.

Moreover, Angelo Rao, another executive at ST, emphasized that silicon-based IGBTs and silicon carbide technologies are not in competition; rather, they complement each other within varied scenarios, industrial conditions, and development phases“Silicon carbide devices outperform in harsh operational settings, enabling prolonged operation times and improved efficiency within constrained spaces,” he remarkedThis efficiency allows vehicles designed with silicon carbide components to achieve energy targets unattainable by other technologies.

For instance, under controlled test conditions involving electric vehicles with identical power ratings and battery capacities, replacing traditional silicon-based IGBTs with silicon carbide devices resulted in a significant increase of 16% to 19% in overall driving range, as verified through extensive simulations by ST and their partners.

In discussing applications for both primary compound semiconductors, Muggeri pointed out that silicon carbide is best suited for high-voltage, temperature-sensitive environments, while gallium nitride excels in high-frequency, relatively low-voltage applications

Practically, silicon carbide serves as a staple within automotive power supplies and motor control, whereas gallium nitride finds its niche in adapters and inverters.

In tandem with the booming demand for data centers driven by large AI models, concerns arise regarding their significant energy consumptionHere, the application of compound semiconductors plays a critical role in decarbonization effortsZhou Guangzu, head of ST's Industrial Power and Energy Technology Innovation Center, introduced ST’s strategy leveraging silicon carbide to enhance efficiency and power density, employing their STGAP isolated gate drivers to mitigate the energy demands of data centersThis innovative solution is designed to offer a 5.5 kW power supply for AI-centric data centersAdditionally, transitioning infrastructures to high-voltage direct current (HVDC) systems reduces transmission losses, further optimizing data center efficiency.

“While enhancing power density is essential, effective heat management within limited spaces is equally crucial,” Zhou added, proposing that advancements in passive component technology would complement power design initiatives and facilitate multi-phase converter structures, showcasing gallium nitride's unique advantages in size reduction and density enhancement in future applications.

The infusion of intelligent robotics into manufacturing workflows is also rising, with not just robotic arms but humanoid robots penetrating factory settings, signifying a new era of automation in smart factories.

Allan Lagasca, leader of STMicroelectronics’ Smart Industrial Global Segment and Automation Technology Innovation Center, shared insights into the burgeoning influence of humanoid robots

“Unlike traditional robotic arms restricted to monotonous tasks, humanoid robots are increasingly embraced for their capability to partake in decision-making processes and connectivity within factory equipmentThis shift pushes forward the smart factory initiative,” he explained.

Breaking down humanoid robots reveals their reliance on traditional functionality: composed of motors, sensors, connectors, alongside AI components serving specific roles“We believe ST’s microcontrollers will play a pivotal role in propelling humanoid robotics into smart production facilities,” Lagasca concluded, highlighting the company’s commitment to developing collaborative solutions with clients aimed at heightened productivity.

Furthermore, Arnaud Julienne, another executive at ST, categorized humanoid robots into two distinct varieties“One category embraces advanced humanoid designs with powerful processors needing extensive AI resources; the other is simpler in conception, which currently aligns with ST's focus.” He expounded on the simpler humanoid robots that integrate embedded microprocessors for system management, providing oversight over sub-systems such as motors and actuators.

Artificial intelligence has emerged at the nexus of edge nodes, granting immediate feedback and action capabilities in automated systems without waiting for final process evaluations