In the wake of the rising global semiconductor demand and strategic technological competition, the U.S. enacted the CHIPS Act four years ago, aiming primarily to revitalize domestic semiconductor manufacturing. With an impressive budget of $59 billion, two-thirds was earmarked for grants and loan guarantees designated for foundries. This investment catalyzed significant developments in the industry, with giants like Intel and Taiwan Semiconductor Manufacturing Company (TSMC) establishing state-of-the-art plants in Arizona, while Samsung made substantial commitments in Texas, and Micron aimed to make strides in New York.
However, the latest updates regarding CHIPS Act funding indicate a crucial pivot in strategy. Instead of merely counting chip production as a metric for success, the focus has expanded to encompass a broader range of artificial intelligence computing capabilities. Funding now includes cutting-edge areas such as advanced packaging, memory solutions, interconnect technologies, materials science, photonics, and secure electronics.
This strategic pivot is timely and necessary. The U.S. acknowledges that it will not achieve complete self-sufficiency in semiconductor supply chains. Rather, the focus is on fortifying U.S. capabilities at critical bottlenecks while avoiding the pitfalls of overburdening taxpayers with every layer of a complex global supply chain. Particularly in countering advances from China, collaboration with trusted democratic allies emerges as an essential strategy, transforming competitors into cooperative partners.
The strategic shift is further evidenced by the recent announcement of a groundbreaking $874 million investment aimed specifically at enhancing the connective tissue of the AI compute stack. Companies like GlobalFoundries may receive up to $300 million for the development of optical data transfer equipment, crucial for overcoming data movement bottlenecks in massive AI models. Other potential beneficiaries include Kepler, slated for up to $245 million for innovative AI memory solutions, and Multibeam, which may be awarded up to $140 million for advancements in chip stacking and connectivity.
The remaining proposed funding will support various other initiatives, including low-power computing through Extropic, development of specialist materials by Aeluma, and software solutions to detect counterfeit or malicious components spearheaded by OBSIDIA.
This funding initiative serves as a pivotal reminder that leadership in AI does not hinge solely on miniaturizing transistors. Instead, future innovations will require an intricate web of coordination among processors, high-bandwidth memory systems, optical technologies, and specialized chips. Advanced packaging technologies facilitating the high-density stacking and interconnection of multiple chiplets and co-packaged optical systems that minimize energy and latency in data transfer are becoming quintessential. Additionally, innovative memory solutions serve to alleviate key constraints hindering AI training and inference.
By channeling resources into durable leverage in AI computing rather than attempting to replicate Taiwan’s well-established foundry ecosystem, these new chip subsidies aim for sustainability and resilience. Even though the $874 million figure may seem modest compared to industry norms, this investment is significant. It targets small to mid-sized businesses where public funding can act as a catalyst, thereby attracting private capital while mitigating risks associated with first-mover advantages in strategically vital sectors.
Historically, the U.S. has excelled in chip design but often lacked the manufacturing might embodied in nations like Taiwan, Japan, South Korea, and parts of Southeast Asia. The importance of cooperation in this space cannot be overstated. TSMC’s ambitious $165 billion expansion into the U.S. fosters leading-edge capacity, drawing upon Taiwanese expertise and extensive supplier networks, thus reinforcing the U.S.–Taiwan industrial partnership.
Analogously, Germany’s Bosch has begun production of silicon-carbide chips in California, enhancing U.S. supply security in power electronics without the need for the U.S. to build its own competitor from ground up.
Furthermore, initiatives like the US-JOINT, which connects a consortium of 12 American and Japanese firms, show promise in developing next-generation packaging technologies. These centers of innovation will create the backend capabilities crucial for transforming well-designed U.S. chips into comprehensive AI systems, rather than solely striving to establish a domestically-owned champion, the government is fostering an allied R&D infrastructure.
Yet, it’s crucial to note that the $874 million commitment remains a preliminary announcement, essentially letters of intent, which require due diligence and further negotiation. Success in this endeavor will hinge on disciplined execution. The U.S. must continue to welcome foreign investment from trusted allies while building its strategic advantages not through subsidy competition with China or replication of Taiwan’s prowess but by owning the next-generation technologies defining AI’s future.
Christopher Cytera CEng MIET is a senior fellow with the Tech Policy Program at the Center for European Policy Analysis and a technology business executive with over 30 years of experience in semiconductors, electronics, communications, video, and imaging.
Bandwidth is CEPA’s online journal dedicated to advancing transatlantic cooperation on tech policy. All opinions expressed on Bandwidth are those of the author alone and may not represent those of the institutions they represent or the Center for European Policy Analysis. CEPA maintains a strict intellectual independence policy across all its projects and publications.
