The CHIPS Act promises to boost U.S. competitiveness, innovation and national security with multi-billion-dollar funding for semiconductor manufacturing, leading-edge technology programs and advanced R&D. A commendable goal of course, but the CHIPS Act is helping the wrong companies.
The dilemma facing U.S. semiconductor companies is well-known. The U.S. makes only 12% of the world's chips, a steep decline from the 1990s when it produced 37%. And in the more critical and profitable advanced designs, foundries in Asia produce over 90% of these chips. This geographic imbalance became all too obvious during the pandemic when significant supply chain disruptions had a major impact on the U.S. economy.
I am fully aligned with the goal of strengthening American manufacturing and supply chains, enabling U.S. companies to compete more effectively and reverse these trends. What’s wrong with this picture? The CHIPS Act funding is going to the nation’s largest chip companies, the same companies that lost their competitive edge to TSMC and Samsung in the first place. At best, this funding might help these large companies catch up to their Asian counterparts, rather than help smaller U.S. companies that provide the innovation behind key technologies of tomorrow to manufacture domestically.
Technology Comes Second
Yes, the CHIPS Act has the potential to bring clear benefits to the U.S. in increased semiconductor manufacturing and a reduction in chip shortages. And yes, the program will generate new jobs. But will it create innovative semiconductor technology that will surpass foundries in Asia? The answer is no.
The U.S. semiconductor industry was out-executed years ago by the likes of TSMC, which focused on state-of-the-art technology and a culture of customer service, establishing key partnerships with companies and listening to their customers. TSMC never had internal design teams competing with their customers and never dictated design methods to their customers. They focused on manufacturing, and as a result produce more silicon wafers than anyone else. Everything that simplifies adoption of their technologies, they automated and ensured designs quickly got into their fab.
Meanwhile, TSMC has announced the opening of its second plant in Arizona, raising its investment in the state from $12 billion to $40 billion—making it one of the largest foreign investments in U.S. history. If protecting the supply chain and making semiconductors in America is the only objective, this would already achieve that.
The Limits of Moore's Law
For U.S. investment in semiconductor technology to be sustainable in the long term, we need to surpass parity with existing leaders. The semiconductor industry is dynamic and fast-moving, with foundries announcing new process nodes every 18 months. To regain leadership, we need to solve the core issues that now limit Moore’s law, the decades-ago observation by Intel founder Gordon Moore who observed in 1965 that the number of transistors in a microprocessor would double every two years. This precept accurately predicted an exponential increase in computer performance—a phenomenon held true for decades. Today this is no longer the case. The design of today’s transistors has added extreme complexity to the design process and Moore’s Law has started slowing down, which raises both concern and opportunity for U.S. semiconductor makers .
First, the concern. The advancement predicted by Moore’s law has been interrupted by physical phenomena in semiconductor technology, such as increases in heat generation, which impedes processor performance and interconnect technology. This has now become the barrier for scaling the performance of SoC (System-on-Chip) architectures needed in increasingly necessary artificial intelligence capabilities. These phenomena hinder the performance advancements promised by Moore’s law, and so hinder an entire industry.
Now here is the opportunity: What if we could overcome these roadblocks by reclaiming the trajectory of Moore’s law?
The U.S. needs to fund semiconductor innovation and solve core problems, such as the heat and interconnect issues I mentioned today, limiting Moore’s law. We need to regain industry leadership, and I believe we can do it.
Chips off the New Block
America has always been about leadership and has historically funded crucial projects intended to win the war against the enemy. Today, new companies with innovative semiconductor technology will help America win the chip war.
Companies solving the semiconductor heat problem include Frore (one of our portfolio companies), which has developed the world’s first solid-state chip for active device cooling. A company called JetCool provides semiconductor liquid cooling via microjets to pass liquid coolant directly to the chip’s surface.
There are also novel approaches to solving the interconnect problem, ranging from optical solutions such as Avicena (another one of our portfolio companies), offering microLED optical interconnect and Lightmatter, providing photonics supercomputers. Ayar Labs is developing multi-wavelength lasers and optical chiplets. And Astera Labs is creating a new purpose-built hardware solution of smart cable modules that overcome the reach, signal integrity and bandwidth of copper interconnect.
These are just a few companies that are ready to advance U.S. semiconductor leadership.
We can innovate again and reclaim semiconductor technology leadership by funding new companies just as we did in WWII. In 1941, U.S. President Franklin D. Roosevelt created the Emergency Shipbuilding Program to rapidly build cargo ships and send much-needed supplies to the Allies for the war effort. He eventually funded industrialist Henry J. Kaiser, who had never built a ship before in his life. Between 1941 and 1945, Kaiser built 2,751 Liberty cargo ships, the largest class of ships ever built … and he built them fast! Kaiser shipyards reduced the industry average of 230 days to 45 days and ultimately to less than three weeks! How did he do it? He revolutionized shipbuilding by introducing mass-production techniques, instituting modular construction and assembly techniques in which ship sections were welded together instead of riveted. The sheer number of Liberty ships at sea overwhelmed German forces. American ingenuity helped the Allies win World War II.
Inventiveness, resourcefulness and sheer brilliance are part of America’s history. We can again be the leader in semiconductors by funding truly innovative technology, and the CHIPS Act presents an opportunity to make a choice. Let’s make it a good one.
Rajeev Madhavan, serial entrepreneur and found of several startups in EDA and chip design, is co-founder and partner of Clear Ventures, which invests in IT, including semiconductors.
