Blog

Senate passed a $280B CHIPS and Science bill to boost the technological competitiveness of the United States. In this essay, I will provide the key takeaways from the bill, go into detail about why this bill was required, and whether it will help solve the microchip crisis. Since Intel gave its earnings this week, it made a lot of sense to me to add a section about it in this essay.

This essay is not about the funding distribution of the bill but more about the crisis in the chip industry and where the CHIPS bill comes into the picture. Critical causes I discuss in detail include:

• The drought of startups and funding in the US chip industry
• Legacy companies are focusing on reducing R&D and CapEx and increasing dividends and share buybacks to attract shareholders/keep existing ones happy

What is CHIPS Bill?

Creating Helpful Incentives to Produce Semiconductors (CHIPS) bill aims to establish investment and incentives to support US semiconductor manufacturing, R&D, and supply chain security of next-generation microelectronics. It provides an income tax credit for semiconductor equipment or manufacturing facility investment through 2026.

Dollar distribution of the bill:

• $52B in grants, loan guarantees, and other support to increase US semiconductor production.
• $39B to be spent over five years to build, expand and modernize domestic manufacturing capabilities.
• $6B for direct loans and loan guarantees.
• $11B to support R&D and workforce development programs.
• 25% investment tax credit for investments in semiconductor manufacturing property.
• $200B in future federal spending to promote research programs based out of the US.

Why was CHIPS Bill necessary?

Semiconductor startups used to rule the roost in Silicon Valley. The very name, Silicon Valley, comes from the birth of the semiconductor industry in the San Francisco bay area 60+ years ago. The US has been losing its dominance over the semiconductor industry over the last couple of decades. The share of modern semiconductor manufacturing capacity located in the US has eroded from 37% in 1990 to 12% today.

SemiEngineering.com tracks startups in the semiconductor industry. Look at the worrying trend for the US in the table below. Clearly, China has well-funded startups in the chip industry.

This does not necessarily mean that all Chinese startups will succeed. Considering the regulations and scrutiny in the US means US startups are more likely to succeed. But the gap in the number of startups is a big issue. So the question arises, why are US-based venture capitalists only investing in “tech” companies? Forgetting that companies such as Apple, Google, Amazon, and Microsoft would not exist without semiconductors.

Semianalyis, a blog by Dylan Patel, did a great article on the dearth of VCs in the chip arena. The core reason behind ignoring hardware and semiconductor space is high startup costs. These are capital-intensive businesses and the upside is low compared to tech companies. “Tech” companies can be run with a few employees who can work from home and a few million dollars to build a prototype. Whereas, It takes roughly $50M of financing to even get to a product and another $200M or more to reach production.

The problem is not only with startups but also with legacy companies. The share of R&D investment has gone down from 40% to 15%. At the same time, China has provided a tax-friendly environment with loose regulatory policies and massive subsidies. Recent policies in the US, until the CHIPS bill was passed, both at the Federal and State levels, did not create a conducive environment for manufacturing capacity and fab units to be built.

It’s no surprise that even though all the critical components required for building semiconductors are US based or owned, still the dominance has shifted outside of the US. The largest semiconductor equipment, design, and software companies are based in the US. Lam Research and Applied Materials are based out of the US. EUV (required for chip manufacturing, basically one cannot manufacture a chip without them) source and EUV collector are based out of the US. Critical software needed for chip design called EDA (without this software you cannot design modern chips) is in the US.

The cost of constructing and operating a semiconductor fabrication facility in the US is 25 to 50% higher than outside the US. The primary reason is the outsourcing of manufacturing capacity. Coming back to the CHIPS Bill, how will this bill help? Here is my understanding.

Will CHIPS bill help?

Countries across the world including China, Japan, South Korea, and Europe are rolling out major federal incentive programs to provide a strong foothold to semiconductor companies. In this scenario, the US federal incentive program is needed to reverse the decline of US manufacturing leadership and re-establish the US as a competitive location for semiconductor manufacturing.

I believe the step taken with the CHIPS bill is in the right direction but it may not be enough. Constant updates and amendments are required to support the long-term vision so as to keep the US as the leader in the chip industry. But the first major win is the passing of the CHIPS bill which was introduced two years ago.

Let’s talk about the pros and cons of the major outcomes of the CHIPS bill.

Manufacturing Funding: $52B is like a drop in the bucket compared to $250B given by the Chinese government in subsidies to its companies in the form of tax rebates, grants, and subsidized loans.

Research and Development: The bill provides $11B in R&D investments. This includes funding for the National Science Foundation (NSF), National Institute of Standards and Technology (NIST), Department of Commerce (DoC), and Department of Energy to engage in R&D partnerships. It is great to see funds going to the NSF to create startup incubators and accelerators targeted at the semiconductor industry.

Con: I believe this is not enough. Unless you help big companies like Intel, Micron or Nvidia with R&D dollars, the US dominance may not return.

Tax credits: The bill should have included provisions for chip companies to get tax credits as they increase their R&D programs. The government does provide tax credits via the FABS act for construction, expansion, and modernization of the fab plans.

Con: The 25% manufacturing investment tax credit offered by the bill may not be sufficient. My argument is for tax credits on R&D which is a big portion of chip companies’ expenditure (usually around 20-25%).

Expansion in China: Putting guardrails and provisions that restrict companies receiving CHIPS bill funds from reaching companies that are expanding in China. But the guardrails are very limited and Intel is still buying Chinese developed and manufactured tools despite the availability of alternative suppliers and US subsidies.

Con: The exceptions to expanding business in China are only limited to existing factories and legacy semiconductors such as the 28nm chips (nanometer refers to the line-width between transistors on a chip, the smaller the number more advanced the chip. For perspective, iPhone 13 uses 5nm chip). It still permits the new generation of semiconductors to be made in China. Having said that, 28nm is the most widely used chip in the industry today, and it’s not advanced technology. China can easily make them.

Summary of Intel earnings

Intel for the first time in 30 years has reported a quarterly loss. The global economy is slowing and there is a slowdown in the PC industry. Outside of that, Intel has been losing its share in the data center business.

I will very quickly summarize the points where I believe Intel went wrong.

• Firstly, Intel paid $1.5B in dividends and remains committed to growing the dividend over time.
• Secondly, Intel reduced its CapEx forecast to $23B ($4B less than previous guidance).

I am not against dividends, but if you are going to cut down CapEx from the new fab buildings which will play a huge role in Intel’s turn-around plan then it is a mistake to prioritize shareholders’ short terms pocketbooks.

But you may argue that TSMC has also pushed out some of its CapEx to next year citing lower demand. Let’s not forget that Intel is behind compared to TSMC in manufacturing capacity but yet is reducing its investment to just over half as much as TSMC going forward because Pat Gelsinger wants to prioritize dividends.

Conclusion

I am not a huge fan of governments’ involvement in free markets and trade. As it often highlights the dangers and inefficiencies of government interference. However, at the same time, I also understand that key supply chains including those of rare earth metals, solar panels, and certain pharmaceuticals are not free from geopolitics. South Korea, Japan, and the EU have outlined aggressive plans to woo semiconductor factories. Taiwan is at risk of invasion and has become an untenable security threat.

Semiconductors are necessary to power other key technologies, like quantum computing, IoT, AI, fighter jets, cars, computers, and even coffee makers. Chips are really important for national security. CHIPS bill is the beginning for the US to pursue an industrial strategy to prepare for the future.

Government should ensure that money given through the CHIPS bill is not a blank check and is not free money. They should control some of the expenditures of these public firms, especially the actions that are taken to please the shareholders in the short term.

CHIPS bill is a move in the right direction but if not followed by amendments that address the startup crisis and legislation to protect against state-sponsored espionage and forced IP transfers, it will merely just be a band-aid.

Cheers