R&D Market Pulse: GM scales down; Amazon to build AI super computer; AI firms partner with Pentagon

While some R&D sectors declined, others expanded. GM sold an EV battery plant and is the second biggest automaker withdrawing from China. At the same time, Nvidia expanded in Vietnam, Amazon built an AI-based supercomputer, and OpenAI joined Anduril to boost the U.S. military’s drone technology.

The R&D World Index (RDWI) for the week ending December 6, 2024, closed at 3,967.30 for the 25 companies in the RDWI. The Index was up 2.99% (or 115.14 basis points). Sixteen RDWI members gained value last week from 0.27% (Bristol-Myers Squibb) to 8.61% (Meta Platforms). Nine RDWI members lost value the previous week from -0.06% (Sanofi SA) to -13.01% (Intel).

Automotive R&D

General Motors to sell EV battery plant

RDW Index member General Motors, Detroit, announced last week that it was backing out of a nearly completed $2.6 billion electric vehicle (EV) battery plant in Lansing, Michigan. GM will sell its share of the battery plant to its partner, LG Energy Solution, in Seoul, South Korea. The sale is expected to close in the first quarter of 2025. The Lansing plant was to be the third battery plant operated by GM and LG, according to an agreement signed in 2019. The other two plants, Warren, Ohio, and Spring Hill, Tennessee, are unaffected by the Lansing sale. Automakers are generally scaling back on their EV production plans due to a lack of consumer demand. President-elect Trump’s plans to eliminate the current $7,500 tax credit on EVs and the lack of government support to produce batteries for EVs.

General Motors takes $5B in charges due to China business’s weakness

General Motors announced last week that it expects to take more than $5 billion in non-cash charges in the fourth quarter because of weakness in its China business that will force the automaker to close plants and offer fewer models. GM said it will write down the value of its stake in partnerships with China’s state-owned SAIC Motor by $2.6 billion to $2.9 billion or nearly half of their value. The move is in light of the company’s weaker long-term outlook. GM has been losing money throughout the year, a significant change from the past decade when it added about $2 billion annually from the country. Home-grown BYD has increasingly captured the market with affordable and tech-friendly vehicles. Considering these changes, GM has stated that it will take action to make the business sustainable and profitable.

AI R&D

Amazon Web Services to build AI-based supercomputer

Amazon.com, Seattle, announced last week that its cloud-computing arm, Amazon Web Services (AWS), will build a massive artificial intelligence-based (AI) supercomputer, termed an “ultracluster,” which will consist of hundreds of thousands of its in-house developed Trainium chips, as well as a new server. This is a development of Amazon’s AI chip design lab based in Austin, Texas. This chip cluster, Project Rainier, will be used by AI startup Athropic, San Francisco, in which Amazon and Google, Mountain View, California, have invested $6 billion. Project Rainier is expected to be completed in 2025 and will be one of the largest in the world for training AI models. AWS also announced that Apple is one of its newest customers for its Trainium chips. AWS is positioning the Trainium chip as an alternative to Nvidia’s graphics processing units (GPUs).

Nvidia to expand Vietnam presence

Last week, Nvidia, based in Santa Clara, California, announced plans to expand its presence in Vietnam by collaborating with the Vietnamese government to establish a joint AI R&D center and a data center. Additionally, Nvidia has acquired VinBrain, a healthcare AI startup under Vietnam’s largest private conglomerate, Vingroup. Nvidia will use the R&D center to focus on software development, capitalize on the country’s strong talent pool of STEM engineers, and engage industry leaders, startups, government agencies, and universities.

OpenAI and Anduril Industries join up to help DOD counter drone attacks

OpenAI, based in San Francisco, announced last week that it has agreed to collaborate with Anduril Industries, based in Costa Mesa, California, to add its technology to the U.S. military’s systems to counter drone attacks. The partnership marks OpenAI’s most profound involvement with the U.S. Department of Defense and its first tie-up with a commercial weapons maker. It is the latest example of Silicon Valley’s shunning of the Pentagon a few years ago to forge deeper ties with the national security complex.

Quantum computing funding update

Bipartisan U.S. senators authorize funding to advance quantum R&D

A bipartisan group of U.S. Senators, Maria Cantwell, D-Wash., Todd Young, R-Ind., Dick Durbin, D-Ill., and Steve Daines, R-Mont., introduced a bill last week that would authorize $2.7 billion in federal funding through fiscal 2029 to advance quantum R&D efforts at federal science agencies, including the National Institute of Standards and Technology, the National Science Foundation (NSF) and NASA. According to the bill sponsors, the National Quantum Initiative Authorization Act strengthens America’s competitive edge through critical R&D investments. Cantwell is chair of the Senate Commerce, Science, and Transportation Committee. Under the bill, the White House Office of Science and Technology Policy would be required to create an international quantum cooperation strategy to coordinate R&D activities with U.S. allies.

Academic R&D

U.S. colleges and universities spent 108.8B on R&D

Total R&D expenditures at American colleges and universities rose to $108.8 billion in FY 2023, according to the NSF’s latest Higher Education R&D Survey, released last week. This represents an increase of just over $11 billion, or 11.2 %, compared to FY 2022, the most significant annual growth since FY 2003. As it has for decades, Johns Hopkins University led the list of academic institutions performing R&D. JHU spent $3.8 billion on R&D in FY 2023.

Semiconductor R&D

Commerce Department changes export rules for chips

The U.S. Department of Commerce (DOC) has changed its export rules for advanced semiconductor chips. The latest regulations prevent makers of advanced memory chips, also known as high-bandwidth memory, from shipping their products to China without permission from the DOC. Such chips work with AI processors in the computations behind generative AI systems. Three major manufacturers of these memory chips are SK Hynix, Samsung Electronics of South Korea, and Micron Technology, which is based in the U.S.