Nvidia’s runaway success in the AI market is fueling a furious race for innovations and discoveries by enterprises seeking to set themselves apart in the semiconductor industry.
Semiconductor companies are famed already for spending a large chunk of their revenue on R&D. It dates to the history of the industry. The market was built on technological innovations and continues to thrive on the discoveries of outstanding inventions by individual enterprises as well as their engineers and scientists. That foundation is getting reinforced in all areas as the market undergoes a tectonic shift that placed artificial intelligence in the center of many economic activities. With AI an increasing reality, but one that requires spending a ton of money to advance its promises, chipmakers hoping to cash in on the opportunities are increasingly willing to spend more on innovations that will give them an edge in the market, according to industry sources. “Artificial intelligence underpins the industry’s near-term growth and revenue expectations,” said Mark Gibson, head of US technology media and telecommunications at KPMG Global. “The upward trajectory for the industry in the short-term is clear, but the companies that can manage their supply chains and attract and retain talent will be the ones well-positioned to sustain and benefit from the AI boom.”
With Nvidia Corp. and others leading the charge, companies watching the explosive sales growth of the leading enterprises serving the AI market are jacking up investments in product innovations, they said. Analysts said they expect spending on semiconductor R&D to increase significantly in 2025 and for the next several years because of competitive pressures and the constant need to improve existing products and introduce new ones. “Innovation has been called for, making research and development (R&D) a core focus for all within the industry, said Jordan Lorence, head of marketing at technology recruitment firm MRL, in a post. “As demand continues to grow for semiconductors, new approaches to development, design, and deployment will be needed.”
The industry already spends a huge amount of money on R&D with figures ranging from a low of $150 billion to as high as $200 billion annually, including commitments by governments. Much of this amount is being spent by the big publicly owned semiconductor companies but a lot of resources are also being poured into the industry for innovation programs by OEMs, hyperscalers, and start-ups. Countries and regional bodies have also entered the fray, doling out billions of dollars in grants and tax incentives to accelerate semiconductor innovations. It is impossible to determine how much is being voted for R&D activities, though, as some of the funds are being spent quietly on innovations that investors would like to keep mum about. Additionally, figures on R&D spendings by countries like China are not typically publicly available due to concerns about disclosing what the companies are working on. Industry analysts estimate chipmakers should be spending as much as 30 percent of sales on new product development although most companies spend well below this amount. Still, the actual figures can be huge depending on the position of a company in its market segments, the nature of its products and requirements for future market share gains.
Staying ahead
Some enterprises feel compelled to spend double-digit percentages of their annual sales on R&D in a bid to stay ahead of or catch up with the competition. “To remain competitive in the semiconductor industry, firms must continually invest a significant share of revenue in both R&D and new plants and equipment,” said the US Semiconductor Industry Association (SIA), in its 2023 Factbook. “The pace of technological change in the industry requires that companies develop more complex designs and process technologies, as well as introduce production machinery capable of manufacturing components with smaller feature sizes.” The SIA estimates American IC vendors as a group on average spend about 15 percent of annual revenue on R&D, making them one of the economy’s biggest spenders. “The ability to design and produce state-of-the-art semiconductor components can only be maintained through a continual commitment to keeping pace with industry-wide investment rates of roughly 30 percent of sales. The rapid pace of industry innovation requires large outlays in capital spending to continue to produce more advanced devices.”
Intel Corp. is in the category of companies that maintain high R&D and capital equipment expenditures even during market recessions. In 2023, for example, the company shelled out $16 billion on R&D, accounting for almost 30 percent of its revenue of $54 billion. Intel doesn’t typically break out how it spends its R&D budget, but observers say this must include programs aimed at bridging the gulf between the company and rivals in artificial intelligence. “It’s not surprising that the company is cagey about exactly what its R&D budget is being used for,” notes MRL’s Lorence. “But in broad strokes, it has acknowledged investment in artificial intelligence (AI integration), intelligent edge infrastructure, and semiconductor security and traceability. The brand is also looking at semiconductor design and programming, with an aim to ‘accelerate the pace of innovation’ within the industry.”
While Intel represents one of the industry’s biggest spenders on R&D, many of its peers are also known for laying out billions of dollars on research and development activities. In 2023, Advanced Micro Devices Inc., for example, spent $5.9 billion on R&D, representing 26 percent of its $23 billion revenue. The latest figures indicate that the numbers have increased. In the first nine months of 2024, AMD spent $4.7 billion on R&D. Its sales during the period were $18.1 billion, which means R&D expenditure accounted for more than 26 percent of revenue. Qualcomm Inc. is another big spender. In its last fiscal year, the company spent $8.9 billion on R&D, accounting for 22 percent of revenue of $39 billion. Other companies spend much less on R&D. As a group, though, the US semiconductor industry dedicates a larger portion of their revenue to research and development activities than any other manufacturing sector of the American economy, according to the SIA. “Annual R&D expenditures as a percent of sales have exceeded 15 percent over the past 20 years,” the association said. “This rate is unprecedented among major manufacturing sectors of the United States economy. The rapid pace of technological change requires constant advancements in process technology and device capabilities.”
Intel’s numbers are larger than most of its competitors because the company remains one of the industry’s IDMs, forcing it to devote substantial amounts of money to product development as well as improving manufacturing processes. Fabless chipmakers spend less on R&D than their IDM counterparts, leaving foundry partners to conduct their production process improvement workload. This is why Taiwan Semiconductor Manufacturing Co. Ltd., for example, devotes as much as 8 percent of its revenue to R&D, leveraging this to keep its leadership of the pureplay foundry business.
Signs are emerging that R&D expenditures are heading higher. The market remains sluggish and the outlook for 2025 is not as rosy as analysts would like as the industry enters the first phase of its recovery. Nevertheless, chipmakers are likely to pour more funds into R&D over the next several years, observers note, because of rapid technology changes, the impact of artificial intelligence and the need to ensure they can remain competitive. Nvidia Corp., with its runaway success, may have unwittingly poured fuel on the industry’s already intense innovation fire.
Outstanding hardware and software innovations paved the way for Nvidia’s rise to the apex position in the global semiconductor market. The company is today the world’s most valuable chipmaker, one of its fastest-growing enterprises, and one of the biggest based on sales, all built on the breakthroughs Nvidia has notched over the years in GPUs and on the dominance gained by its CUDA parallel computing platform. Nvidia describes CUDA as “revolutionary” noting it “makes it possible to use the many computing cores in a graphics processor to perform general-purpose mathematical calculations, achieving dramatic speedups in computing performance.” Nvidia’s competitors – and even companies outside its sphere but focused on the semiconductor market – are searching for a similar Holy Grail, the technological innovation that can set them apart and power their growth. “The pace of technological change in the industry requires that companies develop more complex designs and process technologies, as well as introduce production machinery capable of manufacturing components with smaller feature sizes,” said the SIA.