Ho Chi Minh city, 20 April 2023 8:00 am
If history is any guide, the enormous amount of research — this time in artificial intelligence — will result in new products in medicine, new materials, climate, and other fields. Investors would do well to keep attention.
Large corporations' research laboratories are expanding the capabilities of artificial intelligence in areas such as image recognition, natural language processing, and more. In a competition to develop these capabilities, a number of large companies are allocating increasing amounts of capital. For instance, Meta Platforms Inc. claims that its AI Research SuperCluster system will contain 16,000 Nvidia Corp. graphics processing units by midsummer, making it the world's quickest engine of its kind. In addition, the DeepMind Technologies lab, which is owned by Alphabet Inc., the parent company of Google, has announced the development of a new language model, stating, "The level of high-school reading comprehension approaches human-rater performance."
Microsoft Corporation, Amazon.com, Oracle Corporation, International Business Machines Corporation, and others are also competing aggressively in the AI race. Despite their different approaches, these companies have the potential to accelerate future advancements in drug discovery, new materials, climate change solutions, and more.
Consumers and investors, more concerned with stock market fluctuations, may not be paying attention to projects that are not explicitly related to business lines or quarterly results. However, research and development frequently result in products that exceed a laboratory's original objectives. For example, Xerox Holdings Corporation's Parc facility was a pioneer of personal computing. Bell Labs, a predecessor to AT&T Inc., developed the transistor and a prototype of an early mobile phone.
The amounts of money being spent are massive. At Alphabet, R&D expenses rose to $31.562 billion last year from $27.573 billion in 2020; Meta spent $24.655 billion on R&D last year, rising from $18.447 billion the previous year, company filings show. Alphabet said R&D spending in 2021 amounted to 12.3% of revenue, while Meta reported it was 21% of its sales. According to Morningstar Inc., which calculated a revenue-weighted average for index constituents in their most recent fiscal years, S&P 500 companies spent an average of 2.82 percent of revenue on research and development.
Meta, a Facebook subsidiary, occupies the AI race lane dedicated to scaling data and computing power. Deep learning is a type of artificial intelligence designed to imitate human neurons. The ImageNet project demonstrated in 2010 that two GPUs developed by Nvidia could be used to train a large AI model to recognize labeled images, according to Kyunghyun Cho, co-founder and associate professor of computer and data science at New York University.
Meta stated that its AI Research SuperCluster currently contains 6,080 Nvidia GPUs, but that number will increase to 16,000. The computing power will be used for language processing and computer-vision research with the goal of constructing the immersive world known as the metaverse, according to Shubho Sengupta, a software engineer working on the project. According to the company, the supercomputer will aid in the development of augmented-reality tools and rich, multidimensional experiences in the metaverse, as well as the creation of AI models that can operate across hundreds of languages, analyze text, images, and video, and develop augmented-reality models that can function across hundreds of languages. The technology will also make it easier to identify detrimental content, according to Meta. DeepMind, a London-based subsidiary of Alphabet, employs a technique known as reinforcement learning. In this method, an algorithm gains knowledge through trial and error.
Last month, DeepMind introduced AlphaCode, an artificial intelligence system that competes with human software developers. DeepMind announced in December the development of a new language model that, according to the company, reduces scope without sacrificing performance. This is in addition to DeepMind's claim that it has made a significant advancement in comprehending protein structures, which is essential for drug discovery.
Shuman Ghosemajumder, the chief of AI at F5 Inc. and a former Google employee, commented that the innovative AI techniques employed in the research could lead to significant performance improvements without requiring extensive computational power. However, it is too early to predict the potential commercial applications of this research.
Investing in AI technology may not have immediate business outcomes, but in the long term, it could yield significant benefits. Top-tier tech companies, with a substantial investment in AI, will likely inspire other companies, including those in the pharmaceutical industry, to follow suit. This could lead to significant advancements in a range of fields and provide benefits to society as a whole.