Fully six months after the intelligent chatbot ChatGPT was unleashed on the world, the hype around artificial intelligence (AI) has still not abated. On the contrary: more and more companies are conducting research in this area, developing new solutions and making more use of them. For experts, it is clear that AI heralds the next quantum leap in technological progress. The further it penetrates, the greater also the growth potential of the industry. The experts at Precedence Research reckon that the global AI market will climb from USDbn 119.78 in 2022 to a likely USDbn 1,597 by the end of the decade, corresponding to a hefty 38.1% mean annual rate of increase.
Whether this growth is actually achieved depends to a critical extent on the computing power of the processors. And here there is no getting past Nvidia. Not only is the semiconductor specialist heavily involved in ChatGPT, but the US group is also in the leading pack in autonomous driving thanks to its high-tech chips. That's because the flood of data that is created can only be analysed in real time if the computing speed is extremely fast. The Californians spotted the trend early, developing the “Drive PX” AI platform based on deep learning and presenting it at the Consumer Electronics Show in Las Vegas way back in January 2015. Since then Nvidia has collaborated with a range of car-makers, such as BYD, Mercedes-Benz and Toyota. The latest generation of the platform, named “Thor”, achieves a theoretical power of up to 2,000 teraflops, a huge leap forward from the predecessor version, the Drive PX2, that managed 8 teraflops. To explain, one teraflop represents 1 trillion computing operations per second. While Nvidia's automotive division currently only accounts for a small proportion of its overall business, it is very fast-growing: in the 2022/23 financial year segment sales jumped 60% to USDmn 903. The group sees the car market having long-term potential of USDbn 300.
In this exciting and fast-growing segment, however, competitors are not leaving the field entirely to Nvidia, which currently controls around 80% of the AI chip market. A few days ago, for instance, leading US processor manufacturer AMD presented an AI-based chip to go head to head against its competitors. “MI300X” is the name of the most advanced graphics processor from the research lab of Advanced Micro Devices, and is to start being shipped to customers before the end of the year. For AMD-CEO Lisa Su, AI is the company’s “largest and most strategic long-term growth opportunity”. The boss expects the market for AI accelerators for data centres to expand from about USDbn 30 this year to more than USDbn 150 in 2027, a compound annual growth rate of over 50%.
While AMD and Nvidia are helping AI achieve the necessary speed with their graphics processing units (GPU), Palantir is working on intelligent applications. The data analysis software developer has developed a generative AI platform that is based on similar technology to that behind ChatGPT and, for instance, could help the military fight hostile forces. Interest in the new software, according to CEO Alexander Karp, is “unlike anything we have seen before”. The US group also has an AI solution for the private sector. At the end of May Palantir won the custom of fashion giant C&A, which is using the software to optimise its supply chain. Just a few days later lithium-ion battery manufacturer Panasonic Energy also decided in favour of the same AI software, even simultaneously launching a multi-year partnership with Palantir for the operation of a “Smart Factory”.
Source: Precedence Research