Semiconductor industry driven AI revolution: Opportunity beckons for India
Semiconductor industry & the memory and storage technology will drive the AI revolution as AI systems need large amount of data to be processed at high speeds.
In the modern history, every technological revolution coincided with the rise of a powerful nation state, whether it was Great Britain with the first industrial revolution in the 18th century or America with the second industrial revolution in the early 20th century. More recently, China has leveraged the power of technology, cheap labor, and concerted government policy to become one of the largest economies in the world with a global influence second only to the United States. With the advent of Artificial Intelligence, we are at yet another inflection point in the human evolution where if done right, India can be the beneficiary and become one of the global superpowers. For it to happen, we do have all the right pieces in place – a large number of skilled and experienced IT personnel, government and regulatory policies which are promoting and incentivizing new ideas, and a burgeoning middle class. To quote Rajeev Chandrasekhar, Union Minister of State for Electronics and Information Technology, India will be a significant pole in the universe of AI going forward.
According to an estimate released by the Central Government’s Ministry of Electronics and Information Technology, AI is expected to add US $967 billion to the Indian economy by 2035 and will account for 10% of the projected GDP target by 2025. While US and China lead the race when it comes to investment in AI, India is not too far behind. According to Stanford University’s 2023 AI Index report, India stood at the 5th position in terms of the investment capital received by startups offering AI-based products and services.
In addition to the large language models and progress in software, semiconductor industry and more specifically the memory and storage technology will drive the AI revolution, as all the AI systems require large amount of data to be processed at high speeds. According to an estimate released by the Fortune Business Insights, the combined memory and storage market will triple from what we had in 2021 to the projected 2030 data. A significant portion of this growth will be driven by the demand in High Bandwidth Memory (HBM). HBM is an ideal choice for high performance compute and AI because it has higher latency, smaller footprint, consumes far less power, and is able to keep up with the higher and higher core counts than a typical DRAM memory. HBM accomplishes all this by stacking multiple DRAM dies vertically using through-silicon vias (TSV). Complimenting the HBM memory speed, high-speed storage solutions will enable quick access to training datasets, while data lakes provide high performance repositories for data. Thanks to the efforts by the Government of India, we are already becoming a key part of the advanced semiconductor manufacturing base as the recent announcements by the likes of Micron and Tata electronics show, with a few more ventures under discussion.
Use of AI to further optimize the manufacturing flow in today’s advanced semiconductor fabs, which can entail more than a thousand steps for each silicon wafer, is the next organic step. AI-enabled smart manufacturing techniques can deliver higher-quality products on a faster cadence. For example, one takes images during the entire three- to four-month chip manufacturing cycle and use them to discover deviations and fix them before they can have detrimental effect on chip yield and productivity. Similar strategies are used for assembly and test operations. All these steps rely on the AI system and would be beyond the scope of humans to process in a reasonably fast turnaround time. This was an extension of the big data analytics, and the next step now is the use of Generative AI which can provide tangible and proactive actions even before these deviations can occur. Generative AI will also be the key in improving cycle time and quality issues for architectural and circuit design during the VLSI chip design and developmental phase. Generative AI is clearly bringing new paradigm to the memory industry with several innovative solutions and approaches to ensure optimal and efficient data transfer and data compute. Outside of the semiconductor world, generative AI applications which are built on neural networks with data with billions of parameters or more have potential to create virtual assistants and customized experiences for users and automate the routine tasks.
While there is a lot of excitement around the use of AI, and justifiably so, there are also some areas of concern in terms of what it all means for future employment opportunities and the ethics of using AI including on data privacy. According to a Government of India study, the socio-economic consequences of AI will depend on how well it is managed. Central government has already introduced multiple platforms to make citizens aware of AI including its flagship program ‘AI for All’. Furthermore, the Skill India Mission and the National Education Policy initiatives are being launched to upskill and reskill the workforce so that they are ready to reap the benefits of AI. Regarding the data privacy, we all have already seen news articles of how deepfake is used to steal identities and for extortion and this is just the beginning. Here again authorities must be very vigilant that the citizens do not get impacted and that they have proper recourse to safeguard their identity and data.
It was Steve Jobs who said back in the 1980 that computer is “bicycle for the mind”. Extend that analogy further, and Artificial Intelligence will do to your productivity what adding a powerful motor will do to the speed of that bicycle. The AI revolution is coming, whether we like it or not. India cannot afford to miss this bus and must leverage its strong strengths through a top down, problem oriented, multi-disciplinary approach with the involvement of academia, industry and startups. Every sector is going to be deeply impacted by AI, be it space, education, healthcare, sustainable living or Industry 4.0. Societal opinions are going to be shaped by AI, which we can ill afford to leave to foreign players. India is already facing major security risks because of our complete reliance on import of semiconductor chips. If we do not wake up in time, the same is going to happen with AI. Opportunity is knocking for the world, and India in particular, and now it is up to us to make the most of it.
(Anand Ramamoorthy is Managing Director, Micron India and Prof. V Ramgopal Rao is Group Vice Chancellor, BITS Pilani. Views expressed are personal.)
