Sandip Patel, managing director, IBM India & South Asia region; Photo by: Neha Mithbawkar for Forbes India
As soon as Sandip Patel enters the spacious conference room at IBM India’s office in BKC, Mumbai, he settles in and orders a strong cup of coffee. “It has been a hectic day,” he explains. “I’ve been shuttling between Worli, Nariman Point and here.” Patel has been busy managing IBM’s India operations, especially since the company made a slew of announcements at its flagship IBM Think 2024 event in Mumbai, in September.
IBM significantly shifted its strategy during the Covid-19 pandemic, when Arvind Krishna took charge as company CEO in April 2020, with a focus on hybrid cloud and artificial intelligence (AI). “IBM India operates almost like a microcosm of the IBM Corporation. While we have a thriving domestic business, we also have three R&D powerhouses in India: India Software Labs [ISL], India Research Labs [IRL], and India Systems Development Labs [ISDL],” explains Patel who was appointed managing director of IBM India and South Asia in January 2020, only a few months before Krishna took charge.
“I see our continued growth coming from three main vectors: AI, semiconductors and quantum computing,” says Patel. Without revealing any numbers, he claims it has been three years of consecutive growth for IBM in India. And he is bullish on seeing this momentum in these new technologies, working with various governments and MSMEs in emerging markets. “We operate in India where we innovate in India, for India and the world,” he adds.
Q. What have been some of the key factors behind the growth of IBM India?
As of the second quarter of 2024, which we closed in June (we go by calendar year) we completed about 10 quarters of double-digit growth. We completed almost over three years of growth, consecutive growth.
I attribute a lot of this to three things. One is our strategy of hybrid and cloud AI, underpinned by a strong layer of security. Second, we pivoted to the higher value business that we had started to build, which included making people changes, having the right offerings, connecting with the right clients, among others. In 2021, we announced the separation of our managed infrastructure services business to Kyndryl. So, I think that focus has really worked for us in terms of growth. Lastly, operating as ‘One IBM’, and leveraging all the different parts, including a lot of our announcements with the government. A lot of this was possible because we can leverage our global teams that are doing cutting-edge research and innovation.
Q. What are the factors that have made India a global AI innovation hub?
About 59 percent of Indian enterprises reported they are actively doing something with AI, as per our Global AI Adoption Index 2024. And 27 percent are actively exploring using it. This is the highest among all the countries we surveyed. So, India has been ahead of the game. I’ve always found that in any emerging technology, India is sort of at the forefront. What has helped is the focus the government has had on emerging technologies through various ongoing initiatives. As we see growth in India, a lot of our domestic firms are trying to become global. It is in the DNA of Indian enterprises to be ahead of the rest.
The other important aspect is that India has a lot of great talent, that’s driving innovation. However, the skills gap remains the biggest barrier to AI adoption in India—limited AI skills and expertise was cited by 30 percent of organisations studied.
Q. What makes IBM’s platform watsonx standout?
watsonx is our data and AI platform. We launched it after a lot of deliberation, and with some key design principles in mind. First, it is a platform and not an application which has three key components. The first is watsonx.data, which is a mechanism to organise data to be used by AI for AI. But this is where you take care of ensuring you have the right data, it is organised properly, it’s being used properly, and so on. Second is watsonx.ai, an AI studio where you can use language models (LMs) from anywhere, including IBM proprietary models, but it is multimodal by nature, and you can build what is best suited to your business.
The third component is watsonx.governance, which ensures compliance with data governance, how your models are being used, how they are governed, etc. This platform is hybrid-by-design, which means you can use it on the cloud, on your private cloud, inside your firewalls and on-premise. This means it will take compute power, because large LMs (LLMs) take compute power, but it enables democratisation of whatever models you use.
The next is where this notion of small LMs (SLMs) came up. For example, you probably don’t need a trillion-parameter AI model to compose poetry, but you probably may need a lot more to enable real-time business analytics and computations. So, we started to build the Granite family of models, which are SLMs, use a lot less computing power, a lot more fit for purpose. We believe this is where the world is going, because you’re not going to need your Llamas and other LLMs for everything.
Q. How is IBM moving towards a more open-source ecosystem for AI?
We’ve taken a bunch of our Granite models—flagship series of LLM foundation models based on decoder-only transformer architecture—and moved them to open-source. We take care of data provenance and we indemnify it for copyright infringement. Once you make it open-source, we believe these models are going to become more and more democratised, which is why IBM has taken a step further in this direction.
Our hybrid cloud platform is built on Red Hat, a software company that provides open-source software and services for businesses. With Red Hat we’ve created an open-source project called InstructLab, which enables developers to pick up these models. It’s a platform for enhancing, extending, and driving more out of these models. We’ve also opened an InstructLab-based GenAI Innovation centre in Kochi, Kerala. This gives you a sense of how we are driving lower cost and higher democratisation.
Also read: IBM: A walk in the cloud
Q. Tell us about the plans IBM has in the area of semiconductors?
We got out of the fab and packaging business a while ago. However, I do believe today we are probably the world leaders in semiconductor R&D. We have established a state-of-the-art research centre in Albany, and the key to this is not just being leading-edge in what you are doing with semiconductor chip technology, but creating an ecosystem that can enable you to bring that to life. A bulk of our researchers are based in Albany, but there is collaborative work with India Research Labs as well.
Recently, IBM announced the world’s first 5 nm chips. At the Hot Chips 2024 conference in Palo Alto, California, IBM announced the next generation of enterprise computing for the AI era with the IBM Telum II processor and a preview of the IBM Spyre Accelerator. Both are expected to be available in 2025.
We are open to working with the right governments and players to provide knowledge and technology transfer. In 2022, for instance, IBM and Japan’s Rapidus Corporation announced a joint development partnership to advance logic scaling technology as part of Japan’s initiatives to become a global leader in semiconductor R&D and manufacturing. They brought us in as an R&D partner and they’ve created an ecosystem to have a sprint towards a 2 nm chip, which they want to start fabricating. Additionally, at Semicon India 2024, we announced our R&D collaboration with L&T Semiconductor Technologies to design advanced processors. This is the first that we’ve done, and we’re open to doing more.
Also read: L&T Semiconductor has a pipeline of 20 automotive clients so far: CEO Sandeep Kumar
Q. IBM is moving certain research jobs from China to India. What are your thoughts on this India versus China opportunity?
It’s not about India versus China. IBM adapts its operations as needed to best serve our clients, across different regions. IBM’s local strategy is focussed on having the right teams with the right skills to help them co-create those solutions.
Q. In the first quarter 2024 earnings report, there was a $400 million workforce rebalancing announced. Is AI the driving force behind this?
So globally, we have always taken the position of aligning our resources to where our client needs are, and that continues.
Q. How large is the skilling issue in India?
There are two aspects of skilling within AI that you have to keep in perspective. The first aspect is the need for technical skills. There is a whole new slew of skills that people will have to get trained on, such as prompt engineering, how to extend these models, and so on.
But there’s also the element of how you train existing people such as managers, mid-managers and other employees. Upskilling and getting used to having this technology as an augmented intelligence is also as big an issue, to which people will have to get acclimatised. It’s a change management scenario. We also need to remember it’s not rocket science. Both these skill sets need to proliferate.
There is an ongoing effort from the government and tech companies to ensure more skilling takes place. And we have been one of the pioneers in it. About six years ago we started to work with CBSE, where we introduced the AI curriculum in the final years of high school. In my first year here, the first cohort had just completed the course, and we showcased their final projects. I was so impressed! We bring in the winners for an internship, so they go through real-life training.
Through our IBM SkillsBuild programme, we are doing a lot of vocational training in remote and rural areas. For example, we conducted a programme in Nagaland where we taught schoolgirls to build drones. That is the kind of skilling that is needed at the grassroots.
Q. Apart from skilling, what’s the other biggest challenge with AI adoption?
A lot of CEOs are also talking about the ‘trust factor.’ Can we trust AI? And where does that trust issue come from?
Q. The data breaches…
Well, it’s not just the data breaches. So many infringement lawsuits have been filed. Where did the data come from? Did you really own the data that’s in your model? Are you ensuring the new data is owned by you? Do you know the provenance of the data you are using? These are becoming very fundamental questions.
This is why at IBM we have been very strong advocates of what we refer to as ‘Responsible AI’. For AI to be trustworthy, it has to be responsible—responsible AI is responsible business. Another premise that we should keep in mind is the concept of data stewardship. This essentially means, if you are my client, your data and your insights are yours. Data belongs to the creator. You should never take ownership of it. Additionally, whatever comes from AI should be trustworthy. IBM has put together “pillars of trust”, which include explainability, fairness, robustness, transparency and privacy. Putting all this together is when responsible AI becomes relevant.
So, the notion of data stewardship, ensuring the provenance of data is integral, and that integrity is maintained becomes critical in AI and that is going to become one of the biggest enablers to scale up AI.
Also read: Shaping India’s future with AI skills and jobs
Q. The third big vector that IBM is working towards is Quantum computing.
Quantum computing is the next frontier technology that is going to become relevant fairly quickly. India has been brilliant in terms of establishing the India Quantum Mission. What you need to understand is that quantum technology enables us to solve problems that today’s technology doesn’t allow us to. You can’t think about it as a hammer looking for a nail, which will solve everything.
Today’s technology works on bits and bytes. Quantum technology works on qubits. It’s a different form of programming. So, it enables you to solve more complex problems like mining research or risk models, and some clients are experimenting with that now. In India, we’ve been working with academic institutions to ensure the right cohorts are being trained in Qiskit, which is going to be the quality of the quantum programming language.
We have signed a memorandum of understanding (MoU) with the Centre for Development of Advanced Computing (C-DAC) and others, where we are providing a lot of thought around quantum research and what needs to happen to ultimately bring quantum capability to India. IIT Madras became the first Indian institution to join the IBM Quantum Network globally, where it has access to real quantum cycles on our quantum machines over the cloud, which it is using for exploring use cases and others.
Q. Where do you see IBM India in the next five to seven years?
I see us becoming more relevant as a technology organisation and innovation powerhouse in India’s journey. If India’s growth trajectory continues, we expect to continue to grow ahead of the market.
My objective would be to ensure we continue to transform and innovate ourselves to grow, be more and more relevant to where India is growing. For instance, we are consciously expanding to emerging markets, beyond metros in India, since there is growth of MSMEs all over the place. When you bring together both the availability of skills, as well as where there are market needs, I think it would be important for us to be where our clients are as well.
We opened two software labs over the last two years. One in Gandhinagar and one in Kochi. We’ve also opened delivery centres in Bhubaneswar, Mysuru, Coimbatore, and Gandhinagar. It’s not just about opening a software lab; we are setting up an IBM office with multiple work capabilities, depending on the skill sets we get. It also enables us to create a vibrant ecosystem around the region. I want IBM to be relevant to where India’s growth is, since that’s where we can help innovate and be leaders in innovating in India, for India, for the world.