Liquid cooling is the biggest game changer as less power consumption will make AI solutions more affordable: Amit Luthra, MD-ISG, Lenovo India

In a recent interaction with Express Computer, Amit Luthra, Managing Director – ISG, Lenovo India, discusses Lenovo’s integration of AI and machine learning into its infrastructure solutions. Luthra emphasises Lenovo’s success in various sectors like manufacturing, BFSI, and retail, and the company’s focus on new-age workloads such as AI, cloud-native applications, and edge computing. He highlights the growing importance of AI-driven systems, the role of hybrid cloud solutions, and Lenovo’s efforts to differentiate itself through reliability and cutting-edge infrastructure offerings. Luthra also touches on the challenges and opportunities in India’s expanding data centre market and Lenovo’s strategy to meet the country’s growing digital demands.

How is Lenovo integrating AI and ML into its infrastructure solutions to cater to the growing demand for intelligent systems? What specific use cases or verticals are you focusing on?

So, the use cases we are focusing on, if I go back and reflect, we have success stories in manufacturing, banking, financial services, and insurance. We also have success stories with cloud service providers and in retail. We’ve achieved success in various sectors, including network equipment providers, oil and gas, and so on.

Where are we succeeding? If I break it down, organisations typically have traditional workloads and cloud-native workloads. Cloud-native workloads are where new spending is happening. In traditional workloads, we’ve been very large in SAP, ERP, and collaboration tools like Microsoft.

However, for new-age workloads, especially in AI, our approach is to go back to organisations and customers, along with our channel partners, and offer a complete stack. AI doesn’t happen just by selling a GPU server or a server for inferencing. It needs software, applications, services, and more. Often, these capabilities are brought to the table by us, through OEM partnerships, local channel partnerships, or ISV partnerships. It’s about driving outcome-based discussions.

Do you think that with a growing demand for AI, the relevant skill set is a big challenge?

Skills, if you ask me, will continue to be a significant challenge across organisations. I also face this challenge of retaining skills; it’s not just about bringing in new skills. It’s equally about keeping the existing ones. Now, this is a challenge because you want to move into areas that are outside your comfort zone. You’re trying to accomplish tasks that have never been done before, which involves a learning process.

The positive aspect about India, our channel partners, and my employees is that their attitude is aligned correctly. This means the learning cycle is quicker because the fundamentals are strong. If your fundamentals in IIT, software, and services are solid, the learning process won’t take much time.

I believe the skills gap is evolving, largely because outcome-based requirements are also changing. No one can claim that we have fully defined AI; there is no fixed template for it. New templates and use cases are emerging every day. Since the foundational skill sets available in India are very robust, I think the learning curve is much steeper.

Do you believe that potential temporary job losses can be mitigated through reskilling and upskilling of current employees, making the gap more favourable?

There is a job for everyone in the AI space because, at the end of the day, you need people to operate these machines. There may be certain jobs that are labour-intensive, which machines might take over due to machine learning. It’s about reusing the same processes—what ABC did, now XYZ will do. However, jobs will not be lost in this process.

Given the increasing adoption of edge computing, how is Lenovo positioning itself as a provider of edge infrastructure solutions? What are the key challenges and opportunities in this space?

I believe edge computing is the biggest game changer. Today, if you want to perform analytics or implement AI, you need to consider where to do it. For time-sensitive data, such as when visiting a retail outlet to enhance the product experience, immediacy is crucial. Whether you’re on an e-commerce platform or in a showroom, you want that experience right away. Thus, for any time-sensitive analytics, edge computing becomes very important. 

Conversely, for analytics that are not time-sensitive, like traditional workloads, data centre analytics play a significant role. It’s not that edge computing will replace data centres or vice versa; rather, both will coexist hand in hand. Time-sensitive data necessitates edge analytics, which in turn requires high-end Wi-Fi 6 and 5G infrastructure.

What we have done is create a purpose-built AI infrastructure for edge computing, which includes GPU cards that allow for running hybrid clouds and containers. These systems can operate in harsh environments, with temperatures ranging from minus 10 to plus 60 degrees Celsius. You don’t need to have a dedicated facility; you can deploy this infrastructure on a shop floor, in a refinery, or anywhere else where time-sensitive analytics are needed. 

For instance, in a manufacturing line, you can monitor IoT sensor data to assess the quality of products and reduce defect rates. The answer to whether we can build an edge facility with GPU capabilities is yes. This is where we can have a comprehensive discussion, connecting the edge to the cloud.

Lenovo has been actively promoting hybrid cloud solutions. Can you elaborate on your strategy and vision for hybrid cloud adoption in India? What are the key benefits and challenges of this approach?

Hybrid is definitely the way forward. For example, there are certain workloads that are suited for the public cloud. These workloads have matured to the point where companies package them as products, typically as software as a service or a consumption model. 

However, there are new-age workloads, such as machine learning and deep learning, that require a different approach. High-performance computing tasks, like genome sequencing and simulation, will need to remain in your data centre. 

When we consider why hybrid is the way forward, it’s clear that some workloads will continue to reside in the public cloud, while others will stay in data centres. Therefore, organisations cannot adopt a one-size-fits-all approach or rely on a single solution. This is how the industry is evolving.

Can you discuss Lenovo’s latest data centre solutions, including servers, storage, and networking? How are these solutions differentiated from competitors, and what are the key value propositions for customers?

I would say that differentiation comes from reliability. This is our ninth year in operation, and we are rated the best in reliability. Differentiation also comes from benchmarking. If you take any SAP benchmark, we have the highest number of these benchmarks available. The difference lies in our ability to run the top supercomputers. If you look back, we are the vendor that operates the highest number of supercomputers in the world. 

That’s where our differentiation comes from; we present material differences compared to our competitors in the industry. Regarding our portfolio, as I mentioned, we have a complete range for edge computing, including edge PCs, edge servers, edge GPU servers, and edge cloud servers. We offer a full gamut of compute platforms, such as tower servers, regular rack servers, AI servers, and dense servers. Additionally, we have a very strong portfolio in storage and hyper-convergence. We work closely with ecosystem vendors on the network side, which enables us to deliver full-stack capabilities to customers through our channel partners.

What do you see as the main challenges for expanding data centres in India, given the projected data structure demand of around $10 billion by 2027 and the fact that 20% of global data flow is through India, but the country currently has only 3% data centre capacity?

I don’t think it is a challenge; I believe it’s more about being a little late. Everyone is gearing up for it. In just a year or so, India will be the most lucrative destination.

Can you quantify what is the untapped market in the data centre industry in India?

The untapped market for data centres in India aligns closely with the untapped market for PCs and mobile devices. As soon as every Indian has a PC or mobile phone in their pocket, the demand for data centres will increase exponentially. This is because applications will require infrastructure to operate, and that infrastructure must be supported by data centres—whether public, hybrid, colocation, or private. Therefore, I believe that as device penetration increases, the demand for data centres will grow significantly.

Beyond AI and edge computing, what other emerging technologies are you keeping an eye on? How is Lenovo preparing to capitalise on these opportunities?

One definite technology is liquid cooling, as it is essential for managing power consumption, which is becoming a significant challenge in the boardroom that needs to be addressed.

You have suddenly increased the demand for power with this kind of infrastructure, so the question becomes: how are you going to supply it? That’s where our sixth generation of liquid cooling comes in. We’ve extended liquid cooling to our workstations as well.

We have a legacy of implementing liquid cooling in mainframes since our IBM acquisition, and we are now applying that to our GPU systems. We believe liquid cooling is the biggest game changer because it will make AI solutions more affordable.