‘Sify is playing a critical role today with the AI infrastructure’: Aman Mehra, Head – Strategy & Transformation, Sify Technologies

Sify Technologies inaugurated an AI Lab & Experience Centre at Sify Noida 02 on August 19 in the presence of Alok Kumar, Principal Secretary, Planning, Information Technology & Electronics, Government of Uttar Pradesh. On the sidelines of the launch, Aman Mehra, Head – Strategy & Transformation, Sify Technologies spoke exclusively to Rajneesh De, Group Editor, APAC Media & CXO Media about the new AI data center and the critical role played by Sify in ensuring AI infrastructure for India.

What was the aim behind Sify setting up the AI data center in Noida?

Sify has launched this AI Lab & Experience Centre at Sify Noida 02. This is North India’s largest AI data centre campus, and it marks a significant step towards making enterprise AI more accessible and practical. This center is designed to bring together AI experimentation, industry use cases, skilling and hands-on engagement with emerging technologies such as Agentic AI.

This AI Lab & Experience Centre is positioned not merely as a demonstration facility, but as a platform to help build the skills, experimentation culture and enterprise ecosystem needed to translate AI ambition into practical adoption.

How does it differentiate from a regular data center in terms of its infrastructure and how would you describe its use case?

There are differences infrastructurally and there are differences technologically between the two. The time we are going to run any AI workloads, we call that an AI ready data centre. At that time, power consumption is going to be on the highest side. The time power consumption is going to be on the highest side, we need to have bulky servers. Once the weight of the servers increase, we need to design data center buildings looking at all these factors. 

Accordingly, the floor load bearing capacity of this Noida AI data center is 2100 kgs per square metre. In a standard data center building, this is going to be 400 kgs per square metre with the maximum floor load bearing capacity being 1500 kgs per square metre. In a standard data center building, the slab-to-slab height will go from 3 metres to 4 metres with maximum going till 4.5 metres. The power that can be used in this campus currently is 130 megawatt but this is scalable and if required, it can go till 180 megawatt. The objective is to ensure that as AI workloads grow, the underlying infrastructure is ready to scale with them.

Which are the verticals for whom specific AI cases will be demonstrated here?

We are going to showcase AI use cases to 260+ industry users in this very center where we are sitting. We have a platform called Sify Infinit Aizen where industry vertical use cases are already available for different industry customers to come here and experience and then eventually adopt. There are use cases say for banking, for insurance, for manufacturing, for healthcare, for retail, and for logistics. These use cases are either for businesses or customer centric services.

These use cases either come from our ISV partners or some of these are developed on our own platform on the Sify Infinit Aizen. There are multiple use cases of different industries but then it depends whether they are functional use cases. 

For example, you are offering a use case for HR where mass interviews have to be conducted. Now there are companies in retail space who do mass hiring. Logistics courier companies also do mass hiring. They have to interview 200 – 500 youth at the same time. Earlier it used to be done by human HR interviewers. Now the first round is conducted by the AI agent and the candidate at times does not even realise he is talking to an agent. This AI agent does the interview and the first level of scanning. This is a functional use case. 

Another functional use case can be of an insurance company involving claims processing and claims evaluation. Earlier you used to have humans to do that job but now because of the data of too many claims coming, the AI agent gets trained over a period of time. The AI agent then starts giving response of whether this claim has a deficiency or if it can be cleared. This is where insurance companies are benefiting. 

Overall, how would you trace Sify’s evolution as an infrastructure provider for different industries?

Today we have 170+ service providers coming into the facility. But when you talk about Sify as a group from a network experience standpoint, you have to trace our journey from 1998 from being the first privatised ISP to where we have come today. 

We have the latest in terms of our network infrastructure where we have created ultra-low latency multi-terabyte network across the country. We are one of the very few companies to have multiple open cable landing stations both on the Transatlantic and the Transpacific. We have one landing station in Mumbai, one in Chennai, one in Kolkata and one upcoming already announced in Vizag. These undersea or subsea cables are coming in to India. We are providing them an open cable landing station where they come and download. 

We as a service provider have also now created an inter-city network which is in-built across the country in 28+ cities. For example, we are sitting in Delhi NCR region where have our own fibre distributed starting from Manesar on one side, to Dadri on the other side, to Sonipat on the third side and to Muradnagar and onwards up to the fourth side. We cover this on a fibre network.

Now all this comes back and ties back to our data centres. That is the kind of mesh network which we have created and this is an advantage at an infrastructure level and it is very important.

In light of this critical infrastructure, how will you evaluate the successes of AI deployments?

I will come specifically to the lab and experience centre which is the subject of the day. But to make AI successful in any industry or anywhere in the world in any domain, what is most important is the foundational layer of the infrastructure. And that is where Sify is playing a critical role today with the infrastructure I mentioned.

You would also have been reading that a lot of AI PoCs are happening. But those PoCs are not translating into production because these are being done in a very defined secured environment or a perfect environment. When it comes to taking AI to the production stage, there are two points of failures. 

One is at the data layer or the data level which has to come from the client side. The data today is not very organised to train the AI agent. Therefore, when it comes about PoC it is a small data set which is taken to conduct that PoC. But the second very important layer which Sify is addressing today in the market is the infrastructure layer.

When you are taking an AI PoC to the production grade level to make that AI adoption successful, you need a very robust network. You need a very robust infrastructure layer in the form of data centres, in the form of private cloud environment or a sovereign cloud environment in the form of hyperscalers. This we also classically call the hybrid IT.

How has been Sify’s experience with setting up data center infrastructure in Uttar Pradesh?

The talent potential in Uttar Pradesh is enormous. Through our new initiative, we are setting a precedent to tap into this promising demographic. The aim is to build AI-readiness in the region as well as the local talent.

We built North India’s first and the largest data centre with Noida 01. Then we built the largest AI data centre which is Noida 02. We have recently come up with an edge data centre in Lucknow. The launch also reinforces Sify’s contribution to Uttar Pradesh’s growing technology and deep-tech ecosystem. The state government’s support and ease of doing business have enabled Sify to build and expand critical digital infrastructure in the state, including the Noida data centre campuses and the edge data centre in Lucknow.