2026.09.23 | TB | Datacenter | Pi Data Centers Expert

Primer • 28 Aug 2026

Companies: E2E IN

Expert Former C-level Executive

Reann Dsouza (Analyst) — Third Bridge Sector Analyst

Agenda

• Over eight years’ experience in the data centres space
• Over eight years’ experience at Pi Data Centers (left within the last three years)

Introduction

Analyst: If you could please state I agree or I disagree to the following statement: You understand the definition of material non-public information and agree not to disclose any such information, or any other information which is confidential, during the Primer Interview today.

Expert: I agree.

Questions

Could you describe what Pi Data Centers does and the products and services it provides?

Expert: They are basically a data centre company. It's been 10 years since they started. They have their location in a place called Mangalagiri. It's in a state called Andhra Pradesh in India. Their built-up area is about 2 lakh feet squared and they have about 1,100 racks on the floor. Their annual revenue would be, in terms of rupees, INR 0-250 crore. Their typical clientele industry-wise ranges across almost all segments, finance, education, medical, government and so on. Their largest clientele spread is in finance, fintech, small banking and other clients, primarily. They are also into certain manufacturing companies and 11-20% is in government. They have about 60 employees and they have headquarters in the city of Hyderabad, India.

What would you say is the value proposition of Pi Data Centers' products and services?

Expert: Primarily, co-location services. That'll be the biggest product area and revenue-gathering area. They're not into cloud. They're not into services as such.

Analyst: What do you mean by co-location?

Expert: It means that if you want to rent out, suppose if a company has some servers on their own and they want to put it in an area where it gets all the electricity, cooling and maintenance. They put it into one of these co-location centres and they run it from there. The role of the co-location centre is to maintain the uptime availability of those servers. The client maintains the remaining maintenance of the servers, builds applications on it, connectivity to connect to it remotely and all that.

What would you say is the full range of industries and business sizes that could use Pi Data Centers' products and services? What might be the company's TAM?

Expert: Mid and large would be the range. Small industries, it's not geared towards small segments. Mid-to-large and enterprise-grade clientele it would reach, from companies who would have a monthly invoicing ranging from close to INR 10 lakh-25 lakh to over INR 50 lakh per month billing rates. They would be able to afford that. Why does this matter? Renting out from cloud providers, the same infrastructure, would be very expensive and many of the co-location companies here aren't owned by one single entity. There are several entities that own and they take a part of it. They rent it again. It passes through many intermediary companies. That's not the case here. That is one of the attractions towards this company.

Analyst: Which industries might the company target?

Expert: Because they are a tier 4 certified company, which clears out a lot of banking regulations. They have several certifications and they have also banking customers. Banking, finance, fintech is a major reach for them. At the same time, they're into other manufacturing also because where the volume is high, they're able to handle that capacity also.

Could you profile Pi Data Centers' typical target customers across industries and business sizes? What might be the company's sweet spot based on other metrics?

Expert: The sweet spot is the cooperative banking sector because they are not too heavy on regulations and they're not too heavy on paying for high availability and all that, because they don't mind taking a downtime of five minutes or something like that. What happens is that when you go for the other regulated banking sectors, or say, fully regulated sectors need very high availability environments, which will be very expensive. When you have that sector, in terms of profits also, it is reduced. That's why they go after the semi-regulated banking sectors, where they're able to make them happy with the current requirements as well as there's a good profitability margin also in that sector.

Analyst: Could you provide the business split between cooperative banking vs other industries in percentage ranges?

Expert: 41-50% comes from cooperative banking, small banking and such people. Comparatively.

Analyst: Might there be a close second industry?

Expert: The closest second would be one level up from cooperative banking, the other private banks. That would be the next segment. Beyond that would be the manufacturing segment.

Analyst: How do you determine the business sizes Pi Data Centers caters to? What might make an ideal medium and large company?

Expert: Based on their annual revenue growth, they would target INR 500 crore-plus companies, which most of the banking sector easily achieve that goal because what they're looking at is good paymasters with the monthly billing. They relax up to three months for the billing, 45 days. Many clients pay up within the 60-day period, so that is one take on why they target those kinds of companies.

Could you share any customer names that are either on Pi Data Centers' website or in the public domain?

Expert: I remember cooperative banks [in] Maharashtra and a couple in south India also. They're pretty large. In the cooperative sector, they're pretty large banks. There are some banks that run in other cooperative sectors, like the retirement cooperative sector and the small, SME cooperative sector, they are also there. In the women's SME area, they have their own cooperative banks and all that.

What would you say are the key regions Pi Data Centers derives revenue from geographically?

Expert: Geographically, north India, like Maharashtra, Gujarat, Mumbai, Delhi, Pune, is higher than Hyderabad, Bangalore, Chennai and Kerala. 31-40% vs 61-70% north vs south because they're centrally located.

Might 91-100% of Pi Data Centers' revenue come from India?

Expert: Yes. 91-100%.

What would you say are Pi Data Centers' core revenue-generating product or service segments?

Expert: Core segment would be co-location services. They're also into building a data centre from the ground up. Greenfield build of data centres. They're also into building in-house data centres, like brownfield maybe.

Analyst: Does the company have any cloud or other managed services segments?

Expert: No. The clients build their own private clouds on their own hardware. They don't have a shared environment.

Analyst: Does the company have three major segments?

Expert: Yes.

What might be the relative revenue size of each segment – 0-10%, 11-20% or higher, up to 91-100%?

Expert: 51-60% comes in co-location services. 21-30% from [greenfield build] and 0-10% from [in-house build].

What might be the revenue model for each segment? Is it transactional, reoccurring or recurring for co-location services?

Expert: Recurring.

Analyst: How is the revenue model recurring for co-location? Could you elaborate on the contract?

Expert: Monthly billing.

Analyst: Is it a subscription every month?

Expert: Yes.

Analyst: What might be the revenue model for the greenfield projects?

Expert: This is transactional.

Analyst: Might the revenue model be the same for the in-house projects?

Expert: Yes.

Analyst: How is the price decided for co-location services? What factors come into play? Are there any services that are provided while the infrastructure is with Pi Data Centers?

Expert: Pricing is quoted in terms of per rack. INR 6 lakh-15 lakh a year or a monthly price of INR 10,000-15,000 per rack per month. Depends on the power it consumes. Depends on the amount of the size of the rack and all that. They come to that conclusion. They have managed racks and unmanaged racks. Sometimes customers bring their own men into the data centre to manage their racks. We provide that. Some other people will expect the data centre to manage that. That'll be the cost of managed services on that.

What might be a use case for each segment? Could you discuss a user journey of a customer starting with Pi Data Centers? What value and services are added to them? How do co-location services work on the back end until the end of the engagement?

Expert: Typically, the customer would want to host their infrastructure. They would give something called a bill of materials. Based on the bill of materials, Pi Data Centers will prepare a quotation to host them. What would it cost per month in terms of power, cooling and support? That is resource support. Then, there would be a few levels of something called requirements meeting criteria based on, we want 80% availability, 90% availability or 99% availability. Based on this, they will architect how the co-location should happen. They will start from there. Based on customers' cost, requirements, necessity, regulatory compliance and all that, they arrive at the final solution.

Analyst: What might be a use case for the greenfield projects? How do they work?

Expert: Greenfield basically is very much a non-IT project. It will be a civil contract build of the structure, engaging service providers and internet service providers, contracting with the electricity board of the locality for power and the water supply department for the water requirements and then all the permissions and all the clearances they need to get, whether they need to have a generator capacity with a diesel tank, clearances from fire department, planning for what size it should be, what is the availability of the local grids for electricity, manpower, all those things come into the picture. Based on the size, do they want 1-2MW up to 30MW or 60MW capacity of the data centre? They will size that sitting with the client and they have a team of experts who come into the meeting and begin the discussions. They move on to different areas and implementation. Many partners also come into the picture who have specialised departments. There would be a specialised diesel manufacturer who would come. There would be a UPS manufacturer who would come and so on. Once they build that, then the IT segment, how many racks of capacity did they plan for, installing the racks, giving the power to the rack all the way and then hosting the servers. The IT part, then again, if the client wants further into the IT, they will also provide services on maintaining those servers also at the client premises, wherever the data centre is built. That is the end-to-end greenfield approach.

Analyst: What might be a use case for the brownfield projects?

Expert: Brownfield is, suppose the client already has a data centre but is unable to manage that or has run out of capacity. That's where you come and you expand it or modernise it even. There are many modernisation projects there. You can modernise it, throw away all the old stuff and bring in the new stuff and then complete it to a client's specifications, basically.

Analyst: Is the design and architecture of the greenfield and brownfield projects completely handled by Pi Data Centers, or do customers come into play?

Expert: It depends. If the customer has got a very reputed architect for the physical building infra, he would want them. Then the design comes from us. It could be vetted by their known architect. To maintain certain tier levels, like tier 1, 2, 3, 4, there is a specification requirement that has to be met. Based on that, that's why we have our own teams. Basically, they engaged because they have been working with them for such a long time. They have their own teams who would, without wasting much time and without R&D, they would straight away go and build a plan.

Analyst: What do you mean by tier 4 data centres? What are the regulations?

Expert: That is very highly regulated. It has to meet a lot of requirements. What it means is the data centre can never go down. It's 99.999% availability. You should have fully redundant systems, like four diesel generators. Two of them fail. Two should pick up and if one of them fails, one should always be there. Redundancy is built into every part, from diesel generators to power, to racks, to servers, to the very end, very expensive in the end. There is the Uptime Institute, which is in the UK, that certifies them. They certify them all the way from the physical, all the way to the very last. They should meet certain criteria. Every year, they should renew the certification to maintain tier 4. That is why it is a double-edged sword, basically. You build a tier 4, but you are competing in the market with tier 3 plus segments where the cost would matter a lot. You already have a tier 4. There's past 10 years' experience, the clients really don't need a tier 4 unless they're hyperscalers wanting to rent from you. When hyperscalers come, they will handle most of the tier 4 requirements in a different way. That is why nowadays nobody is truly building a tier 4 data centre.

Analyst: How might customers build their own cloud? What might be the process? Does Pi Data Centers factor in anywhere during that?

Expert: Some clients want us to build a private cloud on their own servers. Basically, it's like virtualising the whole server and having more people, giving more environments to people than the physical server. One physical server can give about 15 virtual servers, for example, for 15 applications to run. That is nothing but the cloud. You're lifting up the whole physical infra into a virtualised environment. It could be compute, storage, network and now GPUs, basically, are all virtualised and presented so that more people can work with more environment reach. That is called cloud, basically. There are some automation features that come into the cloud which are restricted if they stay on the hardware itself. The platform layer, which we call PaaS, platform-availability-as-a-service, is nothing but the cloud layer, where if the client doesn't have the expertise, there is a cloud team that will plan and build for them. If it goes beyond the capability of this cloud team, there is a cloud build partner also who is engaged to build a private cloud for the client, especially when it comes to an AI cloud, which is a little bit complicated because there are GPU servers, there is capacity planning and there is something called GPU slicing and then the containerisation and all those things come into the picture, which is readily available in a public cloud. Due to certain restrictions, people want an AI cloud in a private cloud, in a private environment.

Which of Pi Data Centers' three segments do you anticipate being the fastest growing and potentially driving the largest revenue share in future?

Expert: Right now, co-location would be their biggest bet. Both greenfield and in-house builds are also coming, but the problem is that, with both, the second two segments, like greenfield and in-house, many requirements come from state governments, like the Assam government or remote areas, remote states, because they have already built a state data centre here and many governments approach them to replicate that. For that, what happens is the process of tendering and putting forth something called surety amounts, all those is a huge process, which really doesn't make business sense, so they don't take much. Though there are many such opportunities, they don't really go after them. Their focus is more on co-location, which is simple, fast, continuous generation of revenue, no headache and only expansion, basically. One room gets over, the next room gets filled and then they expand and all that.

Analyst: Does the company increase area when customers come in? How is that managed?

Expert: Yes. As they come in, they have a certain room for expansion. They're also planning to build the physical build because they're seeing something in the pipeline. They are able to expand that also and they have a lot of land available. They can keep on expanding to that because it's a flat data centre. It's not a floor-by-floor thing that you see in many cities because they're in a tier 2 city, they have a lot of land available. They can expand horizontally rather than vertically. That's why they do that and so they're able to go after that expansion.

Could you specify the industry that best describes Pi Data Centres' core operations?

Expert: Any company that needs co-location services would be the industry.

Analyst: Which industry does Pi Data Centres belong to?

Expert: Data centre industry.

What has been the data centre industry's AAGR over the past three years – less than 0%, 0-5%, 6-10%, 11-20% or more than 20%?

Expert: More than 20%.

What do you think have been 1-2 key drivers of the industry's growth over the past three years?

Expert: AI growth has pulled a lot of other requirements to grow the industry as such. AI drives this growth and hence, requirements for data centres have grown exponentially. I would attribute the main factors to AI and the solutions and the industry growth around that.

What do you think the market's AAGR will be over the next three years in a percentage range?

Expert: Greater than 20%.

What do you think will be the key drivers of the market's future growth?

Expert: Same thing, AI expansion.

Analyst: Might there be other key drivers? Are there any regulations, changes in the environment, etc?

Expert: Not regulations, but as the public cloud players, the hyperscalers become more expensive, the smaller players, the need for them will grow. Even investors will be willing to put more on the smaller players.

Analyst: Do you think any specific area within AI will increase or transform the data centre industry?

Expert: It's all got to do with the GPU servers and GPU infrastructure. That's the biggest driving unit. The more GPU servers are coming out and being installed and provisioned, keeps growing exponentially. That needs a lot of cooling, a lot of electricity, more floor space, expanded floor space and a lot of environmental effects, that kind of growth. That'll be the area.

To what degree do you think the industry is correlated to economic cyclicality? Is it counter-cyclical, very cyclical, somewhat cyclical, somewhat resilient or resilient, and why?

Expert: It is a resilient market now. It's no longer cyclical.

Analyst: What are the reasons for it being resilient? What must have changed over the past years?

Expert: Because the expansion of AI is what has changed. It's an upward trajectory, basically. Earlier, it was on a more of a semi-cyclical, like there were ups and downs. Whenever some government regulations change and one sector saw growth, it would fluctuate that way. But now AI is impacting every segment, every sector. That's why it's exponential growth all around, basically.

Based on public information, what impactful technological evolutions or innovations can we anticipate in the industry over the next 3-5 years?

Expert: It would be the proliferation of AI and AI technologies, GPU servers and the way they are being used by various AI models and so on, becoming more and more in the market. More and more requirements for physical infrastructure and virtualisation techniques and innovation in general. The more the innovation is, that person will be upfront. He will have an edge over the others. For that innovation, you need to have a lot of capacity to depend on research and development. That means the capital that you put in may not get immediate returns. Anyone who has that vision to wait for the returns would eventually have a better advantage.

Are there any downside risks you believe might materially impact market growth over the next three years?

Expert: I would see a shortage of hardware as the biggest risk right now. Delays in hardware. What is happening is that as AI technologies run away, the hardware is unable to cope up and is playing a catch-up game, which will result in higher cost and availability of order, which will impact the cost of the environment in which AI is going to run, which will impact the AI products that are being developed. Unless that is met, the next three years would see that as a huge risk.

How fragmented would you consider the data centre industry to be? Is the majority of the market share captured by 2-5 strong players? Is it a monopoly? Is it fragmented or highly fragmented?

Expert: It's fragmented at this stage.

How protected is the market? What would you say are the barriers to entry and to scale in the market?

Expert: Barriers to entry are the lack of knowledge of how to run a data centre. That is the biggest thing I have seen. Very big companies have also failed in managing data centres. It's an industry by itself. It cannot be an also-have thing. It's not a protected market. Anyone can enter it. Lack of knowledge is what leads to many people not making it.

Analyst: What are the barriers to scale? Is knowledge also a barrier to scale?

Expert: No. Scaling is hugely dependent on capital and physical land availability. That is the biggest scaling. Many people would expect faster return on capital. Many people would find the acquisition of land to be more expensive and not feasible.

Who would you say are Pi Data Centers' top three competitors?

Expert: I don't have the current exact comparison, but as far as my knowledge goes, there is Yotta, one of the biggest competitors. There is E2E Networks and then there is BharathCloud, there is CtrlS, which would be the direct competitors at this stage.

Analyst: Might Yotta Data Services, CtrlS Datacenters and E2E Networks be the company's top three competitors?

Expert: Yes.

What would you say are Pi Data Centers, Yotta Data Services, CtrlS Datacenters and E2E Networks' market shares – less than 5%, 5-10%, 11-20% or higher, up to more than 90%? Alternatively, could you rank them?

Expert: Yotta would be one. CtrlS would be two. E2E will be three. [Pi Data Centers, four].

What were Pi Data Centers, Yotta and CtrlS's revenues for the last financial year – INR 0-250 crore, INR 250 crore-500 crore, INR 500 crore-1,000 crore, INR 1,000 crore-2,000 crore or more? E2E Networks is publicly traded.

Expert: [Pi Data Centers], INR 0-250 crore. [Yotta] would be INR 500 crore-1,000 crore. [CtrlS], INR 250 crore-500 crore.

What were Pi Data Centers, Yotta and CtrlS's revenue growth rates for the last financial year – less than 0%, 0%, 0-5%, 6-10%, 11-20% or more than 20%?

Expert: [Pi Data Centers], 0-5%, I would say. [Yotta] would be more than 20%. [CtrlS], 11-20%.

What has been each player's growth driver for the last financial year? Has the growth been organic, acquisition-based or mixed?

Expert: [Pi Data Centers], organic growth. Yotta is mixed, basically. [CtrlS], also mixed. [E2E], also mixed.

In your opinion, what is the average EBITDA margin for the players – less than 0%, 0-10% or higher, up to more than 60%?

Expert: It would be 21-30%.

Would you plot each player above, in line with or below the 21-30% industry average?

Expert: [Pi Data Centers], I would say on par. [Yotta], they're above industry. [CtrlS], they're also above industry. [E2E], they're also above industry.

How do you think each player differentiates itself? What might be each player's USPs? What does Pi Data Centers do differently?

Expert: Unique selling point would be single-owned and strategically positioned, from a geolocation.

Analyst: What is Yotta's differentiator?

Expert: Yotta's differentiator is that their AI expansion would be the biggest.

Analyst: What is CtrlS's differentiator?

Expert: CtrlS's biggest point is their physical build capability, like building data centres.

Analyst: What is E2E's differentiator?

Expert: E2E Networks, their biggest thing would be non-AI cloud. Traditional cloud, that is.

Analyst: Do any players have a strong hold in any sectors or industries? You mentioned Pi Data Centers focuses on the financial or banking sector.

Expert: No. They don't.

Analyst: Are the other players covering all industries?

Expert: Yes.

Analyst: Do all players offer co-location and building, or does any player have more or fewer services than that?

Expert: More or less the same.

Analyst: Do any players get into cloud?

Expert: Yes. Except for CtrlS.

Analyst: Do any players position themselves as more premium or cheaper than the others?

Expert: Yotta is premium, but they're cheaper internationally. CtrlS, it will be almost the same. Pi Data Centers also play in the same game as being competitive.

What would you say are customers' top three purchasing criteria? How would you rank them by importance?

Expert: Cost would be the first. Service delivery and location.

How important would you say price is to customers in the market? Is it commoditised, somewhat commoditised or not commoditised, and why?

Expert: It's commoditised. The reason is it's a price-conscious market, basically. No one wants to pay much and they are able to live without it.

What do you think has been the key driver of Pi Data Centers' historical growth?

Expert: Their focus on the banking sector and government would be the driver.

Analyst: Are there other factors, eg, services that the company has been doing?

Expert: They have been giving very specialised services committed to the clients, so they like that.

How fast do you think Pi could grow annually over the next three years, in a percentage range?

Expert: They would grow at 0-5%.

Analyst: Why would you plot the company much below the more than 20% industry growth?

Expert: Basically, they need to have a lot of investment capital to expand and modernise the products they have. Like, for instance, they don't host any AI servers, at least to the best of my knowledge. They should be able to make some changes in their environment to accommodate or even have some customers with their servers. Until and unless they do that, they won't reach that 20%.

What could be some of the most interesting ways to grow Pi Data Centers organically if you were running the company?

Expert: I would say they should change their product line and not depend solely on co-location. They should go more with AI cloud and all that. Bringing a lot more investment into hardware and modernising their product line and getting on par with what others are doing. They have lost the race because doing it now would involve very large amounts of investment.

Analyst: Are there any industries that the company could target?

Expert: Yes. They could target the physical build and government contracts. That would be coming up a lot. A lot of small data centres want an experienced person to build, manage and give away.

Analyst: Do you think that it makes sense for the company to get into cloud servers or other technology-based services?

Expert: Yes, that's what I said. They should get into modernisation and expand the product line into cloud.

What would you do to improve Pi Data Centers' margin and operational efficiency?

Expert: They should have more automation in place that'll give better client services. They can take up advanced client requirements, which they can't do now. They can do that.

What would you do or change for Pi Data Centers via M&A?

Expert: There are a lot of chances for them to do M&A, acquire small companies that specialise in data centre products and then cloud products and AI technologies. Instead of building it from scratch, they can acquire smaller companies.

Analyst: Are there any company names or specific areas that the company could benefit from?

Expert: No. I don't have the names, but [earlier], they were looking at security-based start-ups. That was a huge requirement.

Based on publicly announced acquisitions, how would you describe Pi Data Centers' M&A strategy? Has the company carried out any M&A in the past?

Expert: Not much.

Analyst: Why has the company not gone ahead with any M&A?

Expert: Synergies were not very feasible, I should say.

In your opinion, could it make sense for Pi Data Centers to acquire, merge with or be acquired by any players in or outside the space?

Expert: Definitely. There have been many people who wanted to have a majority stake. Their valuation wasn't on par with the expected valuation in them being acquired.

Analyst: Do you think that being acquired might be a possibility? What might make most sense?

Expert: Yes. They're actively looking to have a majority stake given to them.

Analyst: Which companies is Pi Data Centers looking at to be acquired by?

Expert: No idea. That keeps changing.

Analyst: Might the company be open to investment or be acquired by a corporate company?

Expert: Either way.

What would you say are the 1-2 biggest risks that Pi Data Centers faces?

Expert: They are very risk-less. They don't have debt. They don't have any major outstandings or anything like that. The client churn is there, but still new clients keep coming. I wouldn't give any major risks.

Analyst: To confirm, are there no internal risks to the company?

Expert: No.

In your opinion, is there anything the market does not properly understand about Pi Data Centers? Are there any common misconceptions about the company?

Expert: Yes. Location is one huge misconception, which they have overcome a lot this past 10 years. Cost, another one. These are two primary ones.