Picks and Shovels for India's AI and Data Centre Boom: The Smarter Way to Play the Theme

During the California Gold Rush of 1849, the men who made the most consistent money were not the prospectors digging for gold. They were the merchants selling picks, shovels, denim trousers and provisions to every miner who arrived whether that miner struck gold or went home empty handed. The suppliers won either way.

India's AI and data centre buildout is the Gold Rush of this decade. Hyperscalers like AWS, Google and Microsoft are committing tens of thousands of crores to Indian data centre capacity. Reliance, Adani and Tata are building their own. The government's digital infrastructure ambition is creating a Construction and power demand wave that will run for years. The question for investors is not whether AI grows — it will. The question is which companies get paid regardless of which AI application ultimately wins.

India does not yet have an OpenAI or an Nvidia. But it has the companies that will build the pipes, lay the cables, cool the servers, supply the power and construct the facilities. These are India's picks and shovels.

The Value Chain: Where to Focus

Break the AI and data centre ecosystem into layers. The further up the stack — AI software, SaaS applications, model builders the more competitive and uncertain the outcome. The further down the stack — power, cooling, cabling, construction, Real Estate the more certain the demand and the more durable the revenue.

The picks and shovels opportunity in India sits in four distinct categories: power and electrical equipment, cables and connectivity, cooling and HVAC infrastructure, and data centre real estate. Each has listed proxies with genuine exposure rather than just aspirational press releases.

Power and Electrical Equipment: The Most Certain Bet

A large-scale data centre consumes as much electricity as a town. A hyperscale facility running thousands of GPU clusters for AI inference requires uninterrupted, high-quality power at a scale that strains local grid infrastructure. Every data centre being built in India needs Transformers, switchgear, backup power systems and grid interconnection equipment.

This is the strongest picks and shovels category because the demand is not discretionary you cannot run a data centre without power infrastructure, and you cannot build power infrastructure without the companies that make the equipment.

The listed proxies here are Siemens India, ABB India, Hitachi Energy India, CG Power and Industrial Solutions and Cummins India. Siemens and ABB are established global players with deep transformer and switchgear portfolios. Hitachi Energy has specific exposure to high-voltage grid equipment that becomes critical as data centre power loads scale. CG Power has benefited structurally from India's broader capex cycle in electrical infrastructure. Cummins supplies diesel generator sets for backup power every data centre requires redundant power systems that activate within milliseconds of a grid failure.

On the renewable energy supply side, Tata Power and JSW Energy are relevant because hyperscalers globally are committed to powering their data centres with renewable energy. Long-term power purchase agreements with renewable developers is the mechanism and the developers who can offer 24/7 reliable renewable supply near major data centre clusters are positioning for contracts that last decades.

Cables and Connectivity: The Overlooked Layer

Data centres are not just buildings full of servers. They are buildings full of servers connected to each other, to the internet backbone and to end users through thousands of kilometres of fibre optic and copper cabling. The data centre boom is a cable boom.

Polycab India, KEI Industries, Apar Industries and Finolex Cables are the primary listed beneficiaries. Polycab is the largest organised player in India's wires and cables market with an expanding presence in industrial and infrastructure-grade products. Apar Industries has specific exposure to speciality cables and conductors used in transmission infrastructure the grid upgrades that data centre power loads will require. KEI Industries has been growing its institutional and infrastructure Order Book steadily.

The connectivity play extends to Tata Communications and RailTel Corporation. Tata Communications operates one of India's largest private submarine cable and fibre networks the backbone over which data centre traffic moves internationally. RailTel runs fibre alongside India's Railway network, with route diversity that makes it a natural candidate for data centre connectivity in tier-2 cities where land and power costs are lower.

Cooling and HVAC: 40 per cent of the Energy Bill

Cooling is not a peripheral concern for data centres it is approximately 40 per cent of total energy consumption and the primary determinant of whether a facility can operate reliably. A server room that overheats shuts down. A data centre cluster that overheats loses hyperscaler contracts. The precision cooling requirements for AI compute facilities which generate significantly more heat per rack than conventional server farms are more demanding than anything that has been built in India before.

Blue Star and Voltas are the primary listed plays. Blue Star has been growing its commercial and industrial HVAC segment steadily and has the technical capability to serve precision cooling requirements. Voltas, despite its consumer appliances profile, has a project business that handles large-scale commercial installations. Johnson Controls-Hitachi Air Conditioning India and KRN Heat Exchanger are smaller listed names with more direct thermal management exposure.

The cooling category is structurally underdiscussed relative to power and cables which means it may be less crowded in terms of valuation premium.

Data Centre Real Estate: Own the Land, Collect the Rent

This is the category closest to direct data centre exposure in the listed space. The model is straightforward own land in the right locations, build the shell infrastructure, lease to hyperscalers on long-term contracts. The hyperscaler brings the servers and the operations. The landlord collects contracted rental income for ten to fifteen years.

Anant Raj is among the earliest listed plays in this category with a genuine commitment rather than just an announcement. Its NCR land Bank gives it proximity to Delhi, which is emerging as a significant data centre hub alongside Mumbai, Chennai and Hyderabad. The company is building toward an integrated data centre ecosystem rather than simply leasing empty shells.

Macrotech Developers (Lodha) is making a serious push into data centres with Mumbai focus. Company plans to develop a capacity of around 2.5–3 GW, with an estimated investment of approximately Rs 1.3 lakh crore. Mumbai is India's largest data centre hub by existing capacity and by hyperscaler preference because of submarine cable landings that make international connectivity cheaper there than anywhere else in the country. Lodha's institutional capital relationships give it credibility in attracting anchor tenants.

Among the broader real estate names — Prestige, DLF, Brigade, Mahindra Lifespace data centre exposure is currently optionality rather than core business. They are worth monitoring but the thesis requires watching whether capital allocation follows the announcements.

Netweb Technologies deserves specific mention in the hardware category. It is the closest India has to an AI infrastructure hardware company building servers, workstations and HPC systems domestically. It is small, niche and carries the valuation premium of a pure-play theme stock, but the business is real and the addressable market as Indian data centres demand locally assembled server infrastructure is genuinely growing.

The Hyperscale Builders: Who Is Actually Building

The demand that makes every picks and shovels thesis real is already committed. These are the groups currently deploying capital at scale into India's data centre infrastructure.

Reliance's Jamnagar campus is being built with Nvidia chip infrastructure at a committed capital of USD 12 to 15 billion larger than most sovereign infrastructure projects. AdaniConneX has AWS as a confirmed anchor tenant at Chennai, meaning contracted revenue before the facility is fully operational. Nxtra Data under Bharti Airtel already commands approximately 25 per cent of India's colocation market and is scaling with USD 1 billion of institutional private equity backing from Carlyle.

Tata Communications owns one of India's most extensive submarine cable and domestic fibre networks — the physical layer over which data centre traffic travels internationally. Its listed status makes it one of the more direct ways to access data centre infrastructure exposure in India without going through real estate developers or power equipment manufacturers. Lodha's Palava Data Centre Park with 2.5 to 3 GW of planned capacity and AWS as a confirmed anchor is the most significant listed real estate play in this space by scale. Rs 1.3 lakh crore of committed investment into a single campus near Mumbai — with STT alongside AWS as anchor tenants — is not optionality.

These groups are not making announcements. They are deploying capital that commits them to years of construction, equipment procurement and power infrastructure buildout. Every gigawatt of capacity they build requires transformers, cables, chillers and grid connections. The picks and shovels companies sell into this demand regardless of how the competitive landscape among these builders eventually plays out. That is precisely the point.

The Framework for Choosing

The picks and shovels test is simple. Ask one question about any company: if AI in India grows ten times from here, does this company's revenue grow regardless of which AI application wins?

Power equipment — yes. Cables — yes. Cooling — yes. Data centre real estate — yes. IT services companies building AI applications — maybe, depending on which client wins. AI SaaS startups — dependent entirely on product success.

The companies that pass the first question are the ones worth building positions in. India may not produce the next large language model. But it will absolutely need transformers, cables, chillers and server halls to run whatever models the world produces.

That demand is not speculative. It is already in the order books.

Disclaimer: This article is for informational purposes only and not investment advice.