As the global race to build artificial intelligence infrastructure accelerates, the biggest constraint may not be advanced chips but electricity. According to a new report by Shriram AMC, India’s strongest investment opportunity in the AI boom lies not in software or AI startups, but in power generation, transmission and electrical infrastructure.
The global AI boom has largely been viewed through the lens of semiconductor companies such as Nvidia and frontier AI developers like OpenAI, Anthropic and Google. However, Shriram AMC argues that investors may be focusing on the wrong bottleneck.
In its report, “The AI Bubble Debate: A Unit-Economics Lens,” the fund house contends that while GPUs account for nearly 60% of an AI data centre’s capital cost, the real limiting factor is power. A single AI-grade data centre with one gigawatt of capacity can cost $20-50 billion and consume as much electricity as 750,000 homes, making energy availability—not chip supply—the ultimate constraint on AI expansion.
Power, not chips, could decide the AI race
The report notes that the world’s four largest hyperscalers have committed around $1.08 trillion in AI-related capital expenditure between 2021 and 2025, while 2026 spending alone is expected to reach approximately $725 billion, up 77% year-on-year. Goldman Sachs estimates cumulative AI infrastructure investment could touch $5.3 trillion by 2030.
MUST READ: HDFC Bank bets big on AI with in-house GenAI platform ‘Neev’; plans to transform customer service, lending, operations
Despite this unprecedented spending, Shriram AMC believes the industry’s biggest challenge is no longer access to GPUs. Instead, securing sufficient electricity to power AI data centres is becoming the key hurdle.
“Whatever the resolution of the unit-economics debate, electricity will get consumed,” the report says, arguing that every AI model, regardless of which company develops it, ultimately depends on reliable power infrastructure.
India’s AI opportunity is infrastructure
Unlike the United States, India does not have listed frontier AI model developers or hyperscale cloud providers. As a result, Indian investors have limited direct exposure to companies building foundational AI models.
Instead, Shriram AMC believes the country’s investable opportunity lies in the physical infrastructure supporting AI.
MUST READ: AI assistants close in on search engines as India’s online discovery habits shift: YouGov
The report maintains an overweight stance on India’s power sector, including power generation companies, transmission utilities, transmission EPC firms, power financiers, diesel genset manufacturers, transformers, switchgear, cables and cooling infrastructure. These businesses are expected to benefit as global investments in AI data centres continue to expand, irrespective of which AI platform ultimately dominates the market.
The report also identifies electrical equipment manufacturers involved in data centre infrastructure as potential beneficiaries of the AI investment cycle.
A safer way to play the AI boom?
Shriram AMC argues that investing in AI infrastructure may be less risky than betting on individual AI companies.
The report says the debate over whether AI is a bubble should focus on returns on invested capital, not valuations. It estimates that the industry needs $600-650 billion in additional annual AI revenue just to generate a 10% return on current investments, while today’s AI revenue is estimated at only $50-150 billion.
Whether companies choose to rent computing power or sell AI services, demand for electricity remains constant.
MUST READ: India tops global AI talent, but trails in meaningful outcomes
The India investment case
According to the report, this makes India’s power ecosystem relatively insulated from the uncertainty surrounding AI business models. If AI demand continues to surge, electricity consumption and investment in transmission infrastructure are likely to rise. Even if AI monetisation takes longer than expected, many planned data centre investments are backed by cash-rich technology companies, meaning the build-out is expected to continue, albeit at a slower pace.
For Indian investors, the report concludes, the AI opportunity is less about identifying the next ChatGPT and more about owning the infrastructure that powers every AI model. In other words, the biggest winners from the AI revolution in India may not be software companies, but firms supplying the electricity, transformers, cables and cooling systems that keep AI running.
MUST READ: China’s AI price war is exposing India’s biggest weakness: Why India needs its own DeepSeek
#AI039s #biggest #bottleneck #isn039t #Nvidia #India039s #opportunity #software1783768706
