Google’s India AI hub cleared for 2.51GW, not 1GW, the Guardian reports

Google's $15 billion AI datacentre project in Andhra Pradesh has received state environmental clearance for 2.51GW of capacity, according to clearance documents cited by the Guardian. The approved capacity is more than double the 1GW initially announced, indicating a substantially larger potential buildout of AI computing infrastructure in India. The expansion could materially increase regional power demand and reinforces Google's investment in AI data-centre capacity.

Analysis

The relevant equity implication is not the project’s headline capex but the probability that Google is securing power, land and permitting ahead of a materially larger inference footprint. If utilization ramps, incremental capacity improves Google Cloud’s ability to sell AI workloads in a power-constrained global market; however, the near-term accounting effect is adverse to free-cash-flow conversion as servers, networking and power infrastructure are funded well before revenue utilization. For GOOG, this is supportive of the durability of Cloud growth over 12-36 months, but unlikely to change the next one or two quarters absent revised capex guidance.

The more actionable spillover is Indian power procurement. A multi-gigawatt baseload load increases the value of dispatchable generation, transmission access and renewable-plus-storage contracts; it may tighten regional power economics and expose the project to commissioning delays if grid upgrades lag. Suppliers with exposure to hyperscale electrical equipment—ETN, VRT and GEV—are better positioned than semiconductor names already priced for broad AI demand, although the project-specific revenue allocation is not yet independently quantifiable.

Consensus may treat overseas AI datacentres as evidence of unconstrained capacity expansion. The constraint is increasingly electricity delivery rather than chips: delays in interconnection, water availability, or local opposition can defer revenue while capex continues. That asymmetry argues against chasing GOOG on this item alone; the tradable confirmation is a step-up in Alphabet capex guidance paired with Cloud backlog/RPO acceleration, not permitting documents.

Key Decisions for Investors

• Maintain a 12-18 month overweight in GOOG versus META rather than add outright on the clearance: Google gains AI capacity optionality, but require Cloud growth to reaccelerate and capex-to-revenue conversion to remain stable before increasing exposure.
• Watch for Alphabet’s next earnings call: add GOOG only if incremental capex is accompanied by Cloud revenue growth above consensus and management identifies capacity as a binding constraint. Falsifier: a material capex increase without Cloud margin or backlog support, which would raise multiple-compression risk.
• Build a 6-12 month basket watchlist of ETN, VRT and GEV for disclosed India hyperscaler orders or raised backlog guidance; do not underwrite project revenue before supplier awards are confirmed. Prefer ETN for grid/electrical exposure and VRT for datacentre power-and-cooling intensity.
• Use a GOOG/META relative-value framework over the next 1-3 months: favor GOOG if investor concern shifts toward Meta’s open-ended AI infrastructure spend while Google demonstrates third-party Cloud monetization. Exit the relative trade if Google’s capex guide rises faster than Cloud growth or regional power delays become public.