Macquarie: India could emerge as an ‘AI powerhouse' and names top stocks to watch

Macquarie says India’s AI narrative is shifting, with over $400 billion flowing into the AI ecosystem (mostly data centers and energy); Microsoft and Amazon pledged >$50bn and Google $15bn toward India’s cloud/AI infrastructure. The report highlights 10 semiconductor projects totaling $18.2bn, Reliance’s $110bn AI infrastructure commitment, and a 12-stock ‘picks-and-shovels’ basket (including Reliance, Bharti Airtel, TCS) as early beneficiaries, plus private plays such as Sarvam AI, Zoho and Yotta (planning an IPO by early‑2027).

Analysis

Macquarie's narrative shift is less about India suddenly inventing a new stack and more about institutionalizing the non-glamorous infrastructure that underpins AI economics. Expect cloud-GPU demand to bifurcate: hyperscalers will monetize software/IP, while a broad supplier base (power, cooling, wiring, backup gensets, mature-node fabs, OSAT) captures durable, low-beta cashflows that compound over multiple 3–7 year capex cycles.

Second-order supply-chain effects matter: rapid hyperscaler builds stress local grids and create sustained demand for diesel/UPS gensets, large-format batteries, copper, and chillers — a structural uplift to industrial OEMs and upstream commodity suppliers rather than immediate fab-equipment leaders. Simultaneously, early-stage onshoring (mature-node fabs, packaging) will pull a slice of wafer demand away from advanced-node foundries over years, creating relative winners in legacy-node capacity and packaging/testing vendors.

Key risks are timing and concentration: GPU supply cadence (Nvidia cadence and allocation policy), policy/regulatory delays (land, water, incentives) and a macro growth shock could move a multi-year runway into a multi-quarter disappointment. A reversal catalyst would be sustained GPU scarcity or a renewed capital flight back to China/Taiwan driven by higher-risk premia; conversely, fast rollouts of sovereign models or a large local hyperscaler monetization event would compress time-to-payback materially.

For portfolio construction, treat Indian AI as an infrastructure-led reallocation opportunity, not a quick growth multiple re-rating. Use hyperscaler equities and GPU exposure to capture upside in utilization and monetization, pair that with industrial/energy hardware exposure for defensive, cash-flow-positive alpha, and keep allocation to private/state builds opportunistic and event-driven (IPOs, carve-outs).

Key Decisions for Investors

• Buy NVDA directional exposure via 12–18 month call spreads (e.g., buy calls financed with higher strikes) sized to 3–5% of the liquid options sleeve. Rationale: captures incremental GPU demand from new AI hubs; target 80–120% upside if Blackwell cycles through hyperscaler adoption, max loss = premium (100%).
• Pair trade (6–12 months): Long MSFT (5% tactical OW) financed by short or underweight TSM (equal notional). Expect MSFT to capture higher SaaS/AI monetization in India while TSM faces slower re-rating if capex shifts to regional mature-node builds; target relative outperformance 6–12% with symmetric downside if global foundry cycle re-accelerates.
• Buy industrial infrastructure exposure: accumulate GEV and CMI on 10–20% dips with a 12–24 month horizon. Risk/reward: downside ~15% on commodity/energy shocks, upside 20–35% as data-center power/cooling and backup power orders compound over multiple deployment waves.
• IPO/event play: monitor Yotta Data Services and Sarvam AI closely; set a watchlist and be ready to deploy up to 1% portfolio on IPO if valuation implies <8x EV / projected 2028 revenue from contracted AI compute (or if lockups/insider selling is light). This is an event-driven asymmetric beta with high upside if they secure long-term hyperscaler contracts.
• Hedge/contingency: maintain a 2–3% tail hedge in liquid China/Taiwan semiconductor exposure (e.g., short TSM or a Taiwan semi ETF) to protect against a rapid reallocation of global AI capital back to East Asia — acceptable cost if it limits drawdowns from a policy or supply shock.