Macron and Modi turn on personal charm offensives as France and India race to secure AI investment

Macron and Modi are actively courting AI tech CEOs to secure major AI data center buildouts, including SoftBank’s plans for 3.1 GW in France by 2031 and a 75-billion-euro AI data center program (with France committing 3GW vs an initial 2GW). Modi highlighted Amazon’s record $48B investment in India, with $21B for AI and cloud infrastructure, alongside Microsoft’s largest Asia investment and Google’s $15B AI hub build. The reported surge of AI capital commitments is a constructive tailwind for AI infrastructure in both countries, though India’s reliance on foreign chips/models leaves it exposed to external export-control risk.

Analysis

This is less an AI demand shock than a competition for scarce inputs: power, permitting, and political access. The immediate beneficiaries are AMZN, MSFT, and GOOGL because sovereign buildouts create sticky workloads, but the earnings impact is back-ended; in the next 1-2 quarters the bigger effect is on capex guidance and multiple support, not revenue. The second-order winner set is the ecosystem around grids, cooling, and interconnects, while any local software hopefuls risk being squeezed out by the scale economics of hyperscalers.

SOBKY is the highest-beta expression, but also the one with the most hidden execution risk. A 3-5 GW buildout is only valuable if power delivery, financing, and utilization ramp cleanly over 6-18 months; otherwise the market will re-rate it from AI optionality to infrastructure overhang. India looks more likely to capture inference, data residency, and enterprise cloud than frontier model leadership, which means the economic leakage to foreign platforms will be high unless domestic chip and model capability improves.

The contrarian read is that the consensus is overpaying for sovereignty narratives. Tax breaks and charm offensives can pull capex forward, but they can also compress returns and are vulnerable to export-control changes or a slowdown in AI monetization. The key falsifier is simple: if these meetings do not translate into signed capacity, booked cloud revenue, or visible backlog within 1-3 quarters, this remains a sentiment tailwind rather than an earnings upgrade.

Key Decisions for Investors

• Buy AMZN/MSFT/GOOGL on 3-5% pullbacks over the next 1-3 months; this is a backlog-and-geography story, not a near-term EPS re-rate. Falsify if cloud capex keeps accelerating without margin support over the next two earnings cycles.
• Take a tactical long SOBKY only on follow-through strength and size it small; 6-12 month upside exists if the buildout milestones stick, but financing or power delays would turn it into a value trap.
• Relative short INTC vs long AMZN as the cleaner expression: Intel's India optionality is too remote to move earnings, while Amazon can monetize sovereign AI hosting sooner.
• Set an alert for any India/French permitting delay, grid bottleneck, or export-control escalation; those are the highest-probability 1-3 month reversal catalysts.