Macron and Modi court US tech CEOs in the AI infrastructure race

French President Emmanuel Macron and Indian Prime Minister Narendra Modi are running parallel, CEO-by-CEO campaigns to land AI data center and cloud infrastructure, according to a CNBC report on the parallel charm offensives. Both leaders are trading on personal relationships rather than industrial policy alone, with SoftBank's 3.1 gigawatts committed to France, Amazon's $48 billion commitment in India, and a stream of US tech leaders taking meetings in both capitals.

The case for treating the two pushes as a single trend is the pattern. Macron personally convinced SoftBank CEO Masayoshi Son to expand the French project from 2 GW to 3 GW, according to Son's interview with CNBC, and the two exchanged text messages as the deal was hashed out over the spring. Modi met with Amazon's Andy Jassy last Thursday and welcomed the $48 billion commitment, and the same week hosted a stream of US tech leaders at the Global AI Summit in New Delhi. Both leaders are using diplomatic access as the lever that their domestic industrial base does not yet provide, and the resulting deal flow is starting to look like a template that other governments can copy if they can find a similar combination of power, demand, and a leader willing to make the personal call. The cumulative build-out now compares in scale with the CPP-EQT EdgeConneX $2.4B data center deal closed the same week, and the model is starting to look more like a general template than a one-off.

France's Macron bet: nuclear power and CEO access

For France, the bet is on power capacity. The country's nuclear-heavy grid gives it a structural advantage on baseload electricity, the constraint that has stalled AI data center siting in Germany, the UK, and the US mid-Atlantic. SoftBank's 75-billion-euro program, announced in May, will roll out 5 GW of AI data center capacity across Europe, of which 3.1 GW is committed to France. Macron also pulled OpenAI's Sam Altman, Anthropic's Dario Amodei, and Google DeepMind's Demis Hassabis into a working lunch with world leaders at the G7 in June, alongside Mistral CEO Arthur Mensch, Cohere's Aidan Gomez, Black Forest Labs' Robin Rombach, and Synthesia's Victor Riparbelli, and France has staked out the "sovereign AI infrastructure" pitch as a national priority. Mistral AI, the homegrown French frontier-model lab, anchors the demand side; the power grid anchors the supply side; and the personal CEO access is the bridge between them. The result, as the AI Infrastructure resource page frames it, is a country trading power capacity and political access for the kind of AI compute buildout that has so far landed in the US, Ireland, and the Nordics.

The SoftBank deal is the proof point. Macron asked Son for a meeting, personally persuaded him to expand the project from 2 GW to 3 GW, and committed to securing the additional gigawatt of power capacity on the French grid. Son told CNBC that Macron's team worked in close collaboration with SoftBank through the negotiation, and the two leaders kept in touch over text as the details were finalized. The pattern is the one France is now exporting: a single committed CEO meeting, a specific concession on power or land, and a follow-on invitation to the next summit.

India's Modi pitch: demand scale and engineering talent

For India, the bet is on demand scale and talent. India is a laggard on the supply side. The country does not yet produce advanced chips domestically, and it has no frontier-scale foundation model on a par with US or Chinese labs, which is why the recent global AI stocks rally has effectively skipped the country. The trade Modi is offering is: build AI capacity where the users are, not where the labs are. Google's $15 billion India AI hub announcement, Microsoft's largest Asia investment, Amazon's $48 billion, and ASML's commitment to supply Tata Electronics' 300mm fab together cross cloud, semiconductors, and energy, and the cumulative commitment is now north of $200 billion in the pipeline. The CEO courtship is the visible surface of a much larger effort to use India's demand base as a way to import AI capacity faster than India can build its own, and the pitch is built on the size of the user market, the depth of the engineering talent, and the willingness to negotiate country-level deals one CEO at a time.

The semiconductor deal is the test case. ASML's commitment to supply advanced lithography tools for the Tata fab is a signal that the world's most critical chip equipment vendor is willing to bet on the Indian buildout, and Intel's commitment as a prospective buyer for the fab output makes the build a real industrial play rather than a political gesture. The Google and Microsoft deals anchor the cloud side, and Amazon's $48 billion closes the hyperscaler triangle. The same playbook Macron used with Son is now being applied at much larger scale in India, with the prime minister doing the same personal outreach to a much longer list of US tech leaders.

The limits of the personal-pitch model

There are real limits to the personal-pitch model. Both Macron and Modi are betting that the CEOs they have courted will treat their countries as long-term homes for AI capacity rather than as one-off customers, and the recent track record on US data center siting suggests that the political risk in any single country can be priced in quickly. The other question is whether the pitch itself works at scale. Building 3.1 GW of data centers in France or standing up Tata's 300mm fab is a multi-year project, and a change of government in either country could reset the personal relationships that brought the CEOs in. The deal flow so far is encouraging, but the test is whether the AI infrastructure actually gets built, and whether the leaders who secured the commitments are still in office when it does.

The third question is what other countries will do with the template. Germany has industrial demand but a slower power transition. The UK has capital and language but limited grid headroom. Japan has its own version of the demand-side pitch and an explicit shortage of 789,000 software engineers by 2030. None of them have the same combination of power and demand that France and India are offering, but the personal-CEO model is now on the table as a workable playbook for any country that wants to land AI capacity. The most likely next adopters are the ones with at least one strong industrial asset to trade on, and the deals will continue to land with the leaders who can credibly make the personal call and back it with a concrete concession that the CEOs can take back to their boards.