In 1850s Assam, British planters secured rent-free land from a cooperative colonial government, built an export industry on local soil and labor, and sent the profits home. The tea was grown in India; the wealth it generated largely was not. It is a pattern Indian economic historians have spent a century and a half documenting, and it is worth remembering now that Google has announced a $15 billion artificial intelligence hub in Visakhapatnam, on India's eastern coast, and found itself defending the project against water shortages, land disputes, and street protests.

Google says the facility, due to open in 2028, will use air cooling technology to limit water consumption, a direct response to local fears that a data center of this scale will strain a region's drinking and irrigation supplies. That is a real concession, and it should be credited as such. But the dispute over Vizag is not really about chillers and cooling towers. It is about who controls land, who absorbs environmental risk, and who captures the value once the servers start humming. These are not new questions. They are the oldest questions in the economic relationship between the Global North and the Global South, simply wearing new silicon.

The Plantation Logic, Rebooted

Bipin Chandra's history of colonial India describes how foreign-owned plantation and industrial enterprises operated with government favor while Indian businessmen struggled for credit on equal terms. In 1914, foreign banks held over 70 percent of all bank deposits in India; British manufacturers lobbied colonial administrators to actively discourage Indian industrial development because they viewed local textile producers as rivals. The profits of foreign enterprise, Chandra notes dryly, "went out of the country."

The AI data center boom does not involve the East India Company's sepoys or the Raj's land revenue officers. But the underlying structure, capital and technology controlled from abroad, land and natural resources supplied locally, and much of the surplus value repatriated, rhymes uncomfortably with that older arrangement. An AI data center is a facility optimized for the extraordinarily resource-intensive work of training and running machine learning models. That intensity is precisely the point of contention in Vizag: these are not light-footprint software operations but massive physical consumers of land, electricity, and water, built in a water-stressed coastal region whose residents were not meaningfully consulted on the trade-off.

None of this makes Google a colonial occupier in any literal sense. India is a sovereign democracy with courts, elections, and a free press, none of which existed in British Assam. But the economic logic of extraction, take the resource where it is cheap and abundant, process the value where capital and expertise already concentrate, has proven remarkably durable across two centuries of technological change. It survived the transition from indigo to tea to textiles. There is no obvious reason it would not survive the transition to large language models.

What Decolonization Actually Solved, and What It Didn't

The wave of decolonization that followed the Second World War was supposed to end this dynamic. Japanese victories over British, Dutch, and French forces in Asia shattered the myth of European invincibility, and colonized peoples who had fought to expel one occupier were in no mood to welcome back another. Between 1946 and 1995 the number of sovereign nation-states nearly tripled, from 74 to 192. India itself gained independence in 1947.

But sovereignty and economic independence turned out to be different things. Kwame Nkrumah, Ghana's first president, coined the term "neo-colonialism" to describe what replaced formal empire: Western nations and the United States retaining economic control over former colonies without the burdens of actually governing them, what he memorably called "power without responsibility." The 1973 Non-Aligned Movement conference in Algiers, where newly independent states declared themselves a "Third World" seeking neutrality between Cold War blocs, was itself an acknowledgment that political independence had not delivered economic parity. By 1989-90, the average GNP per capita of the industrialized North stood at more than 24 times that of the developing South, a gap that decades of aid and development theory failed to close.

Vizag is a test of whether the AI era will repeat that pattern or genuinely break from it. Andhra Pradesh's government clearly hopes the $15 billion investment brings the kind of durable, high-wage employment and technology transfer that eluded earlier waves of foreign capital. Local farmers and residents disputing land acquisition, and activists raising alarms about groundwater, are effectively asking the question Nkrumah asked in 1965: will the money and the know-how actually stay, or will the profits, like the indigo and tea profits of the 1850s, flow back out while the land and water bear the cost?

Who Adjudicates When Giants and Communities Collide

One of the quieter lessons of the postwar order is how weak international mechanisms are at resolving exactly this kind of asymmetric dispute. The International Court of Justice at The Hague has successfully settled relatively narrow interstate matters, a Dutch-Belgian frontier dispute, an Anglo-Norwegian fishing disagreement. But when the stakes involve a powerful actor unwilling to yield, the Court's authority evaporates. In 1984, Nicaragua successfully sued the United States over the mining of its harbors and won a judgment in its favor; Washington simply refused to comply, and no enforcement followed. There is no tribunal, in other words, where a coastal community in Visakhapatnam could meaningfully compel a trillion-dollar technology company to renegotiate a land deal. The leverage, as in the colonial era, runs overwhelmingly in one direction.

This is not an argument against foreign investment, which Andhra Pradesh's government has actively courted and which could genuinely transform the regional economy. Niall Ferguson has argued, controversially but not without evidence, that British imperialism left behind institutional benefits: free trade norms, rule of law, infrastructure. Foreign capital is not inherently extractive; context and terms determine the outcome. A data center built with enforceable local water-sharing agreements, meaningful employment quotas, and transparent land compensation is a fundamentally different thing from one built on opaque acquisitions and unmet promises.

Compare this to how Gulf monarchies, facing their own legitimacy pressures during the 2011 Arab Spring, responded with rapid, tangible concessions. Saudi Arabia, confronting only mild unrest, rolled out a reform program within weeks: 60,000 new security sector jobs, a raised minimum wage, half a million new apartments. It was an authoritarian government using oil wealth to buy social peace quickly and visibly. Google, a private company rather than a sovereign state, has less unilateral power to do this, but the Vizag protests suggest that visible, fast, and credible concessions, not just cooling-technology pledges, are what will determine whether this investment is remembered as a model or a cautionary tale.

The Choice Ahead

The scramble for African colonies in the late nineteenth century was justified at the time as bringing civilization, infrastructure, and commerce to supposedly backward regions. It also stripped resources and left artificial borders and underdeveloped economies that still shape African politics today. The scramble for AI infrastructure across the Global South, in India, in Southeast Asia, increasingly in Africa itself, is being justified today as bringing jobs, digital infrastructure, and economic modernization. Some of that promise is genuine. Data centers do create construction jobs, tax revenue, and sometimes durable technical employment.

But the residents of Visakhapatnam are not wrong to ask hard questions before construction proceeds, nor are they being reflexively anti-progress by doing so. The history of foreign capital in India, from indigo to tea to textile protectionism, is a history of promises made during the investment phase and value extracted during the operating phase. Google has an opportunity, with billions of dollars and years before the facility opens in 2028, to prove that this cycle can be broken: genuine water stewardship, transparent land dealings, real local employment pipelines, and profit-sharing that does not simply flow back to Mountain View. If it does not take that opportunity seriously, Vizag will not be remembered as a Silicon Valley success story. It will be remembered as Assam's tea gardens with better branding.