A Case of How Not to Tackle the Most Essential Component of a New State -The Capital That Wandered: Twelve Years of Andhra Pradesh Searching for Its Own Address

*An essay on indecision as the most expensive infrastructure project of all*

----------------------

## Where, exactly, is the capital?

Ask the question literally, and the answer scatters. Drive out of Vijayawada across the Krishna and you reach Velagapudi, where an "interim" Secretariat — a cluster of functional blocks thrown up in 2016, meant to last five years — has now housed the government of Andhra Pradesh for a decade. The Assembly meets in a building carrying the same adjective. The High Court sits at Nelapadu in a structure that was always described as temporary. The Chief Minister's residence and camp office are across the river at Tadepalli — technically the capital region, practically a suburb of Vijayawada. A few kilometres away stand the half-risen towers of the original master plan, skeletons that weathered years of abandonment before the cranes returned. Scattered between paddy fields: the campuses of SRM, VIT and Amrita that opened on schedule and kept teaching through the chaos, a seed access road, judges' bungalows, unfinished MLA housing, and thousands of acres of pooled farmland that has spent a decade as neither farm nor city.

So the *physical* capital, after twelve years, is an archipelago of the word "interim." And where is the capital *felt*? Not yet in Amaravati. The feel of a capital — that density of power, money, media and ambition pressing against each other — is felt in Vijayawada's hotels and Tadepalli's corridors, faintly in Visakhapatnam's boardrooms, and, for most of the state's first decade, above all in a city that belongs to another state: Hyderabad, where AP's institutions, elites and imagination remained tenants until the shared-capital clause lapsed in 2024. A state of five crore people spent its formative years with its body in a construction site and its soul at someone else's address.

Only in April 2026 did the question acquire a settled legal answer: Parliament amended the AP Reorganisation Act, the President assented, and Amaravati — the greenfield city on the Krishna's right bank between Vijayawada and Guntur — became unambiguously *the* capital, with works worth roughly ₹58,000 crore under execution on World Bank and Asian Development Bank funding. But notice the tense. In 2026, the capital of a state carved out in 2014 is still *under execution*: a construction site with a constitution. And the honest answer to "where is the epicentre of growth?" is more uncomfortable still — there isn't one yet. There are candidates. Visakhapatnam grew organically into the state's largest economic node, carrying IT, pharma, port logistics and now data-centre investments almost despite state policy rather than because of it. Vijayawada-Guntur remained the commercial and political heartland. Tirupati built a quiet electronics cluster. But no city in residual Andhra Pradesh today plays the role Hyderabad plays for Telangana, Bengaluru for Karnataka, or Chennai for Tamil Nadu: a gravitational centre dense enough to hold talent that would otherwise leave. That absence is not an accident of geography. It was manufactured, year by year, by indecision.

## The wandering decade

The chronology reads like a parable. In 2014, Andhra Pradesh lost Hyderabad — the city it had spent six decades building into an IT and pharma powerhouse — retaining only a ten-year lease on shared-capital status that everyone knew was symbolic. In 2015, Chandrababu Naidu announced Amaravati with genuine ambition: 33,000-plus acres voluntarily pooled from farmers across 29 villages, a Singapore government-linked consortium as master developer, a plan estimated around ₹33,000 crore, and renderings of a riverfront city that borrowed equally from Singapore and the Satavahanas.

Then came 2019. The incoming government halted construction, the Singapore consortium exited, and by December a three-capital formula was tabled: executive capital in Visakhapatnam, legislative in Amaravati, judicial in Kurnool. Whatever the theoretical merits of decentralisation — and there were arguable ones — the effect in practice was to convert a capital into a question mark. Farmers who had surrendered fertile multi-crop land sat on annuity payments and appreciating grievances, protesting for over 1,500 consecutive days. The High Court intervened; bills were withdrawn and re-signalled; investors read the only message that mattered: *nothing this state announces survives an election.*

In 2024 the pendulum swung back. The new government revived Amaravati with a ₹65,000 crore package, re-secured multilateral funding, and restarted tenders. By 2026, the legal cloud lifted. Eleven of the twelve years since bifurcation, in other words, were spent deciding — and un-deciding, and re-deciding — the answer to question one.

## The ledger of loss

What did the wandering cost? No single audited number exists, but the components can be honestly assembled.

**Direct cost escalation.** The same city that was scoped near ₹33,000 crore in 2015 is being built with a ₹65,000 crore commitment a decade later — roughly a doubling, before counting the interest burden on borrowings for assets that sat idle or half-built through the freeze. Partially completed towers and roads deteriorated between 2019 and 2024; some work was effectively paid for twice. A conservative reading puts the pure construction-side penalty of delay at ₹25,000–30,000 crore.

**The investment that never came.** This is the larger, invisible column. Between 2019 and 2024, any firm weighing a South Indian headquarters, campus or plant faced a state that could not say where its own government would sit. Capital is a coward; it went to Hyderabad, Bengaluru and Chennai instead. Telangana, which inherited a settled capital, compounded relentlessly — it now posts the highest per-capita income among large Indian states — while AP's growth stayed respectable but rented: ports, assembly, execution work, little headquarters gravity. If Amaravati had followed anything like the trajectory of comparable planned-city projects — GIFT City's slow-then-sudden curve, or even Naya Raipur's modest one — a decade of agglomeration effects, land-value capture, and jobs (the current plan alone projects tens of thousands in construction and services) was forfeited. Economists who have attempted the counterfactual arithmetic land in the range of ₹1–1.5 lakh crore in foregone output and investment over the decade. The precision is debatable; the order of magnitude is not.

**Human costs.** Twenty-one thousand farming families bet their land on a promise that froze for five years. A generation of graduates from Guntur, Krishna and Godavari districts made the rational choice the state's indecision dictated: Hyderabad, Bengaluru, or a visa. The talent-export pattern the Telugu states already suffered from was aggravated precisely where it might have been arrested.

**Institutional credibility.** Hardest to price, longest to repair. The Singapore consortium's exit was watched globally. Every future negotiation — with the World Bank, with anchor investors, with its own farmers on new land pooling — now carries a risk premium earned by the 2019–2024 reversal. Sovereign-style commitments from AP will be discounted for a generation.

## How it could have been different

The tragedy is that the alternative required no genius — only restraint. Imagine a different 2014–15: an all-party resolution, passed in the first assembly session of the new state, declaring the capital's location a settled constitutional-style question, immune to change without a two-thirds majority. A statutory, professionally governed Capital Region Development Authority with a fifteen-year insulated mandate, board seats for the opposition, and financing ring-fenced from the annual budget theatre. A phased, fiscally sober plan — administrative core first, iconic towers later — instead of an all-at-once vision vulnerable to the charge of grandiosity.

Under that counterfactual, by 2026 Amaravati would be roughly where GIFT City or even a mid-trajectory Naya Raipur stands: a functioning administrative city of a few lakh residents, a completed secretariat and High Court, two or three anchor institutional campuses (the Purdue, King's College and XLRI partnerships signed in 2015–18 would have matured rather than lapsed into limbo), an operating startup and fintech zone leveraging the state's engineering pipeline, and — critically — a land-value flywheel returning money to the pooled-land farmers, converting them from protestors into the project's fiercest defenders. Visakhapatnam, freed from the executive-capital tug-of-war, could have been developed on its own honest logic: as the state's commercial and port-economy engine, its Mumbai to Amaravati's Delhi. Two engines, clearly defined, instead of two half-promises competing for one identity.

The deeper counterfactual is cultural. Tamil Nadu's parties war over everything except Chennai's primacy and the industrial-policy consensus; that continuity is why Hosur gets factories whichever party wins. AP's parties chose to make the capital itself the battlefield — the one asset that should have been fenced off from politics became its central trophy.

## Could a good start have been made by today? Yes — and the proof is painful

The proof that consensus would have worked is that the ingredients all existed: farmers *volunteered* land in 2015 at a scale India had never seen, multilateral lenders committed early, a sovereign foreign partner signed on, and public enthusiasm was genuine. Nothing about the project's fundamentals failed. Only the politics did. Had the 2019 government inherited a legally entrenched, opposition-inclusive capital authority, it could have redirected emphasis — more spending on Rayalaseema and North Andhra, a leaner Amaravati — without detonating the project itself. Decentralised development and a single settled capital were never actually incompatible; they were made to seem so.

## The epicentre, honestly assessed

So where is the epicentre of growth in 2026? The truthful answer: distributed and provisional. Visakhapatnam is the organic economic centre. Amaravati is the declared political centre, now finally under genuine construction. The Vijayawada-Guntur corridor is the demographic centre. Whether these fuse into a coherent growth geography over the next decade depends on one variable above all: whether the 2026 settlement *holds* across future elections.

Twelve years taught Andhra Pradesh the most expensive lesson in Indian federal history: that a capital is not buildings but *credibility compounding over time*, and credibility cannot be poured like concrete after the fact. The state lost a decade, tens of thousands of crores, and an entire cohort of its young to a question that consensus could have closed in a single legislative afternoon in 2015. The buildings can still be finished. The decade cannot. The only fitting memorial to it would be a political culture that never again puts the state's own address on the ballot.