It was just after 5:30 in the morning on 12 November 2023 when a section of the Silkyara Bend-Barkot tunnel in Uttarakhand’s Uttarkashi district gave way, trapping 41 construction workers about 200 metres inside. The tunnel, part of the government’s ambitious all weather Char Dham highway project, had no dedicated escape shaft. Workers who had been on site earlier told reporters that large cement pipes, meant to serve as an emergency exit, had been removed by contractors sometime before the collapse. Outside, families gathered and waited. Sudhanshu Shah whose younger brother Sonu was among the trapped men, camped at the site for over two weeks.
What followed were seventeen days of sheer improvisation. An American auger drilling machine, brought in to bore a horizontal escape pipe, broke down repeatedly and eventually snapped inside the rubble. International tunnelling consultants were flown in to advise. In the end, it was a small team of rathole miners, men who dig narrow coal shafts (declared illegal) by hand in eastern India using techniques with no formal engineering pedigree at all, who cleared the final stretch of debris using shovels and their own bodies. On the evening of 28 November, all 41 workers were pulled out alive, uninjured, and flown by military helicopter to a hospital in Rishikesh for observation. Within days, forty of them were cleared to go home.
It’s tempting to end the story there, just as most news coverage did. A happy ending, proof that India’s disaster response machinery can improvise its way out of almost anything. Prime Minister Narendra Modi called the rescuers afterward to congratulate them, the workers were feted, photographed, and sent home to their families. But Silkyara raises a harder question than ‘could the workers be saved.’ It asks what happens after the cameras leave a rescue site, months later, when nobody is watching and the next tunnel contract is being signed, whether a near miss actually changes anything, or whether the next mountain gets bored through exactly the way the last one was.
If you look at India’s disaster record in the driest possible terms, it reads like a vast, and brutally expensive, research programme. Cyclones hammer the east coast most years. Flash floods and cloudbursts keep coming back through the Himalayan states, where a subsiding town or a severed mountain road has become an almost annual monsoon story. Rivers spill over their banks across the Gangetic plain most years, and in Assam the Brahmaputra floods substantial stretches of the state nearly every monsoon, submerging farmland and villages in a slow, recurring disaster that draws far less coverage than a dramatic rescue, precisely because nobody is trapped on camera while a riverbank erodes. Coastal Odisha loses a little more ground to the sea each decade.
Somewhere in the country, nearly every kind of infrastructure failure that can happen to a road, a bridge, a drainage channel or a tunnel already has happened, more than once, and been documented in some form by engineers, auditors, journalists or courts. That accumulated experience is genuinely valuable. But most of the time, it’s sitting in disconnected reports rather than shaping the next contract, which is the precise gap CDRI was designed to close.
This is the backdrop against which India built the Coalition for Disaster Resilient Infrastructure, the institution this series has been building toward since its first article. Modi first proposed it at a disaster risk reduction conference in New Delhi in 2016. Three years later, in September 2019, it was formally launched at the United Nations Climate Action Summit in New York.
Headquartered in New Delhi, CDRI now counts more than fifty member countries and over ten partner organisations, including the World Bank, several UN agencies and the European Union. It’s grown steadily rather than dramatically. The Economic Community of West African States joined as recently as February 2026, bringing the bloc’s member states in as a group. That same year, the coalition’s governing council passed its rotating co-chair from France to Brazil, a small but real sign that the project has outgrown being an Indian vanity initiative and become something other governments are willing to help steer.
Its current work includes a fellowship programme funding around fifty six researchers across a dozen countries on projects ranging from tsunami evacuation modelling in Japan and Peru to wildfire resistant building materials in India. There are also guidance documents on subjects like protecting schools from extreme heat, plus a published data and technology strategy meant to help poorer member states use satellite and sensor data they currently cannot afford to collect themselves.
None of this makes CDRI a finished success. It’s a coalition still building its own institutional muscle, six years after launch, with a stated ambition to nearly double its membership that it has not yet reached. A great deal of its published output so far consists of frameworks and guidance documents rather than finished, evaluated infrastructure on the ground.
The honest question about an organisation like this isn’t whether its mission is worthy, it plainly is, but whether the country hosting and leading it can actually demonstrate the thing it’s asking others to build. India didn’t invent this dilemma. Every country that’s tried to export a model of good governance has faced some version of it. But CDRI’s entire pitch is technical credibility rather than money or military weight, which makes the gap between claim and practice harder to paper over. Call it the credibility gap, and it has three tests.
To be fair to CDRI, it hasn’t exactly been sitting on its hands. The coalition’s flagship Global Infrastructure Resilience Report, first out in October 2023 and updated in 2025, brought something genuinely useful to the table, the Global Infrastructure Risk Model and Resilience Index. CDRI calls it the first publicly available, fully probabilistic assessment of infrastructure risk anywhere in the world. The 2025 edition put a number on a problem few governments like saying out loud. Global disaster losses exceeding $732 billion a year, with the indirect damage from service disruption running, on average, 7.4 times higher than the direct cost of the wrecked infrastructure itself.
CDRI has also turned that lens on its own host country. A dedicated study, Mainstreaming Resilience into Infrastructure, looked specifically at India’s roads, railways and power sector, and calculated that natural hazards could cost the country as much as $1.02 trillion against the $4.51 trillion it plans to spend on infrastructure by 2030. It didn’t stop at the number: it built a cost-benefit tool to justify resilience spending to finance ministries, and a toolkit meant to help government departments write hazard assessments into contracts from the start, rather than after the tender is signed. All of this sits on CDRI’s own website, freely downloadable, and it deserves to be read before anyone waves the coalition away as a talking shop.
What it isn’t, yet, is the kind of document this article has been asking for. A global risk index and a sector wide gap analysis are diagnostic tools, they tell a government roughly how exposed its power grid or its rail network is in the aggregate, in dollars and percentages. They don’t say why a specific tunnel collapsed, or a specific embankment failed, or which named department was responsible. Nothing in CDRI’s published output, so far, examines Silkyara or Joshimath by name and checks whether the lessons its own reports recommend were actually applied there. That’s a harder report to write, because it names people and departments rather than sectors and probabilities. It’s also exactly the report a credibility gap is closed by, and CDRI hasn’t written one yet.
The second is institutional continuity. Silkyara is a useful test case precisely because its aftermath was neither a triumph nor a cover-up, but something messier and more believable than either. Within hours of the rescue, the road transport minister, Nitin Gadkari, promised a nationwide safety audit of tunnels. Weeks later, an opinion piece in a national newspaper was still asking where that audit had gone, noting that officials had begun arguing over whether an escape tunnel had even been legally required in the first place. That argument was never entirely resolved in public.
But over the following two years, the ministry did issue tunnel safety norms requiring geological risk zoning, minimum diameter escape pipes and rescue containers in high-risk stretches, and mandatory emergency response plans updated on site. When the Sela tunnel in Arunachal Pradesh, built for strategic connectivity to Tawang, neared completion, the government required an independent third party safety audit before it could open to traffic, a precaution Silkyara did not have.
It’s a story of a two-year lag between promise and rule, punctuated by bureaucratic hedging in between, which is a far more honest description of how institutions actually learn than any speech will admit. It’s also, on balance, a passing grade: the rule now exists, it’s already been applied to at least one major project, and it exists because a specific failure was investigated rather than quietly filed away.
The third test comes down to this. Can India point to a specific place where resilience has measurably improved, not just on paper, but in reality? Chennai makes for a useful, if unfinished, case study.
Back in December 2023, Cyclone Michaung dumped close to 45 centimetres of rain on the city in just 48 hours. The images were stark with cars floating down major roads, the airport runway underwater and shut for a day, floodwater unable to drain into the sea because the cyclone’s own storm surge was pushing back. At least four people lost their lives in Tamil Nadu. Afterward, reporters traced part of the damage to a pattern we’ve seen too many times, a residential high rise in Pallikaranai had gone up on what residents and environmental activists called obvious marshland, ground where flooding was, in their words, practically inevitable.
Yet there were also signs that Chennai had learned something from its catastrophic 2015 floods. Reservoir operators let water out earlier and more carefully this time, avoiding a repeat of 2015’s uncontrolled discharge. A partly rebuilt stormwater drain network moved water out of the core city faster than it had eight years before. And weeks after the storm, the central government green lit a ₹561 crore integrated urban flood management project for the Chennai basin.
Will that money actually produce a city that doesn’t flood the next time a cyclone parks itself offshore? As of this writing, that’s still an open question. But that uncertainty is the honest answer, and frankly, a more useful one for other coastal cities than a premature victory lap.
None of these three tests demand perfection. What they do demand is a paper trail, a warning that was recorded, a decision that followed from it, and a way to check later whether that decision actually held. A country that can show that trail, even when it’s embarrassing, has something real to teach. A country that can only show ribbon cuttings? Not so much.
Take Silkyara and Chennai together, and they show you why resilience can’t be reduced to a single formula. Then throw in the Brahmaputra’s yearly floods in Assam, mentioned earlier, and you’ve got a third kind of problem altogether, not a single dramatic failure, but a problem so routine it barely makes a national headline.
A Himalayan tunnel collapses because of geology, construction pressure, and an emergency exit that someone decided to remove. A coastal metropolis floods because reservoirs, drainage, storm surge, and thirty years of construction on wetland all interact in ways no single agency fully controls. A river that floods the same villages every June doesn’t fail because of any single decision, it fails through the accumulated absence of one. Nobody ever quite gets around to relocating a settlement that floods predictably rather than catastrophically.
Here’s the thing, what worked in one place, Chennai’s early reservoir releases or Silkyara’s eventual safety code, rarely transfers automatically to the next. Each geography needs its own engineering answer. But what can transfer is the discipline of writing the failure down and holding someone accountable for fixing it. That’s a process, not a technology, and it’s exactly the kind of thing a coalition of countries can actually share with each other.
This isn’t cheap, and the costs go way beyond what you can see. Every rescue operation, relief camp and reconstruction contract is paid for by the same taxpayers who already paid once to build that road, tunnel or reservoir in the first place. Chennai’s ₹561 crore flood project sits on top of, not instead of, whatever the city already spent on drains that didn’t hold.
Economists have a name for this pattern, socialised risk. Put simply, it means a builder who profits from a high rise on marshland doesn’t pay for the flood it causes. The government does, and by extension, every taxpayer does. And here’s the uncomfortable part, a coalition that wants to export resilience abroad will struggle to be taken seriously if its home country keeps letting private actors profit from exactly the kind of construction its own guidance documents warn against.
None of this requires a research paper to explain. It doesn’t need to be complicated. Start with the money. Every big project, in India and in anything CDRI helps pay for elsewhere, should have its maintenance bill written down and budgeted for before the first shovel goes into the ground. Not squeezed out of whatever is left over once the ribbon is cut. That single habit would have made Chennai’s drains, and the embankments that give way along the Brahmaputra most Junes, someone’s clear job rather than everyone’s vague responsibility.
Which is really the deeper problem. Ask who owns a stretch of national highway, or an embankment, or the tunnel wall at Silkyara. The honest answer is often several departments, none of whom will admit it first. A road, a bridge, a drainage channel should each belong to one named office, the kind that gets credit when it holds and blame when it doesn’t.
Audits need to stop being something that only happens after the mishap. The Sela tunnel was checked by independent engineers before it opened, not after, because Silkyara had just happened and for once somebody decided not to wait for a body count. That should be routine, not an exception worth a newspaper column years later.
Contractors need the same discipline extended to them, answerable for what they build years afterward rather than just while the concrete sets. And none of it holds if the one engineer who understands why a particular hillside keeps sliding gets rotated to a desk job the moment she starts to matter. Technical people need to stay put long enough to remember what happened last time.
CDRI should hold itself to the same standard it quietly asks everyone else to meet. That means collecting its own members’ failures, India’s included, and publishing what it learns rather than only what flatters an annual report. It means putting an Indian official and a Kenyan or Fijian official in the same room, learning from the same case studies, instead of treating the arrangement as India teaching and everyone else taking notes. And it means judging itself by whether a hillside stays put the next monsoon, not by how well a workshop was attended.
None of this argues for slowing India down, or treating every new road as something to be resisted on principle. It argues for spending money India already spends a little more carefully, and for saying out loud what still hasn’t worked.
Set against China’s much larger overseas infrastructure lending, India’s smaller, resilience-branded coalition is a genuinely different kind of offer. China’s projects have repeatedly drawn criticism from independent researchers and recipient governments for prioritising speed and scale over climate and disaster resilience. India’s pitch is different: less steel, more standards, fewer ports, more post-disaster audits. That is a real strategic choice, not an accident of India having less capital to lend, and it is a defensible one on its own terms.
It will only be believed, however, if the standards it exports visibly apply at home first. A country cannot credibly sell the idea of an enforced maintenance schedule to a partner government while its own public works departments treat maintenance as the first line item to be cut when budgets tighten.
There is real evidence to build on, and it would be its own kind of dishonesty to pretend otherwise. When Michaung threatened Tamil Nadu and Andhra Pradesh in the same week Chennai flooded, more than 15,000 people were moved out of coastal districts in the two states combined and thousands more sheltered in relief camps. This was part of a now-routine evacuation drill that Indian coastal states run before nearly every major cyclone, a capability that barely existed in most of the country’s coastal administrations a generation ago.
Further up the same coast, Odisha has refined that same evacuation discipline into something close to a zero-casualty model for the cyclones, even where it has struggled with the slower, harder problem of a coastline that keeps eroding regardless of how well the evacuation goes.
CDRI’s own fellowship programme is, in its small way, evidence that India is willing to fund people studying its own failures rather than only its successes, and to put them in the same cohort as researchers from Japan, Vietnam and Turkey doing the same work on their own hazards. None of this adds up to a finished system. It adds up to a country with the administrative scale, the engineering talent, the digital reach and now the accumulated failures to build one, if it chooses to reward the parts of government that prevent disasters as generously as it rewards the parts that respond to them.
The forty one men pulled from the Silkyara tunnel went home to families who had spent seventeen days camped outside a mountain, not knowing whether an escape pipe removed for convenience would end up costing a life. It did not, this time.
Tunnels built after Silkyara are now required to plan for the possibility that it might. At least one, hundreds of kilometres away on a militarily sensitive road to Tawang, was made to prove it before it ever opened to traffic. That is a narrow, specific, unglamorous kind of progress, a design rule buried in a government circular and a single audit clause in a single tunnel’s approval file, unlikely to feature in any speech about India’s global leadership. But it is precisely the kind of progress a country can actually export, because it survived contact with a real failure rather than a hypothetical one, and because someone, somewhere, can still point to the file and check.
Article One of this series showed how India builds for the photographs and maintains for no one. Article Two showed how it allows settlements, generation after generation, in places the water and the mountains had already marked as dangerous. The test that remains is whether a country that still does both of those things can nonetheless become the place other nations turn to for advice on how to stop doing them.
It can, but not by saying so at a summit. It will need every future Silkyara to end not just with forty one men walking out alive, but with a rule written down, applied to the next mountain, and checked. That will depend less on what India says in New York or New Delhi than on whether the next tunnel has an escape shaft nobody removes.
Lower per-tonne freight costs, sharply lower fuel use, multimodal terminals, Jalvahak incentives and private investment are shaping a system in which heavy cargo can move by river for the longest, most efficient leg of journeys.
India’s goal now should not be just to become a ‘Builder’s Navy’, but a ‘Fast Builder’s Navy’. This means freezing designs, placing large orders together and never letting shipyard assembly lines sit empty.
India notified an offshore wind policy in 2015 and sits on about 70 GW of potential off Gujarat and Tamil Nadu. 11 years later, not one commercial turbine stands in the water. This article tracks why the offshore turbine tenders had no bidders.
Selected issues, clear explanations and thoughtful perspectives—delivered when we publish.
