At 9.10 am on August 26, a school staffer burst into a classroom in Nuwakot, Nepal, shouting that a wall of water was hurtling down the valley. Rajendra Dawadi, the principal, ordered nearly 900 students on campus to run for higher ground, put fresh arrivals back on buses, and directed drivers to leave immediately. By 9.23 am, just 13 minutes after the first alarm, the school was gone, swept away alongside the valley floor. Every student survived.
Most people in the valley were not as fortunate. High above, a mass of rock and glacier ice, estimated at up to seven billion cubic feet, broke off the north face of Langtang Lirung, near the Nepal-China border, and thundered downhill. It is likely the largest glacier collapse recorded in the Himalayas.
More than 7,000 people are dead or missing, many buried under feet of debris. As Nepal mourns and begins rebuilding, it is confronting a grim reality: such a disaster could easily recur.
Climate change is causing Himalayan glaciers to retreat and permafrost to thaw, making the range increasingly unstable. What if mountain communities received warnings long before a glacier collapsed? International scientists could use artificial intelligence to monitor satellite imagery continuously, issuing alerts when warning signs appear. Building this system offers wealthy nations a way to help remedy climate disasters created largely by their disproportionate fossil fuel consumption.
Relocating vulnerable communities might seem the obvious choice, but river corridors most exposed to these hazards are also Nepal’s prime hydroelectric sites. At least a dozen hydropower projects were damaged or destroyed in this single event.
Nepal cannot easily forgo the revenue and power these rivers generate. Its per capita electricity consumption is among the lowest globally, and hydroelectricity remains its best vehicle for clean economic growth. The country aims to scale up its current 4,300 MW installed capacity to 30,000 MW by 2035. However, planning can no longer assume these valleys are safe. Every dam, tunnel, and workers’ camp constructed below a glacier is now a gamble against mountain stability, and the odds are shifting.
Here, global science can offer concrete assistance. Satellite radar recorded Langtang Lirung’s slow-motion instability for months. Data from January showed the slope creeping downhill, accelerating weeks before the failure. Yet no one analysed the data until it was too late.
Identifying such signals in one mountain post-disaster is straightforward. Monitoring thousands of glaciers and rock slopes across the Himalayas continuously is a vastly larger challenge. It is precisely the task modern AI tools are designed for. Machine learning can automatically sift through massive satellite datasets daily, flagging the few accelerating slopes among thousands of stable ones.
Designing and testing such a system requires a major coordinated effort by glaciologists and geomorphologists worldwide. Deployed at scale, it could grant communities hours or days of notice rather than 13 minutes, allowing authorities to execute planned evacuations instead of improvising. The network could also mitigate risks in neighbouring countries like China, India, Pakistan, and Bhutan.
Climate risk is not Nepal’s to solve alone. Its contribution to the greenhouse gas emissions warming the planet is negligible, yet the mountain communities living beneath these unravelling massifs are among the poorest people on earth, suffering a disproportionate share of the consequences. Climate finance funds, such as the United Nations Fund for Responding to Loss and Damage, established in 2022 after the torrential floods in Pakistan, were designed with exactly these disasters in mind.
Contributions from wealthier nations could be used to make Nepal more resilient, but the fund is woefully underfunded and slow to process claims. New money is needed for development of monitoring systems, disaster planning and, in the most exposed valleys, new roads and escape routes that let people get away from the river.
The heavy monsoon rains will return next summer, and the glaciers above these valleys will keep retreating as global temperatures rise. The protections Nepal needs will take years to build, and the countries most responsible for the warming can afford to pay for them. The question is how many disasters Nepal will endure before they do.