A dangerous new era of mountain disasters is unfolding across the Hindu Kush Himalaya's (HKH) most vital water systems. The region's glaciated reserves, often described as the Third Pole, store more ice and snow than anywhere outside the Arctic and Antarctic regions. But what we are witnessing is not just the melting of ice, but also rapidly escalating risks for nearly a quarter of the global population. Along with glacial retreat, warming is triggering permafrost thaw and the destabilising of mountain slopes. These changes are also driving a sharp rise in high-altitude cryosphere-related disasters, including glacial lake outburst floods (GLOF), ice-rock avalanches and their cascading impacts such as debris flow, and flash floods. A review of major cryosphere-related disasters between 2012 and 2023 by the International Centre for Integrated Mountain Development (ICIMOD) paints a sobering picture. In 2012, a rockfall from Annapurna IV in Nepal triggered the deadly Seti disaster, killing over 70 people. The 2013 Kedarnath disaster in India killed several thousand after extreme rainfall and rapid snowmelt. In 2016, a transboundary GLOF from Tibet flooded Nepal's Bhote Koshi Valley, damaging hydropower projects and causing tens of millions in losses, with no warning reaching Nepal. Between 2019 and 2022, the surging Shisper Glacier in Pakistan triggered repeated outburst floods, damaging the Karakoram Highway and displacing communities in Hunza Valley. In 2021, an ice-rock avalanche at Chamoli, Uttarakhand destroyed two hydropower plants and killed over 200 people. The same year, the Melamchi disaster in Nepal, caused by cascading landslides, glacial outflow, and debris flows, killed at least 70 people and severely damaged Kathmandu Valley's drinking water system. In 2023, a GLOF from South Lhonak Lake in Sikkim breached the Teesta III dam, caused losses exceeding US$ 1 billion, and left dozens dead or missing. What is more alarming is not just the scale, but the severity and frequency of these disasters. Hazards that once occurred in isolation, are now clustering within single year timeframes. Between 2024 and 2025 alone, multiple high-mountain disasters rocked Afghanistan, Bhutan, India, and Nepal. These included a cascading flood in Thame, Nepal; an avalanche burying parts of Mana, India; a thermokarst outburst in Afghanistan's Andarab district; thermokarst outflow flood in Humla, Nepal; a transboundary flood in Rasuwa, Nepal; a massive flash flood in Dharali, Uttarakhand, India; and a flash flood in Gupis, Gizhar district, Pakistan that killed dozens and destroyed critical hydropower infrastructure. Interestingly, all events are traced back to their origin in the region's cryosphere. But our ability to diagnose the root-cause has been a major blind spot with events being frequently misattributed. Floods have been routinely linked to cloudbursts without supporting evidence, with later analysis revealing causes like glacial lake drainage, ice collapses, and debris flows. This error doesn't only reveal our weakness in how risks are assessed, but also how warnings are issued and how our responses remain reactive rather than preventive. Even more troubling is the transboundary nature of these risks. Many disasters, including the 2016 Bhote Koshi flood and the 2025 Rasuwa flood seen in Nepal, occurred in one country, striking another with little to no warning. What this revealed was that even with the best of intentions, national disaster management agencies may still be overlooking cross-border hazards. The need of the region is a systems approach where coordination, science-based attribution, and sustained investment are urgently prioritised. First, a real-time monitoring and observation network which integrates real-time data sharing, standard protocols for hazard assessments coordinated across the eight HKH countries is essential. Without this, national and local responses remain fragmented and early warnings remain unreliable. ICIMOD's own efforts have shown that while real time monitoring stations are possible, they need sustained community support. Community-based flood early warning systems have also been proved to be scalable and cost-effective but need institutionalisation within local level disaster plans. Second, disaster causation must be better attributed, with rapid, science-based studies prioritised – shifting the approach from reactive crisis management to proactive risk reduction. An example of this was seen in the 2025 Rasuwa flood, where independent researchers identified the glacial source within a day using satellite imagery, which formed the basis for Nepal's official response on the cause of the event. Third, financing flows need to be routed through the region. Current global adaptation funding falls short of the US$ 310-365 billion estimated to support long-term adaptation across the world. While the US$ 800 million Fund for Responding to Loss and Damage is a step in the right direction, mountain-specific concerns remain largely overlooked in global climate negotiations. With the launch of the Decade for Cryosphere Action, the region just might get the global attention it deserves. But what good is awareness raising without a change in our systems? We are already living through a period of intense risk, and our systems are no longer able to keep pace. A major disruption to the HKH would critically undermine water, food, and systems that support the lives of a quarter of humanity. With stakes as high as they are, will we step up and mobilise the political, technical, and financial will to act? Will we continue to respond to disasters after they happen, or finally build systems that uphold the roof of the world? Shrestha is CEO of the Asia Office at Hydrominds and Senior Advisor on Climate and Environmental Risks to ICIMOD