Can Private Players Reinvent India’s Forecasting Future?

Part two of CarbonCopy’s series on India’s weather warnings explores how startups and private players are betting on AI, satellites, and new data streams to plug forecasting gaps. While facing hard questions of scale, trust, and integration

In the first part of this series, CarbonCopy explored why India’s weather forecasts so often fall short. Despite major investments in satellites, radars, and data networks, the India Meteorological Department (IMD) has struggled to predict the intensity of extreme rainfall, flash floods, and cloudbursts,  especially in regions with complex terrain like the Himalayas. Scientists told CarbonCopy that rainfall is inherently harder to model than heat waves or cyclones, and that even with the best global systems, hyperlocal extremes remain difficult to pin down.

The impact of heat and changes in weather patterns have begun to show serious impacts on the economy in India, one of the world’s worst climate-change-affected nations. The rising  and monsoon swings are affecting agriculture, which employs about 45.67% of the population and flooding and landslides caused by erratic monsoons is affecting cities and fragile Himalayan states. This comes at a time when multilateral cooperation is collapsing as was evident in the US cutting funds for National Oceanic and Atmospheric Administration (NOAA) that might affect India’s weather monitoring and forecasting capabilities in the long run. 

That leaves India at a crossroads. IMD, which has been operational for the last 150 years, is still the country’s backbone for weather and climate data. Its improved cyclone predictions have saved thousands of lives, and initiatives like Mission Monsoon, launched in 2012 by the Ministry of Earth Science to improve accuracy of monsoon rainfall forecast, promise further upgrades. But the hardest forecasting challenges demand new approaches, fresh data sources, and tools that move faster than traditional models.

Enter India’s growing band of startups. These companies are not trying to replace IMD,  which remains the country’s only official forecasting authority. Instead, they are carving out space to complement it. Some are building AI-driven systems that refine predictions at the village or city level. Others are launching hyperspectral satellites or using thermal mapping to track heat islands and flood risks. Together, they are positioning themselves as “force multipliers,” capable of filling gaps that the national system, by itself, cannot bridge.

This second part of the series looks at how private players — from early-stage ventures to space-tech firms already working with NASA — are reshaping what forecasting can mean in India. Their promise lies in speed, precision, and new kinds of data. But there are limits. Startups are young and untested at scale. Their models may be sharp in one city, but unproven nationwide, and commercial incentives sometimes push them toward niche clients instead of public warning systems.

The question, then, isn’t just whether startups can innovate. It’s whether they can embed themselves into India’s broader forecasting ecosystem in a way that truly saves lives.

The focus is on Granular Data and Artificial Intelligence

For most startups, the pitch is simple. Artificial Intelligence (AI) can do what physical models alone cannot. Take ClimateAi for example, which builds climate risk models that help businesses plan for disruptions in the long term. Its model can learn the complex relationships like how global phenomena like the El Niño (the periodic warming of the Pacific that can suppress monsoon rains), La Niña (the cooling of the Pacific ocean that amplifies Monsoon rains), and Madden-Julian Oscillation (MJO, causes wet and dry phases every 30-60 days) and Southern Pacific Monsoon affect monsoon, argues founder Himanshu Gupta. AI-driven technology is also cheaper and faster to update as soon as new data is available, according to him. 

Other start-ups are chasing hyperlocal insights. SatLeo Labs, an Ahmedabad based start-up, captures infrared bands through satellite sensors to detect temperature fluctuations on the Earth’s surface – what it calls “thermal intelligence.” SatLeo refines this data with AI to map heat islands and flag risks like wildfires. It is already working with the municipality of Tumakuru in Karnataka.

Academic institutes are also looking into AI solutions. At IIT Kanpur’s Kotak School of Sustainability, Professor SN Tripathi and his colleagues are working on integrating physical models, statistical methods, and machine learning for better accuracy. “We start with what data we have from the past and we try to predict what’s going to happen today and if we have more and more information available at the local level then the AI model will be able to learn it and get more accurate information,” says Professor Tripathi.  

The appeal is obvious. According to Professor Tripathi, current IMD models can gauge weather patterns for tracts going up to 6 sq kms. This is not enough, especially in hilly areas, where conditions can change over a ridge. Start-ups promise to close that gap. 

Going beyond AI

Beyond AI, some ventures are looking to push new domains to observe the earth. Pixxel, a Bangalore based start-up, is building a constellation of hyperspectral imaging satellites. These remote-sensing satellites capture a broad spectrum of light across many wavelengths, producing detailed data on surface materials and vegetation, especially in hilly terrain regions. 

But Pixxel’s focus isn’t short-term weather. “Our focus is climate intelligence: using hyperspectral satellites to monitor and address long-term environmental challenges, strengthening national Earth observation capabilities and climate monitoring,” Awais Ahmed, founder and CEO of the startup, told CarbonCopy. 

Ahmed claims their data can be used to monitor methane emissions and other atmospheric pollutants at a fine-grained level, which traditional satellites cannot capture. 

Who is buying these technologies?

These startups don’t deliver weather forecasts directly to the public yet. Instead, they’re building this data primarily to cater to their clients, which include governments and corporations. IMD itself has signed an MoU with Google for sharing cyclone-related advisories in the Regional Association (RA) II Asia region, which includes areas like East Asia, South Asia, Central Asia, and West Asia, and for the development of short-term or nowcast techniques for prediction of location-specific rainfall. It is also open for Public-Private Partnerships (PPP) in the dissemination of forecast and warning messages related to agricultural services. 

The Indian National Space Promotion and Authorisation Centre (IN-SPACe), under the department of Space, had recently announced the PPP with Pixxel and other companies for climate change monitoring, disaster management, agriculture, infrastructure, marine surveillance, national security, and urban planning. The aim for this partnership is to build India’s first indigenous commercial Earth Observation satellite. 

Private sector enterprises are also a very big market for these companies. ClimateAi, a US-based company, has built their client base in the food, agriculture, finance, and FMCG sectors to help them adapt to climate disruptions.  In India, they have collaborated with Imperial Tobacco Company of India (ITC) to generate precise climate risk forecasts for specific times and crops, providing adaptation insights such as weather based crop-guidance and crop calendar for farm planning. In November 2023, this FMCG conglomerate launched the app called ITC Metamarket for Advanced Agriculture and Rural Services (MAARS), which claims to help farmers as well as ITC’s Agri and Food businesses. 

SatLeo Labs, as CarbonCopy mentioned earlier, is currently working with the Tumakuru municipality to identify heat zones and monitor landfill emissions.

Investors seem to be responding by placing big bets on these technologies. SatLeo raised ₹28 crore ($3.3 million) in their pre-seed round led by firms like Merak Ventures, Huddle Ventures, Java Capital, IIMA Ventures and angel investors like Manish Gandhi and Dheer Baldua. Meanwhile, Pixxel has raised a total of $95 million from investors like M&G Catalyst, Glade Brooke Capital, Google, Radical Ventures and Lightspeed.

According to Grandview Research, the private weather forecasting space in India is currently valued at $114.9 million and is projected to reach $174.3 million by 2030 growing at a compounded annual growth rate of 7.2%.

But the success of these startups is difficult to identify yet as they are still at a nascent stage. They have not had any case study so far in India which can prove that the technology can work on a national scale. 

Limitations

One of the biggest limitations is that many of these technologies focus on heat, emissions, or land-use, which are easier to model than extreme rainfall.  Dr. Krishna AchutaRao, Professor at the Centre for Atmospheric Science in IIT Delhi, says, “Extreme heat events are a lot easier to predict as temperature variations are less sharp as compared to rainfall, making heat distribution more uniform across larger areas. High and low pressures in the atmosphere also help in predicting heat events with a minimum of four to five days of advance warnings.”

Other limitations that come across for these startups are that they are acting in silos. SatLeo Labs which uses geothermal intelligence for their climate forecasting looks solely at the surface temperature. It does not include other forecasting parameters such as ground observations and other data sets. Shravan Singh Bhati, co-founder and CEO of SatLeo says, “Currently, we are solely working on geothermal intelligence and scaling it will take a lot of time.”

Pixxel, which is working on hyperspectral imaging, however, does not solely rely on it. Rather, they integrate it with complementary data sources such as LiDAR, SAR, and Aurora, which is their Earth observation studio, and also validate it with ground-truth data.

Currently, private forecasting startups cater to businesses seeking insights, not the government. Experts believe their insights could only be beneficial if it is complementary instead of replacing state capabilities.

“Privatization must be carefully coordinated,” said  Dr. Raghavan Krishnan, retired scientist from IITM Pune. “Government scientists must remain central, with private players brought in to complement, not replace, public capabilities. Strong coordination, capacity building, and collaboration are essential for success.”

Experts’ Analysis of Private Sector’s Role  

Most experts agree that startups are useful, but only as supplements. IMD Director General Mrutyunjay Mohapatra himself has welcomed AI and private partnerships, noting that the department’s digitised weather records since 1901 could be a treasure trove for machine learning. 

Dr Madhavan Rajeevan, who is a former secretary of the Ministry of Earth Science, says, “They [private players] can fine-tune the data or provide additional data sets to complement the forecasting, but replacing IMD is not going to happen.”

Professor Tripathi argues data from these AI models must be made available in the public domain. He emphasised that these solutions are highly relevant for Global South countries like India, Bangladesh, East Asian countries, and the African subcontinent, which are disproportionately affected by climate change.

However, he is also cautious that technological solutions have their limits. “The AI and forecasting space cannot solve the entire climate crisis,” he said. “Broader changes in consumption patterns, use of resources, and energy dependency, especially in developing countries, are equally crucial.” 

Professor AchutaRao said, “Although AI is helpful in leveraging better data and training models, especially for extreme heat, the accuracy of the model depends on the quality and volume of data, which is currently limited, but expected to improve over time.” He adds that forecasts have natural limits — warnings can reduce losses, not eliminate them.

The promise of these private ventures is real. They give access to sharper insights, faster updates and more granular data. But their limits are real, too, such as unproven accuracy, and the risk of commercialisation. The real test will be whether startups can embed their innovations into India’s broader forecasting ecosystem, not just as niche services for clients, but as effective tools for national resilience. 

About The Author

Paridhi Choudhary

See author's posts