Water-energy intelligence for the AI infrastructure boom.
Maps, data products, and strategic advisory for
AI data centers, semiconductor fabs, and critical infrastructure.
Facility-level water and energy risk for the infrastructure behind AI.
Hydroclimate-resolved data on 9,500+ data centers and 300+ US semiconductor fabs — with the advisory to act on it. The same method has been applied to ammonia, chemicals, and thermoelectric power. Built for credit and capital allocation.
Facility-level water and energy risk for the infrastructure behind AI.
Hydroclimate-resolved data on 9,500+ data centers and 300+ US semiconductor fabs — with the advisory to act on it. The same method has been applied to ammonia, chemicals, and thermoelectric power. Built for credit and capital allocation.
Data Products
We built the facility-level databases that did not previously exist.
An integrated view of where water and energy risk emerges — at the site, watershed, and portfolio levels.
Each dataset combines precise facility locations with climate and hydrological models, energy system data, and sector-specific water use. The result is site-level intelligence that goes beyond simple location data to answer the questions that matter.
The four questions our data answers:
Which operator, REIT, or issuer holds the asset, and how exposure aggregates across their portfolio?
How much water is required?
Who else depends on the same water supply?
How do risks evolve under future climate scenarios?
Data Products
We built the facility-level databases that did not previously exist.
An integrated view of where water and energy risk emerges — at the site, watershed, and portfolio levels.
Each dataset combines precise facility locations with climate and hydrological models, energy system data, and sector-specific water use. The result is site-level intelligence that goes beyond simple location data to answer the questions that matter.
The four questions our data answers:
Which operator, REIT, or issuer holds the asset, and how exposure aggregates across their portfolio?
How much water is required?
Who else depends on the same water supply?
How do risks evolve under future climate scenarios?
9,500+ facilities. Most operators don't know their water risk.
Our global dataset combines facility locations with climate and hydrological models to deliver site-specific water scarcity intelligence — current conditions and future projections.
For each facility: water consumption and withdrawal, power use and source, water quality risk, basin-level scarcity index, and competing sectoral demand from agriculture, municipalities, and ecosystems.
9,500+ facilities. Most operators don't know their water risk.
Our global dataset combines facility locations with climate and hydrological models to deliver site-specific water scarcity intelligence — current conditions and future projections.
For each facility: water consumption and withdrawal, power use and source, water quality risk, basin-level scarcity index, and competing sectoral demand from agriculture, municipalities, and ecosystems.
300+ US fabs. Most sites in watersheds already under stress.
Our fab dataset combines facility locations with water demand estimates, wastewater discharge profiles, energy intensity, and basin-level supply risk — under current conditions and climate projections to 2050.
For each facility: water sourcing vulnerability, competing demand from agriculture, municipalities, and ecosystems, and the regulatory exposure that follows when industrial water use meets constrained supply.
300+ US fabs. Most sites in watersheds already under stress.
Our fab dataset combines facility locations with water demand estimates, wastewater discharge profiles, energy intensity, and basin-level supply risk — under current conditions and climate projections to 2050.
For each facility: water sourcing vulnerability, competing demand from agriculture, municipalities, and ecosystems, and the regulatory exposure that follows when industrial water use meets constrained supply.
Advisory Services
The maps tell you where risk lives. Advisory tells you what to do about it.
Advisory turns data into decisions — from site selection and due diligence to regulatory strategy and long-range planning. We work with operators, investors, and lenders.
Site Screening & Due Diligence
Watershed availability, cooling trade-offs, grid access, and permitting exposure at specific locations — before capital is committed.
Portfolio & Issuer Risk
Multi-site exposure scoring, scenario analysis, and stress testing across holdings, aggregated from facility to basin to issuer.
Policy & Regulatory
Regulatory analysis, impact modeling, and expert testimony for policy design and stakeholder engagement.
Sustainability Strategy
Reduction targets, technology benchmarking, and stewardship roadmaps through to water-positive commitments.
Advisory Services
The maps tell you where risk lives. Advisory tells you what to do about it.
Advisory turns data into decisions — from site selection to regulatory strategy. We work with operators, investors and lenders.
Site Screening & Due Diligence
Watershed availability, cooling trade-offs, grid access, and permitting exposure — before capital is committed.
Portfolio & Issuer Risk
Exposure scoring, scenario analysis, and stress testing across holdings, aggregated from facility to basin to issuer.
Policy & Regulatory
Regulatory analysis, impact modeling, and expert testimony for policy design and stakeholder engagement.
Sustainability Strategy
Reduction targets, technology benchmarking, and stewardship roadmaps through to water-positive commitments.
Methodology
How the assessment is built
Published hydrology applied at the facility level. Methods are peer-reviewed and citable.
Scope & resolution
9,500+ data centers globally, 300+ semiconductor fabs in the United States. The unit of analysis is the individual facility — not the region, country, or company.
Facility locations are precise. Hydrological supply is modeled at 30 arc-minute resolution and assigned by grid cell and basin. Facility water demand is estimated separately and layered onto modeled supply.
Hydrological model
Water availability and water gaps follow the peer-reviewed method published in Nature Communications: a monthly water balance solved from surface and subsurface runoff across climate models.
Results carry the spread across climate models rather than a single point estimate.
Methodology
How the assessment is built
Published hydrology applied at the facility level. Methods are peer-reviewed and citable.
Scope & resolution
9,500+ data centers globally and 300+ semiconductor fabs in the United States. The unit of analysis is the individual facility — not the region, country, or company.
Facility locations are precise. Hydrological supply is modeled at 30 arc-minute resolution and assigned by grid cell and basin. Facility water demand is estimated separately and layered onto modeled supply.
Hydrological model
Water availability and water gaps follow the peer-reviewed method published in Nature Communications: a monthly water balance solved from surface and subsurface runoff across climate models, with environmental flow requirements reserved using the Variable Monthly Flow method.
Results carry the spread across climate models rather than a single point estimate.
Publications
Water risk
Global water gaps under future warming levels
Establishes where water consumption exceeds renewable supply worldwide, under different warming scenarios. Water Positive assessments apply this published method at facility level.
Chemicals & fuels
Water scarcity risks in ammonia-based fertilizer production
Facility-level water risk across global ammonia plants — the chemicals and fuels sector, where site-level water stress propagates through fertilizer trade to food supply.
Thermoelectric power
Hydrological limits to carbon capture and storage
Water risk across global thermoelectric power plants, and the additional cooling demand carbon capture would impose — the same facility-level method now applied to data centers and fabs.
Solutions
Global managed aquifer recharge potential as a solution to water scarcity
Where storing high flows underground can offset unsustainable withdrawals — the supply-side response once exposure is known.
Methodology
How the assessment is built
Published hydrology applied at the facility level. Methods are peer-reviewed and citable.
Scope & resolution
9,500+ data centers globally and 300+ semiconductor fabs in the United States. The unit of analysis is the individual facility — not the region, country, or company.
Facility locations are precise. Hydrological supply is modeled at 30 arc-minute resolution and assigned by grid cell and basin. Facility water demand is estimated separately and layered onto modeled supply.
Hydrological model
Water availability and water gaps follow the peer-reviewed method published in Nature Communications: a monthly water balance solved from surface and subsurface runoff across climate models.
Results carry the spread across climate models rather than a single point estimate.
About
We built the facility-level databases for data centers and semiconductor fabs because they did not exist — and because the siting decisions being made now will bind water and power systems for decades.
We provide integrated water and energy risk intelligence at the intersection of digital infrastructure, hydrology, and climate, combining geospatial analytics, hydrological modeling, and energy systems analysis to assess where water and power are available, sustainable, and compatible — and where they are not.
Most approaches address water or energy in isolation. We map both simultaneously, surfacing the trade-offs between cooling efficiency, power availability, and water sustainability that determine whether a site is viable today and resilient in 2050.
Most approaches also stop at the country or region. Two facilities forty kilometers apart can sit in different basins with materially different exposure, and a regional score cannot tell them apart. We resolve to the individual facility, then aggregate upward — site to watershed to portfolio to issuer — so concentration risk becomes visible before it becomes a loss.
Three claimants draw on the same constrained systems: AI data centers, semiconductor fabs, and the municipal systems that host them. We work with operators and developers, credit and equity investors, lenders, policymakers and public finance analysts, communities, and mission-driven organizations — each with different stakes in the same set of problems.
Lorenzo Rosa, Ph.D.
Founder
Lorenzo Rosa is an environmental engineer and hydrologist working on the resilience of water, energy, and food systems under climate change and resource constraints. He is a Principal Investigator at the Carnegie Institution for Science, where he leads the Rosa Lab, and Assistant Professor (by courtesy) in the Doerr School of Sustainability at Stanford University.
His published research quantifies where industrial water demand collides with constrained supply — and what can be done about it — combining hydrological simulation, systems modeling, techno-economic analysis, and life-cycle assessment. The methods have been applied to ammonia production, thermoelectric power, carbon capture, and the infrastructure behind AI.
He holds a Ph.D. from the University of California, Berkeley, and B.S. and M.S. degrees in Environmental Engineering from Politecnico di Milano, and was a Postdoctoral Fellow at ETH Zurich before joining Carnegie.
Publications
Water risk
Global water gaps under future warming levels
Establishes where water consumption exceeds renewable supply worldwide, under different warming scenarios. Water Positive assessments apply this published method at facility level.
Chemicals & fuels
Water scarcity risks in ammonia-based fertilizer production
Facility-level water risk across global ammonia plants — the chemicals and fuels sector, where site-level stress propagates through fertilizer trade to food supply.
Thermoelectric power
Hydrological limits to carbon capture and storage
Water risk across global thermoelectric power plants, and the added cooling demand carbon capture would impose — the same method now applied to data centers and fabs.
Solutions
Global managed aquifer recharge potential as a solution to water scarcity
Where storing high flows underground can offset unsustainable withdrawals — the supply-side response once exposure is known.
About
We built the facility-level databases for data centers and semiconductor fabs because they did not exist — and because the siting decisions being made now will bind water and power systems for decades.
We provide integrated water and energy risk intelligence at the intersection of digital infrastructure, hydrology, and climate — geospatial analytics, hydrological modeling, and energy systems analysis, to assess where water and power are available, sustainable, and compatible.
Most approaches address water or energy in isolation. We map both simultaneously, surfacing the trade-offs between cooling efficiency, power availability, and water sustainability that determine whether a site is viable today and resilient in 2050.
Most approaches also stop at the country or region. Two facilities forty kilometers apart can sit in different basins with materially different exposure. We resolve to the individual facility, then aggregate upward — site to watershed to portfolio to issuer — so concentration risk becomes visible before it becomes a loss.
Three claimants draw on the same constrained systems: AI data centers, semiconductor fabs, and the municipal systems that host them. We work with operators and developers, credit and equity investors, lenders, policymakers and public finance analysts, communities, and mission-driven organizations.
Lorenzo Rosa, Ph.D.
Founder
An environmental engineer and hydrologist working on the resilience of water, energy, and food systems under climate change and resource constraints. Principal Investigator at the Carnegie Institution for Science, where he leads the Rosa Lab, and Assistant Professor (by courtesy) in the Doerr School of Sustainability at Stanford University.
His published research quantifies where industrial water demand collides with constrained supply — and what can be done about it — combining hydrological simulation, systems modeling, techno-economic analysis, and life-cycle assessment. The methods have been applied to ammonia production, thermoelectric power, carbon capture, and the infrastructure behind AI.
He holds a Ph.D. from the University of California, Berkeley, and B.S. and M.S. degrees in Environmental Engineering from Politecnico di Milano, and was a Postdoctoral Fellow at ETH Zurich before joining Carnegie.
Contact Us
Let's talk about your water and energy risk exposure.
Email: lorenzo@waterpositive.co