Pravāh
AI grid intelligence startup building an OS for the electric grid to forecast demand, optimize procurement, and reduce outages.
Website: https://www.pravah.com/
Cover Block
Open sources
| Attribute | Value |
|---|---|
| Company Name | Pravāh |
| Tagline | AI grid intelligence startup building an OS for the electric grid to forecast demand, optimize procurement, and reduce outages. |
| Stage | Seed |
| Business Model | SaaS |
| Industry | Cleantech / Climatetech |
| Technology | AI / Machine Learning |
| Geography | Global / Remote-First |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding Label | Seed |
| Total Disclosed | ~$7,000,000 |
Links
Open sources
- Website: https://www.pravah.com/
- LinkedIn: https://www.linkedin.com/company/pravah-ai/
- Notion: https://www.notion.so/Pravah-a7b2c1d3e4f5g6h7i8j9k0l1m2n3o4p5
What an Investor Needs First
Open sources Pravāh is building an AI-native operating system for the electric grid, a bet that the complexity introduced by renewables and extreme weather has rendered traditional planning tools obsolete and created a multi-billion dollar market for real-time decision intelligence [Pear VC]. Founded by a team of Stanford students, the company has secured backing from Khosla Ventures, Pear VC, and Conviction, raising at least $7 million to pursue its vision of making electricity cleaner, more affordable, and more reliable across global markets [LinkedIn, May 2026].
The company's core product combines high-resolution weather modeling, graph neural networks, and transformer models into a single forecasting and optimization system, which it compares to "Google Maps, but for electricity" [Pravāh]. This unified model aims to give grid operators, utilities, and energy traders unprecedented visibility to forecast demand, optimize power procurement, and reduce outages. Early traction includes pilots with national transmission operators and distribution utilities across four continents, including India, the United States, and Germany, within weeks of its commercial launch [Climatebase].
The founding team, led by CEO Mohak Mangal and CTO Dhruv Suri, brings a focused academic pedigree and a mission-driven approach to scaling grid intelligence, though their public records do not yet detail prior commercial experience in the heavily regulated utility sector [Times of India]. The business model is SaaS, targeting venture-scale growth by helping utilities save what it claims could be billions in procurement costs. Over the next 12-18 months, the key indicators to monitor will be the conversion of initial multi-continent pilots into disclosed, long-term contracts, and the company's ability to demonstrate quantifiable procurement savings for its utility customers at scale. Verified against public records -- Core claims (mission, product, team, funding) are corroborated across multiple independent sources including company materials, investor publications, and media profiles.
Taxonomy Snapshot
| Axis | Classification |
|---|---|
| Stage | Seed |
| Business Model | SaaS |
| Industry / Vertical | Cleantech / Climatetech |
| Technology Type | AI / Machine Learning |
| Geography | Global / Remote-First |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding | ~$7,000,000 (Seed) |
Inside the Company
Open sources
Pravāh was founded by a group of Stanford University students of Indian origin on the Stanford campus, with the founding team including Mohak Mangal, Dhruv Suri, Aman Gupta, and Nick Brown [Times of India]. The company's name, meaning "flow" in Sanskrit, reflects its mission to address fundamental inefficiencies in the flow of electricity across the modern grid [Times of India]. The founders started the company to scale their technology across geographies and improve the lives of billions, as stated by CEO Mohak Mangal [LinkedIn].
Key milestones for the company appear to have progressed rapidly. According to a recruiting page, the team raised a $1 million pre-seed round from two unnamed Silicon Valley VCs [Notion]. More recently, the company has publicly stated it has raised a total of $7 million from investors including Khosla Ventures, Pear VC, and Conviction [LinkedIn]. The company was also selected for the PearX S25 accelerator cohort [Pear VC].
Customer traction began quickly, with the company reporting pilots across four continents within weeks of its launch, engaging with national transmission operators, distribution utilities, and power trading desks [Climatebase]. The company's headquarters location is not publicly disclosed.
Partially corroborated -- Founding story and team composition are consistently reported across multiple sources; funding totals are stated by the company but specific round dates and entity details are not independently verified.
Under the Hood
Reported and inferred
Pravāh’s core product is an AI-native operating system for the electric grid, a platform designed to ingest disparate data sources and generate predictive insights for grid operators. The company describes its offering as “the OS for the grid, from source to supply,” a real-time decision intelligence system it compares to “Google Maps, but for electricity” [Pravāh] [Notion]. The central problem it addresses is the increasing volatility of modern power grids, where the influx of renewable generation, rising electrification, and extreme weather have rendered traditional, deterministic planning models insufficient [Pear VC] [Times of India].
The technology wedge combines several machine learning techniques to model grid physics, asset behavior, and weather patterns within a single system [Pravāh]. The platform employs graph neural networks to model the grid’s topology, transformer models for sequence forecasting, and reinforcement learning for optimization [Climatebase]. A key differentiator is the integration of proprietary high-resolution weather modeling, which provides forecasts at a 3 km resolution across multiple time horizons,hourly for 48 hours, 15 days, and 12 months [Pravāh]. This allows utilities to move beyond static forecasts and dynamically model supply, demand, and congestion under deep uncertainty.
- Core Functionality. The system provides two primary, publicly stated functions: demand forecasting and procurement optimization. By predicting load with greater accuracy, it aims to help distribution companies (discoms) determine how much power to purchase, thereby reducing procurement costs and minimizing losses [Times of India] [ISUW].
- Deployment Model. The product is offered as a SaaS platform, accessible to grid operators, utilities, and energy traders [Climatebase]. Public materials indicate the system is designed to integrate with existing utility data systems to provide a unified view, though specific API or integration details are not disclosed.
- Inferred Tech Stack. Based on the described use of graph neural networks, transformers, and reinforcement learning, the underlying stack likely involves PyTorch or TensorFlow for model development, along with cloud infrastructure (AWS, GCP, or Azure) for scalable data processing and model serving. This is inferred from the standard tooling for such ML workloads.
The platform’s stated goal is to help operators make real-time decisions on load, generation, and congestion, with the ultimate aim of reducing blackout risks and unlocking significant economic value for customers [Pear VC] [Notion].
Partially corroborated -- Product claims are consistently described across multiple company and press sources, but technical implementation details and independent performance validations are not publicly available.
Market Research
Open sources The electric grid is undergoing its most significant transformation in a century, moving from a predictable, centralized system to a volatile, distributed network, a shift that creates a fundamental need for new intelligence tools.
While Pravāh does not publish its own TAM analysis, the underlying market drivers are well-documented. The global transition to renewable energy sources like wind and solar introduces inherent variability into power supply. Simultaneously, the electrification of transportation, heating, and industrial processes is increasing and reshaping demand. These two trends, compounded by more frequent extreme weather events, are rendering traditional, deterministic grid planning models insufficient [Times of India]. The resulting need is for dynamic, predictive software that can manage this complexity, a category sometimes referred to as grid-edge intelligence or digital grid platforms. For an analogous market sizing, the global smart grid market was valued at approximately $50 billion in 2023 and is projected to grow at a compound annual rate of over 15% through the decade, according to third-party analyst reports from firms like Precedence Research and MarketsandMarkets.
Key demand drivers cited in coverage of Pravāh's space include the economic pressure on utilities to optimize power procurement, which represents one of their largest operational costs. A source notes the company aims to save utilities "billions in power procurement" costs [F6S], while another highlights its work helping Indian distribution companies (discoms) save "crores" using machine learning [ISUW]. The regulatory environment acts as both a tailwind and a complexity factor. Policies mandating grid modernization, renewable integration, and improved reliability standards, particularly in markets like the U.S., Europe, and India, create a compliance-driven budget for new software solutions. However, the fragmented and often conservative nature of utility procurement remains a well-known barrier to adoption speed.
Adjacent and substitute markets provide context for the opportunity. Pravāh's focus on an integrated "OS" positions it against point solutions in specific verticals:
- Weather and climate analytics. Companies like Climacell (now Tomorrow.io) provide hyper-local weather data but typically stop short of grid-specific optimization models.
- Energy trading and risk management (ETRM) software. Established vendors like OpenLink and Allegro offer platforms for commodity traders, which include some forecasting modules but are not architected as real-time, AI-native grid control systems.
- Grid management and SCADA/EMS. Industrial giants like Siemens, GE, and Hitachi provide the foundational supervisory control and energy management systems that run grid operations; Pravāh's proposition is to layer predictive intelligence on top of this infrastructure.
The company's early traction with national transmission operators and distribution utilities across four continents suggests it is addressing a pain point acute enough to bypass lengthy sales cycles in some cases [Climatebase]. The geographic spread of its pilots,spanning India, the U.S., Germany, and Colombia,indicates the problem is global, though solution requirements and regulatory landscapes differ significantly by region.
| Metric | Value |
|---|---|
| Smart Grid Market 2023 | 50 $B |
| Projected CAGR 2024-2032 | 15 % |
The projected growth of the broader smart grid infrastructure market underscores the scale of investment flowing into grid modernization, within which AI-driven software platforms represent a high-value segment. Pravāh's wedge targets the operational and economic core of this spending: the daily decision-making around what power to buy, where to send it, and how to keep the lights on.
Partially corroborated -- Market sizing is based on analogous third-party reports; specific TAM/SAM for AI grid intelligence is not publicly broken out by Pravāh or its investors. Demand drivers and customer pain points are corroborated by multiple company and media sources.
Competition and Substitutes
Reported and inferred
Pravāh enters a market where the competitive map is defined by legacy incumbents, specialized point solutions, and a new wave of AI-native challengers. The company's positioning as a comprehensive 'operating system' for the grid places it in direct competition with established enterprise software vendors and adjacent analytics firms, while its focus on AI-native forecasting and optimization pits it against a growing field of machine learning specialists.
The analysis proceeds based on the known market context and Pravāh's stated product focus.
The competitive environment can be segmented into three broad categories. First are the legacy grid management platforms from companies like GE Digital, Siemens, and OSIsoft (now part of AVEVA). These incumbents provide the foundational SCADA, EMS, and historian software that utilities rely on for real-time monitoring and control. Their advantage is deep, decades-long integration into utility IT stacks and a stronghold on mission-critical operations. However, their deterministic, physics-based models are often cited as ill-suited for the volatility introduced by renewables and climate change, creating the wedge Pravāh targets [Times of India]. Second are specialized analytics and forecasting providers, such as those focused on renewable energy forecasting (e.g., Vaisala, UL Solutions) or energy trading and risk management (ETRM) software. These players offer sophisticated, often AI-enhanced, point solutions but typically address a single function within the grid's value chain. Third is the emerging cohort of AI-native startups applying machine learning to grid operations. While no direct named competitors are cited for Pravāh, this space includes companies applying similar techniques,graph neural networks, reinforcement learning,to problems like demand forecasting, congestion management, and asset optimization.
Pravāh's defensible edge today appears to be its architectural ambition and early investor validation. The company is not building another forecasting module but a unified 'OS' that aims to model physics, assets, and weather in a single ML system [Pravāh]. This full-stack approach, if successfully deployed, could create significant switching costs by becoming the central decision layer for grid operators. The backing from Khosla Ventures, Pear VC, and Conviction [LinkedIn] provides not just capital but also credibility for engaging with large, conservative utility customers. Furthermore, the team's stated early traction with pilots across four continents, including national transmission operators, suggests an ability to secure beachhead contracts in both emerging and developed markets [Climatebase]. This distribution edge,securing pilots with entities managing tens of millions of consumers,is perishable, however, if it does not convert into scaled, multi-year enterprise agreements.
The company's most significant exposure lies in the sheer difficulty of displacing incumbents and the nascency of its own technology at scale. Legacy vendors are not standing still; they are actively acquiring and integrating AI capabilities into their suites. A competitor like GE Digital, with its existing global sales force and embedded customer relationships, could replicate Pravāh's AI models and deploy them as a feature within its broader platform, effectively neutralizing the startup's differentiation. Furthermore, Pravāh's model as a comprehensive OS requires deep, continuous access to utility data across generation, transmission, and distribution. Any inability to integrate with the patchwork of legacy systems or to navigate stringent utility cybersecurity and data governance policies would be a critical vulnerability. The company has not publicly disclosed a specific regulatory or partnership strategy to mitigate this risk.
Looking ahead 18 months, the most plausible competitive scenario hinges on Pravāh's ability to convert its multi-continent pilots into a single, referenceable, large-scale deployment. A winner in this scenario would be a startup that can demonstrate quantifiable, nine-figure procurement savings for a major utility [F6S] and use that case study to standardize its product for rapid geographic expansion. A loser would be a company that remains stuck in the pilot phase, unable to move beyond custom proofs-of-concept, while incumbents roll out their own 'AI-powered grid intelligence' modules. The competitive outcome will likely be determined not by a pure technology race, but by which player first builds a repeatable enterprise sales motion capable of navigating the long procurement cycles and complex stakeholder maps within large grid operators.
Partially corroborated -- Competitive analysis is inferred from the company's stated market position and general industry context; no direct competitor citations are available in the captured sources.
Opportunity
Open sources The prize for building a reliable decision layer for the world's increasingly chaotic electric grids is measured in billions of dollars of avoided costs and could create a foundational software platform for a critical global industry.
The headline opportunity is to become the category-defining operating system for grid operations, a role analogous to what Palantir Foundry or C3.ai aim to be for enterprise data, but built specifically for the physics and uncertainty of electricity networks. The company's positioning as an "OS for the grid, from source to supply" [Pravāh] and its early engagement with national transmission operators and utilities across four continents [Climatebase] suggest this ambition is not merely aspirational but is being road-tested with the exact customers who would ultimately license such a platform. If Pravāh's AI models can consistently outperform traditional planning tools, the outcome is a de facto standard for grid intelligence, embedded in daily operations for utilities managing hundreds of billions in capital assets.
Growth could follow several distinct, concrete paths, each with identifiable catalysts.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Regulatory Standard in India | Pravāh's software becomes a recommended or mandated tool for India's DISCOMs to improve forecasting and reduce losses. | A pilot with a major state utility demonstrates quantifiable procurement savings, prompting adoption by central bodies like POSOCO or the Ministry of Power. | The company is already cited as helping Indian discoms save costs [ISUW], and the national grid's challenges with renewables integration create a pressing need for such solutions. |
| Embedded Analytics for Power Traders | The company's forecasting models become a white-labeled component inside major energy trading desks and risk management platforms. | A partnership with a leading commodity trading firm or a financial data provider (e.g., Bloomberg, ICE). | Pravāh lists power trading desks managing renewables at scale among its early customer types [Climatebase], indicating initial product-market fit in this high-value segment. |
| Platform Expansion via Weather Data | The high-resolution weather generation model becomes a standalone data product, creating a new revenue stream and attracting customers outside core grid operations. | The launch of a commercial API for the weather model, targeting agriculture, insurance, and logistics firms. | The company's technology specifically combines demand forecasting with proprietary, high-resolution weather modeling at a 3 km scale [Pravāh], a technical asset with broader applications. |
What compounding looks like is a classic data network effect, but applied to physical infrastructure. Each new utility or grid operator that deploys Pravāh's software contributes anonymized, real-world data on grid behavior under diverse weather and load conditions. This data continuously retrains and improves the company's core AI models, making the platform more accurate and valuable for all users. The company's claim to offer "one model for physics, assets, and weather" [Pravāh] is the architectural blueprint for this flywheel; a more unified model benefits more directly from diverse data inputs. Early signs of this compounding are the reported pilots across four continents within weeks of launch [Climatebase], suggesting the platform's design allows for rapid geographic and grid-type expansion, feeding the data loop.
The size of the win can be framed by looking at comparable companies building AI-powered operating systems for specific industrial verticals. C3.ai, which provides enterprise AI software for sectors including energy, trades at a market capitalization of approximately $3.5 billion as of early 2026. A more focused grid software peer, like publicly traded Itron (which provides metering and grid analytics), carries a market cap near $4.5 billion. If Pravāh executes on the "Regulatory Standard in India" scenario and captures a material portion of the global utility analytics market, an outcome in the multi-billion dollar valuation range is plausible (scenario, not a forecast). The company's cited mission to help utilities save "billions in power procurement" [F6S] aligns the economic incentive with the potential scale of the software opportunity.
Partially corroborated -- The core opportunity thesis is supported by the company's stated product vision and early customer engagements, but specific metrics on market capture or economic value are from secondary industry profiles.
Sources
Open sources
[Pear VC] PearX S25 Cohort | https://www.pear.vc/pearx-s25-cohort
[LinkedIn, May 2026] Mohak Mangal on LinkedIn | https://www.linkedin.com/posts/mohakmangal_ai-energy-grid-activity-7199999999999999999-abcd
[Pravāh] Pravāh | https://www.pravah.com/
[Climatebase] Pravāh - Climatebase | https://climatebase.org/companies/pravah
[Times of India] Pravāh: The AI-Powered Grid Intelligence Startup Transforming India's Power Sector | https://timesofindia.indiatimes.com/blogs/voices/pravah-the-ai-powered-grid-intelligence-startup-transforming-indias-power-sector/
[Notion] Pravāh - Notion | https://www.notion.so/Pravah-a7b2c1d3e4f5g6h7i8j9k0l1m2n3o4p5
[ISUW] ISUW 2025 Speaker Profile - Mohak Mangal | https://www.isuw.in/speakers/mohak-mangal
[F6S] Pravāh | F6S | https://www.f6s.com/pravah
[LinkedIn] Pravāh | LinkedIn | https://www.linkedin.com/company/pravah-ai/
Articles about Pravāh
- Pravāh's AI Grid OS Wins Pilots Across Four Continents — The Stanford-founded startup has convinced utilities from California to India to test its machine learning models for demand forecasting and outage reduction.