Jua
Jua builds the Earth Physics Transformer (EPT-2) foundation model and the Athena AI agent to deliver AI weather forecasting and decision support for energy traders, utilities, and renewable operators globally.
- Company typePrivate
- Founded2022
- HeadquartersZürich, Switzerland
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Jua does
Jua is a Zurich-based AI company founded in 2022 that builds foundation models and AI agents for physical-world simulation, beginning with atmospheric weather prediction and expanding into adjacent physics domains. Its core technology is the Earth Physics Transformer (EPT-2) family — proprietary AI weather models trained on planetary-scale data and direct satellite initialization — which the company reports outperforms ECMWF IFS, Microsoft Aurora, DeepMind GraphCast, and NOAA GFS on station-level accuracy metrics, with hourly refresh cycles and subseasonal skill out to 60 days. On top of the world model, Jua has developed Athena, an AI agent that ingests weather forecasts, market data (ENTSO-E, EEX, EPEX, BMRS/ELEXON), and portfolio positions to deliver natural language analysis, automated morning briefings, scenario testing, and backtesting for energy trading desks.
Jua firmographics
Firmographics- Name
- Jua
- Legal name
- Jua.ai AG
- Website
- https://jua.ai
- Company type
- Private
- Founded year
- 2022
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Jua builds the Earth Physics Transformer (EPT-2) foundation model and the Athena AI agent to deliver AI weather forecasting and decision support for energy traders, utilities, and renewable operators globally.
- Ownership category
- akta.pro rank
Jua industry classification
Industry- Product category
- AI Weather Forecasting / Climate Tech Software
- NAICS
- Software Publishers (5132), Custom Computer Programming Services (541511), Computer Systems Design Services (541512)
- SIC
- Services-Computer Programming Services (7371), Services-Computer Integrated Systems Design (7373)
- akta.pro primary industry
- Scenario Planning, Forecasting Platforms & Planning Intelligence (HDAAAHAL)
- akta.pro secondary industries
- Forecasting & AI Optimization for Flexibility (Load/Price/Renewables/SoC) (EUAFAKAH), C&I Load Forecasting, Scheduling & ISO/RTO Market Interface (EUAGACAH)
Keywords
Where Jua is headquartered
LocationHeadquarters
- HQ city
- Zürich
- HQ country
- Switzerland
- HQ region
- Europe
Offices1 record
Markets served
Jua business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations
Revenue model
- API/Software Subscriptions: SaaS model providing weather forecasting API access, platform subscriptions with tiered access levels (standard vs early-access subscribers). Enterprise customers get 24/7 support and named contacts.
- Usage-based API Pricing: Volume-based pricing for API access with special handling for large requests. Production access granted per workspace based on agreed volumes and use cases.
- Enterprise Custom Solutions: Custom EPT-2 or Athena adaptations for specific portfolios, regions, or workflows for enterprise partners. Enterprise support includes 24/7 technical support and dedicated contacts.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Subscription | Multi-year contract | Enterprise tier with full support and custom solutions |
Go-to-market motion2 records
Distribution channels3 records
Marketing channels4 records
Jua product offering
Product offeringCore offering
Jua builds and sells proprietary AI weather forecasting foundation models (the Earth Physics Transformer / EPT-2 family) and the Athena AI agent that delivers real-time, natural-language analysis of weather and energy-market signals for energy trading desks, utilities, and renewable operators. Customers access forecasts via REST API, Python SDK, CLI, and web dashboard under subscription, usage-based, and enterprise custom contracts, with the platform powering 100+ GW of customer portfolios globally.
Product overview
Jua's product portfolio centers on its Earth Intelligence Platform, an AI-powered decision engine for simulating and predicting physical world behaviors. The platform is built on the EPT (Earth Physics Transformer) family of proprietary weather models, the most recent being EPT-2 and its variants. EPT-2 is the flagship deterministic model delivering 20-day global forecasts; EPT-2e is the 50-member ensemble extending to 60-day (and 180-day for enterprise) probabilistic forecasting; EPT-2 Early provides earlier dissemination for time-critical trading; EPT2-RR delivers hourly-updating global forecasts; EPT2-HRRR provides 5.5km high-resolution European coverage; and EPT-2.1 Helios and Europa offer solar nowcasting and next-generation regional ensembles. On top of these models, Athena is an AI agent that monitors all weather models, price signals, and portfolio positions, delivering natural language analysis and automated briefings for energy traders. Supporting products include Morning Briefings (automated daily trading summaries), Power Forecasts (MW-level power output predictions), and the StationBench open-source benchmarking library. Access is via the web dashboard, REST API v2, Python SDK (pip install jua), and CLI.
Differentiator
Problem solved
Functional benefit
Brands
- EPT-2: Earth Physics Transformer - a foundation model that learns physics from data for state-of-the-art atmospheric prediction.
- Athena
- Earth Intelligence Platform
Products and services
- Earth Intelligence Platform Jua's flagship AI-powered decision engine that simulates and predicts physical world behaviors for energy, agriculture, aviation, and shipping sectors. Powers the EPT-2 family of weather models and the Athena agent. Provides web dashboard, REST API, Python SDK, and CLI access to all forecasting products.
- EPT-2 (Earth Physics Transformer) Flagship deterministic AI weather model delivering precise 20-day global weather forecasts at 9km resolution. Outperforms ECMWF, Microsoft Aurora, DeepMind GraphCast, and NOAA GFS on aggregate skill metrics. Four daily runs at 00, 06, 12, 18 UTC with hourly resolution for the first 10 days.
- EPT-2e (Ensemble) Ensemble probabilistic weather forecasting model extending to 60 days (and 180 days for select enterprise customers). 50 ensemble members providing uncertainty quantification. Delivers record-breaking accuracy across wind, temperature, and solar radiation at extended ranges.
- EPT-2 Early Early-dissemination version of EPT-2 delivering forecasts ahead of the market, approximately 2.5 hours earlier than the standard EPT-2 runs for time-critical advantage in energy trading. 20-day horizon, global coverage.
- EPT2-RR (Rapid Refresh) World's first hourly-updating global AI weather model delivering 24 forecasts per day vs the industry standard of 4. Direct satellite initialization with 65-85 minute latency vs ECMWF HRES ~5h 45min. 2-day horizon at 1-hour temporal resolution, optimized for time-critical energy trading and grid management.
- EPT2-HRRR (High-Resolution Rapid Refresh) First operational hourly-updating high-resolution AI weather model for Europe at 5.5km resolution. 16-member ensemble with 48-hour horizon and 24 forecasts per day. Covers Europe from 36°N to 72°N, -15°W to 30°E, resolving individual wind farms, solar installations, and local weather phenomena.
- Athena AI agent system for energy traders that monitors every weather model, price signal, and portfolio position, then delivers natural language analysis of what matters. Supports position-shaped queries, scenario testing, backtesting, automated pre-market briefings, and real-time weather impact assessment. In production globally serving more than 50 GW.
- StationBench Open-source Python library for benchmarking weather forecasting models against real-world weather station observations rather than interpolated reanalysis grids. Includes 10,000+ Meteostat stations, RMSE and skill score metrics, multi-region support, and integrations with ECMWF, NOAA, and other major forecast models.
- EPT-1.5 Previous-generation Earth Physics Transformer model providing global 9km resolution weather forecasts up to 20 days ahead. Four daily runs with high hindcast coverage back to January 2022. Still available for existing customers.
Quantifiable outcome
- 16% lower RMSE at 7 days vs ECMWF IFS, 24% lower vs Microsoft Aurora, 21% lower vs NOAA GFS
- +6 more outcomes
Companies that use Jua
Customer profileNamed customers12 records
Segments6 records
Ideal customer profiles2 records
Jua technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration25 records
AI capability7 records
Feature7 records
Jua partnerships and signals
Strategic signalPartnerships
Four partnerships are on record, tiered core and strategic.
- Origin EnergycorePartnership under the Free Electrons program to incorporate AI-driven weather forecasting into Australia's largest energy company's operations. Origin Energy operates across generation, storage, and retail in Australia.
- Free ElectronscoreGlobal energy tech accelerator program connecting startups with major utilities. Jua signed a deal with Origin Energy under this program.
- Vast DatastrategicInfrastructure partner providing storage solutions that enabled scaling of AI weather model training infrastructure. Referenced in case study with Crusoe Cloud.
- Crusoe CloudstrategicCloud infrastructure partner for training Jua's models at scale. Case study published on scaling AI weather models with Crusoe's GPU cloud.
Scale indicators8 records
Recent moves6 records
Expansion highlights6 records
Jua competitors and assessment
Company assessmentDirect peers
- Planetrics (S&P Global): Climate and weather analytics platform purpose-built for energy markets, providing power demand forecasting, renewable output forecasts, and scenario analysis. Directly comparable in the energy trading and utility customer segment with portfolio-shaped weather analytics.
- Climavision: AI/ML-based weather forecasting company combining proprietary radar, satellite, and numerical weather prediction for utilities, agriculture, and government. Comparable business model of building AI-enhanced weather products for enterprise verticals including energy.
- Salient Predictions: Subseasonal-to-seasonal forecasting company focused on long-range weather prediction for agriculture, energy, and water management. Comparable in the subseasonal forecasting niche where Jua's EPT-2e competes, and similar target verticals.
- Tomorrow.io: AI-driven weather forecasting platform targeting enterprise customers in energy, aviation, and logistics with proprietary radar and satellite data. Directly comparable to Jua in product strategy (API-first AI weather models for enterprise verticals) and target customer base (utilities and trading desks).
- Meteomatics: Swiss high-resolution weather data provider serving energy, aviation, and insurance with global forecast models. Direct geographic and product overlap with Jua, also headquartered in Switzerland and targeting European utilities with high-resolution forecasting.
Broad incumbents
- ECMWF: European Centre for Medium-Range Weather Forecasts is the institutional incumbent that defines the field. Its IFS, ENS, and AIFS models are benchmarked against Jua's EPT-2 in every technical report; ECMWF represents both a competitor and a data/compute partner for the European weather industry.
- The Weather Company (IBM): IBM-owned weather data and forecasting business serving media, aviation, energy, and logistics. Established incumbent with broad enterprise footprint, comparable forecasting-as-a-service model, and significant distribution through IBM's enterprise sales channels.
- DTN: Established weather and environmental intelligence provider serving agriculture, energy, aviation, and logistics with operational forecasting and decision-support tools. Broader, more established incumbent in the same enterprise weather intelligence space Jua targets.
- Microsoft Aurora: Microsoft Research's AI foundation weather model that Jua benchmarks against directly. While not a standalone company, Microsoft's AI weather efforts represent a direct competitive threat through Azure cloud distribution to enterprise customers.
Others
- Bureau of Meteorology / NOAA (public sector): Government meteorological agencies provide foundational weather data and baseline forecasts that Jua builds upon. They are upstream data sources and indirect competitors whose GFS and IFS outputs are integrated into Athena's model catalogue but also compete as alternatives for less sophisticated buyers.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat6 records
Key risks6 records
Key highlights7 records
Customer concentration
Jua social profiles
Digital presenceJua financial estimates
Financial estimateRevenue estimate
Valuation estimate
Jua leadership team
Management profileNumber of profiles
Profiles8 records
Jua funding detail
Funding detailFunding overview
Funding rounds4 records
Investors11 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Jua M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Jua
What does Jua do?
Jua builds and sells proprietary AI weather forecasting foundation models (the Earth Physics Transformer / EPT-2 family) and the Athena AI agent that delivers real-time, natural-language analysis of weather and energy-market signals for energy trading desks, utilities, and renewable operators. Customers access forecasts via REST API, Python SDK, CLI, and web dashboard under subscription, usage-based, and enterprise custom contracts, with the platform powering 100+ GW of customer portfolios globally.
Is Jua a public or private company?
Jua is a private company. It is classified as venture growth investor backed and is currently operating.
When was Jua founded?
Jua was founded in 2022. It employs 11 to 50 people.
Where is Jua based?
Jua is headquartered in Zürich, Switzerland, in the Europe region.
How does Jua make money?
Three revenue lines are on record. API/Software Subscriptions are the primary driver. The others are usage-based API Pricing and enterprise Custom Solutions.
Who are Jua's main competitors?
Direct peers on record are Planetrics (S&P Global), Climavision, Salient Predictions, Tomorrow.io and Meteomatics. Broad incumbents are ECMWF, The Weather Company (IBM), DTN and Microsoft Aurora. Bureau of Meteorology / NOAA (public sector) is listed as an others.
Does Jua have an API?
Yes. Jua offers a REST API (v2) for weather forecast data access, a Python SDK (pip install jua) for programmatic access, and a CLI tool. The Query Engine API supports flexible geospatial queries, temporal filtering, aggregation, and multiple response formats including JSON, Apache Arrow, and streaming. Access methods include Web Dashboard, REST API, Python SDK, and CLI. Authentication via developer portal with rate limiting. Early-access subscribers can receive forecast data up to 30 minutes earlier than standard tier. Developer documentation is at docs.jua.ai.
What industry is Jua in?
Jua's product category is AI Weather Forecasting / Climate Tech Software. Its primary akta.pro industry code is HDAAAHAL, Scenario Planning, Forecasting Platforms & Planning Intelligence, with a secondary code of EUAFAKAH, Forecasting & AI Optimization for Flexibility (Load/Price/Renewables/SoC). Its NAICS code is 5132 and its SIC code is 7371.