Everywhen Energy

**Product and the concrete problem it solves.** Everywhen Energy is an unusually early energy-infrastructure company focused on a problem that becomes more consequential as grids add variable renewables, storage, electrification, and large data-center loads: operators and market participants need better forecasts of what power systems will do before they commit capacity, schedule assets, or respond to volatility. Public ecosystem material describes the company as developing AI forecasting for electricity markets and grid operations, rather than selling a generation asset, a battery, or a conventional utility-management package. Its intended value is therefore decision quality. Better forecasts can help a generator decide when to dispatch, help a retailer or aggregator manage exposure, help a grid operator anticipate imbalance, and help an energy developer understand how a project will perform under changing weather, demand, congestion, and market conditions. The company is not publicly claiming a specific installed utility, revenue level, or operational outcome, so the record treats the product as an R&D-stage forecasting platform with credible infrastructure relevance, not as a proven grid-control system.

**Core technology and how it actually works.** The most specific public technical description comes from co-founder Avichai Tendler, who says the team builds agentic systems to conduct research and solve difficult problems related to power-market data. That wording suggests a stack with two connected layers: forecasting and research automation. The forecasting layer would ingest time-series and contextual data such as load, generation, weather, prices, transmission conditions, and market rules, then produce probabilistic estimates or scenario analysis for future operating states. The agentic layer could use software agents to find relevant data, formulate hypotheses, select mathematical questions, run analyses, and compare results instead of relying only on a fixed dashboard. The public record does not disclose the model family, feature set, forecast horizon, backtesting methodology, uncertainty calibration, data vendors, or whether the product writes to operational systems. This absence matters in energy: a useful model must handle regime shifts, missing data, market-rule changes, extreme weather, and the difference between correlation and dispatchable causality. Everywhen’s technical promise is therefore concrete enough to merit deep-tech attention, but independent validation of accuracy, latency, explainability, and failure behavior remains outstanding.

**Market, customers, and go-to-market.** Everywhen sits in a large and strategically important buyer landscape that includes utilities, independent power producers, renewable developers, energy traders, storage operators, demand-response aggregators, grid-service providers, and infrastructure investors. Forecasting can be sold as research software to analysts, as an API or decision-support product to commercial energy teams, or as an operational layer integrated with scheduling and asset-management systems. The company’s Startup Nation Finder profile classifies it under Energy Tech, Energy Transition, Eco-Efficient Transmission and Distribution, and Energy Management & Grid Integration, with a B2B model and an R&D-stage product. That classification is useful because it distinguishes Everywhen from consumer energy apps and from a simple weather-data vendor. A plausible initial wedge is a narrowly defined forecast that directly changes a customer decision, such as renewable-power output, imbalance exposure, or storage dispatch. The go-to-market challenge is that energy buyers demand measurable reliability and often require lengthy integration, auditability, and procurement processes. No named customer, paid pilot, utility contract, recurring-revenue figure, or market geography beyond Israel is publicly confirmed.

**Traction, funding, and third-party validation.** The company was incorporated in Israel on March 5, 2025, under the Everywhen Energy name and company number 517131769, according to public corporate-registration data. Startup Nation Finder reports one funding round in March 2025 from Grove Ventures and Elements VC, with the amount undisclosed, and describes the company as operating with one to ten employees. Grove’s public portfolio sitemap lists Everywhen among its portfolio companies, while Elements’ Israeli-backed energy and climate portfolio page displays Everywhen as an investee and frames the fund around grid resilience and reliable clean-energy deployment. These are meaningful ecosystem and financing signals: they show that the startup has passed an early investor-screening process and is connected to funds with energy, infrastructure, and deep-tech networks. They are not evidence of commercial scale. The available public record does not establish a disclosed valuation, later round, revenue, customer deployment, forecast benchmark, patent family, certification, grant, or award. Founder activity provides a further but still company-controlled signal: Everywhen is experimenting with agentic research systems and reports a roughly one-day run that generated 48 mathematical results, including claimed new proofs, counterexamples, and improved bounds. That experiment is interesting evidence of technical ambition, not proof that power-market forecasts are production-grade.

**Founders and team background.** Public sources identify Neer Saad as co-founder and chief executive officer and Avichai Tendler as co-founder and chief technology officer, although the Finder snapshot foregrounds Tendler and the public team record is still sparse. Tendler’s LinkedIn profile describes a focus on data, difficult problems, and agentic systems for power-market research; the company’s public language connects that work to AI-assisted mathematics and energy data. This combination is directionally appropriate for a forecasting startup because the hard problem is not merely fitting a model to a historical time series. It is building a reliable data pipeline, understanding market mechanics, choosing the right objective function, communicating uncertainty, and converting forecasts into decisions that survive operational scrutiny. At the same time, the public record does not establish the founders’ prior employers, utility experience, academic credentials, exits, or the size and composition of the broader engineering team. The disclosed one-to-ten-person range indicates a compact seed team. That can support fast experimentation, but it also creates key-person and execution risk: a production energy product needs quantitative research, data engineering, platform security, customer integration, and domain support, and it is not yet clear how much of that capability exists in-house.

**Competitive dynamics and potential edge.** Everywhen competes with several layers of established and emerging approaches. Aurora Energy Research and Energy Exemplar provide sophisticated market modeling and planning tools; AutoGrid, Next Kraftwerke, and Kraken Technologies provide software for forecasting, optimization, and distributed-energy operations; utilities maintain internal forecasting teams; and cloud providers offer general-purpose time-series, machine-learning, and agent frameworks. Weather-data companies and specialist renewable-forecast vendors can also compete for the first budget line. Everywhen’s potential edge is the combination of power-market specialization with agentic research automation. A system that can discover relevant relationships, test mathematical assumptions, and update its analysis as markets change could be more adaptable than a static model or a consulting workflow. Its Israeli ecosystem position may also help it recruit quantitative and engineering talent and connect with energy operators through Grove, Elements, and the broader climate-tech network. The edge is not yet a defensible moat. Incumbents have proprietary data, reference customers, integrations, and domain credibility, while open-source models can lower the cost of building generic forecasting tools. Everywhen must show that its agents improve calibrated forecast quality or reduce analyst workload in a way customers will pay for, not merely that they can produce interesting research output.

**Defense, security, and resilience dual-use relevance.** Everywhen’s core product is commercial energy intelligence, and no public source reviewed for this record establishes a defense customer, military deployment, classified program, or critical-infrastructure contract. The dual-use case is nevertheless credible through resilience rather than weapons. Electricity availability, price, congestion, and renewable output directly affect the continuity of hospitals, water systems, communications networks, emergency services, data centers, ports, and defense installations. Forecasting can support more deliberate dispatch of backup generation and storage, improve planning for islanded or semi-islanded microgrids, and help operators anticipate when weather or demand will stress a local network. An agentic research layer may also help a resource-constrained resilience team explore scenarios faster, provided that all recommendations remain auditable and human-controlled. This is a meaningful strategic adjacency because energy security is a precondition for national security, but it must not be overstated: Everywhen is not publicly documented as a defense-tech vendor, and forecast error in a commercial dashboard is not the same as safe control of a military microgrid. Dual-use scoring therefore reflects plausible transfer into energy resilience and mission-support infrastructure, not fielded defense capability.

**Growth stage, trajectory, and key diligence risks.** Everywhen is classified as early because its legal entity is roughly eighteen months old, its financing amount is undisclosed, its employee range is one to ten, and its product is publicly described as R&D-stage rather than as a scaled commercial service. The trajectory is attractive if the company can turn an agentic research capability into a narrow, auditable workflow with measurable economic value, then expand from forecast generation into planning, optimization, and portfolio decision support. The diligence agenda should be specific: (1) obtain out-of-sample backtests across multiple markets and extreme-weather periods; (2) measure calibration, not only average error, because operators need trustworthy confidence intervals; (3) verify data rights, provenance, latency, and protection against poisoned or stale inputs; (4) establish whether agents can explain their research path and reproduce their results; (5) confirm a paid design partner or live operational workflow; (6) test cybersecurity, human-approval, rollback, and model-monitoring controls before any grid-adjacent deployment; and (7) clarify financing runway, founder roles, hiring plan, patent posture, and the route from research software to recurring revenue. The principal risks are thin public traction, strong incumbents, long energy procurement cycles, model drift during regime changes, and the possibility that a capable research assistant does not translate into better dispatch decisions. Everywhen merits a strategic monitoring slot because it targets the intelligence layer of an increasingly stressed energy system, while remaining firmly evidence-seeking at the current stage.