Although battery energy storage systems (BESS) have become a headline topic, the real market transformation will not be driven by hardware deployment alone. Equally important is who operates the assets, with what software, across which markets and at what speed. According to sight‑E analytics, one of the keys to future energy markets is multi‑market optimisation: simultaneously forecasting prices, managing technical constraints, choosing between day‑ahead, intraday and balancing markets, and executing those decisions automatically.
Why the power market is turning into a prediction business
Dolányi Mihály, CEO of sight‑E, says the market structure is changing fast: more weather‑dependent generation, higher price volatility, and a growing number of assets whose operation cannot be managed with static rules. Battery storage in particular forces continuous trade‑offs — when to charge, when to discharge, which markets to serve, how much capacity to reserve for balancing — making trading a prediction and rapid‑reaction problem. Intraday changes (weather, production, consumption, market prices, system needs and activations) frequently invalidate plans, so optimisation must be continuous rather than a one‑time day‑ahead schedule.
sight‑E describes its solution as end‑to‑end: price forecasting, position optimisation, algorithmic trading and automated load management, executing roughly 150,000 to 1,000,000 decisions per asset per year.
Hungary’s position and regional outlook
Hungary is roughly two to three years behind the most advanced multi‑market implementations, in part because of a conservative market culture and past market failures that weakened trust. Yet the Hungarian market’s dynamism also makes it a strong training ground: teams that learn to operate effectively here can gain an advantage when expanding to other markets. Some regional markets, for example Romania, have already seen earlier adoption of similar models.
Software’s growing value and market effects
Possessing an asset is becoming less decisive than operating it well. sight‑E cites estimates that the full optimisation software market could reach about $13 billion by 2030, and that the BESS software market in Europe may grow faster than the hardware market.
Balancing markets, notably aFRR, may saturate as more flexible capacity enters. Some scenarios estimate that by the end of the decade up to 90% of storage revenues might come from arbitrage; other analyses suggest balancing and other revenue streams could still make up 60–70%. sight‑E therefore stresses multi‑market optimisation so strategies can be re‑weighted as markets evolve.
What gains are possible with good optimisation?
Using day‑ahead arbitrage as a reference, sight‑E reports that combining better optimisation with capacity payments and activation energy fees can yield up to 2.5‑fold returns versus simple arbitrage. However, relative margins typically shrink as asset or portfolio size grows: above roughly 100 MW it becomes harder to sustain the same percentage uplifts.
Batteries co‑located with PV plants are especially interesting: a storage unit can provide standalone revenues while sometimes reducing system charges if it charges directly from the PV plant rather than via the main meter. For feed‑in‑tariff (KÁT) and non‑KÁT PV alike, that can be a material revenue booster and also make financing easier by pairing stable KÁT income with the storage project’s upside.
How sight‑E’s system differs from generic “AI”
Dolányi emphasizes that their platform is not a large language model solving the control and trading problem by itself. Instead, sight‑E runs a complex optimisation engine built from mathematical models, neural networks, forecasts and decision logic. Large language models might help indicate directions, but they do not replace the optimisation core. The system’s decision module determines which market each storage asset participates in, how it charges and discharges, and in what cadence — sometimes preferring extended discharge profiles over rapid full discharge because that delivers higher value in a given market situation.
With systems able to make hundreds of thousands to a million decisions per asset per year, manual intervention does not scale. sight‑E therefore aims for automation within predefined risk and operational thresholds, reserving human intervention for cases where tolerances are breached rather than for instinctive overrides.
Learning from mistakes, local adaptation and in‑house development
Learning from algorithmic errors is crucial: it’s not enough to see a suboptimal decision — operators must understand which forecast, market signal or technical constraint led to it, to refine the models. sight‑E’s platform operates adaptively in real time, accounting for RHD, losses, exchange fees, degradation, minimum profit thresholds, technical constraints and warranty limits.
In‑house development matters because local market rules, liquidity, regulatory logic and technical norms differ. Off‑the‑shelf software built for other markets often needs substantial adaptation; owning the software enables faster responses to regulatory or market changes. sight‑E also aggregates individual plant‑level optimisation into portfolio‑level control.
Team and final message for market participants
Such ventures require cross‑disciplinary teams combining energy expertise, maths, data science and software engineering rather than isolated skill silos. sight‑E currently operates with a small team of about 6–8 people.
Their main recommendation to industrial and market players considering storage, flexibility or aggregator partnerships is that passive operation is no longer sufficient. Realising the value of flexibility requires data, forecasting, optimisation, market access, automated execution and transparent settlement. Assets must be adjusted on daily, hourly and even sub‑second timescales to capture value; those who can do this will turn today’s volatile power markets into revenue opportunities.
This article is not investment advice or a recommendation.
