NEWS ANALYSIS Two batteries, two operating jobs The planned Darkhan microgrid pairs 20-minute high-power storage with roughly 7.67-hour bulk storage. Commercial operation is targeted for December 2027. Hero image: AI-generated conceptual illustration, not a photograph of an actual facility. Toshiba is preparing a battery project in northern Mongolia that assigns two different jobs to two different storage systems. At the Darkhan substation, a compact high-power battery will sit alongside a much larger energy-storage unit, helping a planned solar microgrid handle both rapid changes in output and longer shifts between electricity supply and demand. According to October 26 reporting by ESS News, Toshiba has begun manufacturing a 2 MW / 667 kWh power-type battery and is preparing the project's energy management system. A separate 2 MW / 15.33 MWh capacity-type battery will provide bulk storage alongside a 10 MW solar plant. Operation is scheduled for December 2027, so these specifications describe a planned installation, not a commissioned asset. The design deserves attention because the two batteries have the same stated power rating but radically different energy capacities. Rather than asking one battery fleet to absorb every fluctuation and deliver every extended discharge, the project separates fast power control from sustained energy delivery. One site, two very different batteries The smaller system uses Toshiba's SCiB lithium-ion cells with a lithium titanate oxide anode, commonly abbreviated LTO. Its 667 kWh energy capacity divided by its 2 MW power rating gives about one-third of an hour, or 20 minutes, of nominal discharge duration. The larger system's 15.33 MWh divided by 2 MW gives about 7.67 hours. Those calculations are nameplate comparisons, not guarantees of usable output at the grid connection. Operating reserves, conversion losses, auxiliary consumption and control settings can change the energy actually delivered. Still, the contrast makes the intended division of labor clear: one unit is sized for high power relative to stored energy, while the other is sized to sustain output much longer. Together, the published energy ratings add to just under 16 MWh. That does not establish a firm 4 MW export capability for the site. Simultaneous operation depends on the electrical configuration, interconnection limits and dispatch rules, none of which can be inferred simply by adding the two battery power ratings. Why separating the workloads matters Solar plants create more than one storage problem. Clouds can change production quickly, requiring a rapid adjustment in charging or discharging. A predictable gap between daytime generation and later demand is different: it calls for enough stored energy to maintain output over hours. Designing around only one of those needs can leave the other poorly served. At Darkhan, the power-type battery may handle frequent solar-output adjustments while the capacity-type unit handles energy shifting and broader supply-demand balancing, according to the reported project design. In principle, directing the rapid fluctuations to the smaller unit could reduce cycling stress on the bulk battery. That is an engineering objective, not a published operating result. The benefit will depend on how the controls divide actual work. If the larger battery still follows most short-term fluctuations, the theoretical separation may deliver little practical relief. If the smaller unit is repeatedly pushed against its upper or lower state-of-charge limit, it may be unavailable when a new fluctuation arrives. Keeping room to both absorb and release power is part of the design challenge. The energy management system is central Toshiba's role extends beyond the SCiB equipment. ESS News reports that its energy management system will forecast solar generation and demand, coordinate the two battery systems, and support remote monitoring and control from National Power Transmission Grid's headquarters and Darkhan branch. NPTG is Mongolia's state-owned transmission operator. Forecasting matters because the fast battery has relatively little energy behind its power rating. A controller cannot treat it as an unlimited buffer. It needs a strategy for restoring charge after a sustained intervention without creating an unwanted new ramp or consuming energy that the larger system needs for a later dispatch period. Coordination also has to work when conditions differ from forecasts. Developers and operators will need to understand what happens during communications failures, maintenance outages or reduced battery availability. Those are general commissioning questions for a hybrid plant, not disclosed shortcomings at Darkhan. A successful system must remain predictable when one part is unavailable, not only when both batteries are operating as intended. What the announcement does not establish The bulk battery is not identified as Toshiba-made in the project diagram described by ESS News. The outlet suggests it is likely to come from a third-party lithium iron phosphate supplier, but that is not a confirmed chemistry or supplier announcement. Treating the entire installation as a Toshiba SCiB system would therefore overstate the company's disclosed equipment scope. Nor does the report provide enough information to calculate a cost advantage over a single-battery design. Two battery systems can introduce more interfaces, separate maintenance requirements and additional controls work. Any savings from reduced stress on the bulk battery must be weighed against those costs across the operating life of the project. Useful evidence after commissioning would include how much energy each unit processes, how frequently each reaches its operating limits, and whether the hybrid controller meets the site's power-quality and dispatch objectives. Measured availability and degradation would be more informative than a claim that mixing battery types is inherently superior. The appropriate design depends on the duty cycle, not the novelty of the architecture. A specific test for Mongolia and battery suppliers The installation is funded through a Japanese government grant aid package, according to the report. Its substation setting and planned solar connection make it a practical test of how storage can support renewable generation within an existing power network. Mongolia's stated goal of increasing renewable energy to 30% of installed generation capacity by 2030 provides context, but the Darkhan project's contribution should not be confused with a nationwide deployment program. For battery suppliers, the project illustrates a procurement opportunity beyond selling the largest possible number of kilowatt-hours. High-power equipment can have a defined role beside energy-oriented storage when frequent adjustments and long discharge periods compete for the same asset. The value proposition then rests on measurable performance, integrated controls and a clear allocation of responsibilities between suppliers. December 2027 is the next stated operational milestone. Until commissioning and operating data are available, Darkhan is best understood as a planned demonstration of task-specific battery design. Its central question is straightforward: can a small, fast battery and a much larger energy reservoir work together well enough to justify the added complexity? The answer will come from the way the plant runs, not from its combined storage headline. Sources and reporting scope Project specifications and schedule are attributed to ESS News reporting dated October 26, 2026. Duration calculations and engineering implications are CurrentCells analysis. The original project announcement was not independently reviewed; the bulk battery supplier and chemistry are not confirmed. ESS News, October 26, 2026: Toshiba pairs 20-minute SCiB power-type battery with conventional 7.6-hour BESS for Mongolian grid