Every utility-scale BESS answers one structural question early, often without noticing it’s being answered: does the plant assemble from DC blocks — battery containers delivering DC, with power conversion procured and placed separately — or from AC blocks, integrated units where batteries and inverter ship as one product speaking grid-ready AC? It reads like a packaging detail. It’s actually a decision about supply chain, risk ownership, and how your plant will age.

For a visual companion, explore the interactive MV Skid Structure diagram on BESS.Engineer.

The two shapes

DC block architecture. Battery enclosures terminate at a DC connection; PCS units (often skid-mounted with their MV transformer) sit separately, each serving one or several DC blocks. The plant is a marriage you arrange: batteries from one maker, conversion from another, integration engineered by your team or EPC.

AC block architecture. The vendor integrates storage and conversion into one engineered product — DC bus internal, output at low-voltage AC (or with the MV transformer bundled too). You buy megawatts-that-speak-AC; the DC world is the vendor’s private business.

What DC blocks buy you

  • Supply-chain freedom. Batteries and inverters are different markets with different cost curves and different geopolitics. Sourcing them separately lets you pair the sharpest battery price with the PCS that suits your grid code — and swap either relationship at the next procurement without divorcing both. In an era of tariffs and origin rules, that flexibility has hard value.
  • Independent optimization. DC-to-AC ratios become your design variable — oversizing batteries behind a given PCS, or configuring central vs string conversion — instead of accepting a vendor’s fixed marriage.
  • Augmentation room. Adding DC capacity years later to a plant designed around a DC bus is a well-trodden path; the conversion layer can be sized for the future on day one.
The block ends at the DC terminals — conversion is a separate purchase, and so is the interface you then own. The AC block below stops one stage later.
battery racksone enclosure,or severalparalleledDC bus1500 V LFP:size at~900–1,040 VPCSPE gen-3:2×2,295 A/ 4×1,148 AunittransformerLV → MVMV collectionfeeders + busmaintransformerMV → HVHV bay +gen-tiebreaker,disconnectsPOI + meterthe boundaryDC block — stops at a pair of DC terminalsthe conversion stage — often a separate supplier'sbeyond it, nobody's by default — integrator, EPC wrap or the buyerthe DC-side guarantee is referenced here

The bill for the freedom: you own the interface. Battery-BMS-to-PCS communication, protection coordination across the DC boundary, and — most commercially — the seam in responsibility when the capacity test disappoints. Multi-vendor plants make the interface matrix and the contract structure first-class engineering documents.

What AC blocks buy you

  • One throat, one warranty. Performance is guaranteed at the AC terminals by one party; the finger-pointing seam is internalized where it can’t hurt you.
  • Compressed engineering and schedule. Integration, tested at the factory, arrives as a product rather than a site activity — attractive for lean teams and first projects.
  • Validated internals. The vendor tuned that battery to that inverter across a fleet; you inherit the tuning.
The same boundary moved downstream: you buy megawatts-that-speak-AC, and the DC world is the vendor's private business.
one drawing set, one test procedure, N buildsbattery racks~4–5 × 5MWh-class · ~5MVA at 4 hDC busenclosurePCSEPC skid:98.4% CEC-wtdunittransformerEPC skid: 99%typMV collectionfeeders + busmaintransformerMV → HVHV bay +gen-tiebreaker,disconnectsPOI + meterthe boundaryAC block — one nameplate, one drawing set, one fault-isolation boundaryoutside the block1211.5–5.3 MVA— the AC terminal — the nameplate becomes MVA2US: where IEEE 1547 / 2800 duties are measured

The bill here: coupling. Your pricing, your spare parts, your augmentation options, and your firmware roadmap all route through one vendor for twenty years, and the integrated product’s internal choices (cell supplier, PCS topology) are theirs to change between orders.

How the market actually chooses

The industry’s center of gravity has moved toward DC blocks for large projects — the same force reshaping contracts everywhere: battery prices fell fast enough that sophisticated owners wanted direct OEM battery relationships, and integrators’ bundled margins couldn’t hide. Meanwhile AC blocks hold their ground where speed, simplicity, and single-party accountability outweigh basis points: smaller projects, first-time owners, hard-schedule builds. Both are legitimate; what’s illegitimate is choosing by default. The architecture should fall out of an honest reading of your team’s integration capability, your market’s supply-chain exposure, and your augmentation plan — in that order.

FAQ

Is one architecture cheaper? DC blocks typically win on procured hardware cost at scale; AC blocks win on integration and engineering cost. Total-cost honesty requires pricing your own integration competence.

Does the choice affect performance? Well-executed, both meet the same grid requirements. The differences surface in failure modes, serviceability, and how cleanly responsibility resolves when something underperforms.

Which ages better? DC-block plants generally offer more augmentation freedom; AC-block plants age on their vendor’s roadmap. If augmentation is central to your economics, that asymmetry should weigh heavily.

Architecture selection gets a full decision framework in my Grid-Scale BESS: Complete Guide — with the interface checklists that make DC blocks safe to choose.