-
Framework: integrated vs. assembled
-
Dimension 1: System integration
-
Dimension 2: Monitoring and software
-
Dimension 3: Battery selection and true cost
-
Dimension 4: Warranty and support
-
Dimension 5: Future expandability
-
The industry has shifted — but integration discipline still matters
-
How to choose — based on situation, not brand loyalty
If you just got a solar-plus-storage quote — or you're pulling quotes from three vendors for a project spec — there's a good chance you're staring at two very different builds. One puts everything under a single brand: inverter, battery interface, monitoring, all handled by one manufacturer. The other assembles best-of-breed parts from different vendors. Both work. But the things you need to check before signing off are completely different.
Here's the comparison I run through, directly, across five dimensions.
Framework: integrated vs. assembled
The comparison below uses two concrete archetypes.
Archetype A is the SMA ecosystem route — a SMA Sunny Boy 7.7-US grid-tied inverter as the anchor, an AC-coupled storage inverter on top of it, and Sunny Portal handling monitoring end to end.
Archetype B is the assembled route — a third-party hybrid inverter, a set of 48V rack mount LiFePO4 batteries, a separate EV wallbox (the E.ON Wallbox is a common example), with each subsystem reporting through its own app.
I've specified both. Neither is universally better. The question is which one removes risk for your situation.
Dimension 1: System integration
Archetype A (SMA): The Sunny Boy 7.7-US speaks SunSpec, so the handshake between inverter and grid interface is a known quantity. When paired with storage on the same ecosystem, grid state, battery state-of-charge, and backup transfer logic all talk through one protocol. That matters — for anyone signing the commissioning report, that alone eliminates a whole category of rechecks.
Archetype B (assembled): The hybrid inverter and the 48V LiFePO4 battery bank communicate over CAN bus. The wallbox was designed to talk to the grid and a simple solar meter; getting it to read battery data or inverter output is solvable but not guaranteed. "Works on paper" and "works at 2 a.m. on day 400" are not the same thing.
Here's something vendors won't tell you: the most common integration ticket is not a broken component. It's two individually healthy components that don't cooperate. I've reviewed a system where a battery BMS and an inverter both charged correctly under normal conditions — until the BMS firmware and the inverter firmware used slightly different definitions of "end of charge" and drifted apart by 0.3V over three months. Battery bank aged prematurely. Neither vendor considered it their defect. The installer ate the replacement.
I'm not an electrical engineer, so I can't speak to the exact harmonic behavior behind that drift. What I can tell you from a quality audit perspective is that the failure was invisible in the design review and obvious in the site visit six months later. That's the pattern you want to design around.
This doesn't disqualify assembled systems. It just means assembled systems only stay reliable if the person installing them (or checking the design) actively validates the pairings. If that 3-hour commissioning step isn't protected in the schedule, it becomes a future service call with a higher price tag.
Dimension 2: Monitoring and software
Archetype A: Sunny Portal shows production, storage activity, fault history, and remote diagnostics in one interface. One login, one report, one support call.
Archetype B: Separate apps for the inverter vendor, the battery BMS, and the wallbox. Different update cycles. Different alert policies. Different levels of data history.
When a customer calls and says "something seems off," my first 15 minutes go to figuring out which app sent the alert. Not ideal, but workable.
To be fair — the standalone apps in an assembled system are often functionally deeper than a single integrated dashboard. There are things a battery vendor's app exposes that no consolidated portal will match. You just have to accept that you'll be reading three of them.
Dimension 3: Battery selection and true cost
This is where Archetype B genuinely wins on sticker price.
For a turnkey, brand-integrated residential battery system, published benchmarks put installed cost in the $1,000–$1,500/kWh range (based on NREL 2023 cost benchmarks and EnergySage 2024 market data; verify current pricing with your installer). That's the typical band for fully managed, single-vendor systems.
Rack mount LiFePO4 changes the math. Raw 48V rack units (typically around 5.12 kWh per module) price out roughly $250–$350/kWh for the batteries alone, before inverter, electrical work, labor, and permits. Once you add a hybrid inverter and installation, an assembled 10-kWh system often lands at $800–$1,100/kWh. On a 10 kWh build, that's easily $2,000–$4,000 of spread versus the integrated route.
But "cost per kWh" and "cost of ownership" are two different numbers.
I made the classic rookie mistake in my first year on this side of the desk: I assumed "compatible" meant the names matched on a spec sheet. We ran with a rack mount LiFePO4 setup on a light-commercial project because the price was right. Ten months in, the comms protocol between the BMS and the inverter started dropping packets intermittently. No fault code. No alarm. But the backup transfer time went from ~20 ms to over 2 seconds — yes, seconds, not milliseconds — during a grid event. Customer's IT rack dropped. Three weeks of diagnostics later, we replaced half the comms wiring. Original factory position was that everything was "operating normally." The redo cost us about $6,000 plus the downtime. A lesson learned the hard way.
So now I require one thing on any battery spec, regardless of brand: the integrator must give written clarity on the BMS comms protocol version, the firmware update timeline, who is responsible for updates, and who physically shows up if something fails.
Spec sheets describe what the system should do. Contract terms are what protect you when it doesn't.
Dimension 4: Warranty and support
Archetype A: One warranty chain. SMA's Sunny Boy line in the U.S. typically carries a 10-year standard warranty with extended options, and the storage side is tied into the manufacturer's compatible-battery program. One number to call. One qualified field service network.
Archetype B: Three warranties, three support desks. The first question on any service call is "which device is failing?" — and nobody on the customer side knows the answer.
I ran into this last Q4. A customer's assembled system wouldn't charge the EV in backup mode. Inverter OEM said backup output was normal. Battery OEM said discharge curve was normal. Wallbox OEM said no fault log. All three devices healthy, system still non-functional. Root cause, eventually: the wallbox firmware hard-coded a specific zero-crossing detection profile that didn't match the waveform quality on the inverter's backup output. Not a defect on any single vendor's side. The customer waited 26 days for restoration.
Whether you go integrated or assembled, the question "who owns this system end-to-end" matters more than any parameter table.
Dimension 5: Future expandability
Archetype A: Expansion is bounded by the SMA ecosystem. Adding storage or additional capacity has a defined path, but you're committed to compatible hardware. Model-generation changes require a per-case compatibility check. The advantage is that there is an official path forward.
Archetype B: More flexible. Going from 10 kWh to 30 kWh is often a matter of stacking more rack mount modules and staying within the hybrid inverter's AC coupling limits. Past a certain size — usually around the hybrid inverter's total capacity ceiling — you hit a pretty hard wall.
In practice, most residential sites cap out well below where Archetype A becomes restrictive. The expandability edge matters much more on light-commercial jobs or sites with firm backup-power requirements.
The industry has shifted — but integration discipline still matters
If you'd asked me in 2021, my recommendation would have been simple: go integrated, and don't look back. Third-party rack mount LiFePO4 quality was inconsistent, and there was almost no cross-vendor compatibility data to work from.
2025 looks different. Tier-1 cell suppliers tightened consistency considerably since the 2021 era. BMS firmware now handles multi-string configurations in ways that would have been edge-case territory three years ago. Third-party wallboxes like the E.ON Wallbox have opened up their integration APIs (not fully open, but no longer a black box). The landscape has genuinely moved.
The fundamentals haven't changed. A solar storage system is delivered as a system, not a component list. Failure anywhere degrades the whole operating envelope. That was true in 2021 and it's still true in 2025.
SMA's own shipping data reflects the scale of the market that's still moving through integrated channels — the company reported roughly 20.5 GW of inverter shipments in 2023, according to its annual report (SMA Solar Technology AG, 2023). That's a substantial jump from the prior year's figure, and it says the integrated route isn't shrinking. What has changed is that the assembled route is now a genuinely credible alternative for the right project. That's the real shift.
How to choose — based on situation, not brand loyalty
Here's how I split it in reviews:
- Go SMA ecosystem (Archetype A) when: residential or light-commercial, single site at 20 kWh or less, customer wants one support number, interconnection approval needs to be straightforward, and someone on site actually reads the monitoring dashboard. The Sunny Boy 7.7-US is a solid anchor for the grid-tied side — but note it's not a hybrid unit, so if you plan on adding storage within the next 18 months, plan the AC-coupled architecture from day one rather than retrofitting.
- Go assembled (Archetype B) when: you need 30 kWh+ within five years, the customer already has a wallbox brand lock-in (E.ON Wallbox or similar), your electrician will commit 3–4 extra hours to comms validation, the customer doesn't care about app sprawl, and budget pressure is the dominant constraint.
My one hard line: assembled works, but only when someone owns integration at the system level. Installation, commissioning, sign-off, and long-term support responsibility. If that role is undefined, whatever the assembled system saves on day one will come back as a service ticket within two years.
Our project group tracks commissioning cycle performance quarterly. Integrated builds close out about 40% faster on first-pass commissioning than assembled builds — and that doesn't count the callback cost from compatibility issues. That's been consistent across the last six quarters.
That said — if you or your installer know the components cold, your electrician is reliable, and your project timeline has buffer, the assembled route does deliver real savings per kWh. There isn't a universal answer here.
And if the underlying question is "how much does a solar battery system cost" — my first counter-question is: are you pricing a brand-integrated turnkey system, or an assembled system? The gap between those two numbers is routinely 30–45% per installed kWh. Comparing across architectures before you've picked one isn't really a comparison. Lock the architecture, then compare prices.
Prices referenced are as of early 2025, based on NREL cost benchmarks, EnergySage market data, and vendor-quoted ranges. Verify current pricing and specifications with your installer before committing to a design.