Zendure SolarFlow 2400 PRO & AC+ review: the technical deep-dive

Key takeaways

• The SolarFlow 2400 PRO and AC+ share the same chassis and the same 2400W bidirectional inverter -- the difference is on the back
• ZENKI 2.0 AI combines energy prices, weather forecasts, and your consumption pattern -- Zendure claims 73% more yield than standard CT mode
• The RJ45 port is not an ethernet connection -- it's Modbus-RTU for the CT meter
• Round-trip efficiency measured at ~87-88% with our power analyzer -- in line with specifications
• FLIR measurements show even heat distribution under load -- a good sign for longevity
• Hardware is premium (metal, GaN/SiC, IP65, aerosol fire suppression). Software is the real differentiator
• Home Assistant integration is limited -- a consideration for HA enthusiasts

Why this review

Zendure sent us two brand-new 2400 Watt home batteries to test: the SolarFlow 2400 PRO and the SolarFlow 2400 AC+. Important to mention upfront: we received no compensation for this, no agreements were made about the content, and Zendure has not seen this review before publication.

What interested me about these products: two models with exactly the same power, but a completely different approach. One connects directly to solar panels. The other to your existing rooftop installation. I wanted to find out what that means in practice -- not just on paper.

This article is the technical deep-dive with all the data that doesn't fit in video format. The visual review will appear soon on ThuisbatterijNederland.

The problem Zendure is trying to solve

Before we dive into the hardware, some context. With net metering ending on January 1, 2027, self-consumption of solar power is becoming increasingly important. During the day you produce surplus that goes onto the grid -- soon at a fraction of what you currently receive for it.

A home battery captures that surplus and delivers it back to your home in the evening. But most home batteries require a hybrid inverter, professional installation, and an investment of 5,000 to 15,000 euro. Plug-in batteries like the Zendure offer an alternative: plug into a wall outlet, configure the app, done. No installer needed for the basic setup. That lowers the barrier -- but how does it perform in practice?

The unboxing: first impressions

The boxes arrived on a pallet, and that's no exaggeration. The PRO and AC+ each come in a sturdy brown shipping box with thick polystyrene inner packaging. What immediately stood out: the weight. You're lifting about 25 kilos per unit out of the box. That's hefty for something you set against a wall in your shed, but it immediately gives a sense of solidity. This is no plastic toy.

In the PRO box you'll find the unit itself, an AC power cord, the Quick Start Guide, safety instructions, and a set of mounting materials for wall installation. What's not included and what you need to purchase separately: the CT meter (unless you buy a bundle), the AB3000L battery modules for expansion, and the MC4 cables to your panels. The AC+ box is identical, minus the MC4-related documentation.

The first thing you notice when you take the unit out of the packaging: the finish. The chassis is fully metal with a matte black coating that feels sturdy. The top has a subtle Zendure logo. The heatsink construction is integrated into the side panels -- you can feel the fins as you run your hand along them. This isn't a cosmetic detail; it's functional thermal design.

What surprised me in the comparison: the AC+ is not a "stripped-down" PRO. Turn both units around and the difference is immediately clear. The back of the PRO has four MC4 connector pairs, neatly recessed into the chassis. The AC+ has a smooth, finished rear panel there. You don't pay for hardware you don't use -- a deliberate design choice I appreciate.

Form factor and weight vs. predecessor

I placed both models next to the older SolarFlow AC2400. The first thing that stands out: the new units are more compact in depth. The footprint is comparable in width, but everything has been tightened up. That makes placement easier, especially against a wall or in an electrical cabinet. The weight of the unit itself is comparable to the AC2400, but with the standard 2.4 kWh built-in storage you have capacity right away -- with the previous generation you had to buy separate battery modules.

The AB3000L battery modules

The expansion modules deserve their own section, because a large part of the value proposition lies here. Each AB3000L offers 3072 Wh of usable capacity and uses LFP chemistry (lithium iron phosphate). LFP is the gold standard for stationary storage: thermally more stable than NMC, longer lifespan, and no cobalt in the supply chain.

The specs per module: 3072 Wh, LFP, IP65-rated, with a cycle rating of 3000 cycles above 80% remaining capacity and 6000 cycles above 70%. With one full cycle daily, you're looking at 8 to 16 years of lifespan. With partial cycles -- which is what happens most in practice -- even longer.

Physically the modules are solid. I measured them and they each weigh around 30 kilos. Lifting is a two-person job, or very carefully solo. The modules stack on the main unit via lockable connectors. You can stack up to four AB3000L modules for a total capacity of 14.7 kWh (including the built-in 2.4 kWh), or a maximum of 16.8 kWh in the fullest configuration.

The connection between modules is plug-and-play: push in the physical connector, turn the lock, done. The app detects new modules automatically. During my test I added an AB3000L to the PRO and the system recognized the module within two minutes. No firmware update needed, no reconfiguration.

Each module has the ZenGuard fire suppression system built in -- an integrated aerosol fire retardant that activates automatically during thermal runaway. Combined with cell-level monitoring in the BMS, this is one of the more comprehensive safety systems I've seen in plug-in home batteries.

Two models, one platform

Specification comparison

The PRO is built for those who want to connect their own panels that aren't on an existing inverter -- ground mount in the garden, second set of panels on a shed, or a complete off-grid system. With four independent MPPT trackers and 3000 Wp solar input you have serious capacity.

The AC+ targets a completely different scenario: you already have panels on the roof feeding into the grid via an inverter, and you want to add a battery without touching your existing installation. Purely AC-in, AC-out.

Installation: from box to running system

Let me walk step by step through how the installation went, because this is where the difference lies between "plug-and-play" in the brochure and reality.

Step 1: Physical placement (10 minutes). I placed the AC+ in my utility room, on a concrete floor against the exterior wall. The PRO is in the shed, closer to the point where I want to connect solar panels later. Both units have anti-slip feet on the bottom. Wall mounting is possible with the included brackets, but at the weight of these units you want to be sure your wall can handle it.

Step 2: CT meter installation (30 minutes). This was the most time-intensive part. The Zendure Smart Meter 3CT needs to go in your electrical panel, with the current clamps around your phases. If you're handy with your electrical panel this is no problem -- but I understand not everyone is comfortable with this. The CT meter connects via the RJ45 cable to the main unit. Note: this is Modbus-RTU, not ethernet.

Step 3: WiFi configuration and app pairing (15 minutes). The Zendure app walks you through the process. Scan the QR code on the unit, select your WiFi network, enter the password. For me this worked on the first try, but I've read that some users have issues with 5 GHz networks -- the unit only supports 2.4 GHz WiFi. Make sure you have a 2.4 GHz network available.

Step 4: ZENKI activation (5 minutes). In the app you select your energy supplier from a list of 840+ European providers. I use a dynamic contract, so ZENKI could immediately start working with hourly prices. You set your preferred mode: fully automatic (ZENKI decides everything), or manual with preferences. I chose fully automatic to see what the system does on its own.

Total installation time: about one hour. Honestly smoother than expected. The only thing I ran into: the app requested a firmware update right after pairing, which took about eight minutes. Not dramatic, but it felt like an unnecessary hurdle at the moment you're eager to get started.

ZENKI 2.0: the AI brain behind the battery

This is where Zendure sets itself apart from the competition, and the part I dove deepest into.

ZENKI 2.0 is not a marketing term. It's an AI-driven energy management system that analyzes three data sources simultaneously:

1. Current and predicted electricity prices from your energy supplier -- supports 840+ European providers
2. Weather forecasts -- how much sun does the system expect in the coming days?
3. Your historical consumption patterns -- when do you use energy and how much?

Based on those three factors, ZENKI decides when the battery charges, when it discharges, and when it deliberately waits. The system even reserves capacity before a price dip occurs.

A day in the life of ZENKI

To make concrete what ZENKI does, I recorded a specific test day in detail. This was a weekday in May with partly cloudy weather and a dynamic energy contract.

06:00 - 07:30: Morning peak at home. Coffee machine, toaster, shower. Electricity price: 0.28 euro/kWh. ZENKI discharged the battery at about 800W to cover household consumption. Smart: it knew the morning peak was coming and had already charged the battery overnight.

07:30 - 11:00: Quiet morning. Electricity price dropped to 0.12 euro/kWh. Cloudy. ZENKI held the battery at 45% and did nothing. It waited.

11:00 - 14:30: Sun broke through. On the PRO the panels kicked in and produced 1,800W. ZENKI charged the battery from 45% to 92%. It deliberately stopped at 92% -- not at 100%. Why? It expected two more hours of sun based on the weather forecast and wanted to keep room.

14:30 - 17:00: Cloud cover increased. Production dropped to 400W. ZENKI discharged lightly (300W) to compensate for the difference between production and consumption. Battery dropped to 78%.

17:00 - 21:00: Evening peak. Electricity price rose to 0.34 euro/kWh. ZENKI discharged fully -- 1,500 to 2,000W depending on consumption. Cooking, oven, television, lighting. The battery dropped from 78% to 18%.

21:00 - 23:00: Price dropped again. ZENKI did nothing. It knew a price dip was coming overnight.

23:00 - 05:00: Overnight dip. Electricity price: 0.04 euro/kWh (nearly free). ZENKI charged the battery slowly from 18% to 85%. It deliberately chose slow charging -- less wear on the cells.

This pattern repeated daily with variations based on weather and prices. What stood out to me: the decisions were consistently logical. The system made no odd choices. It didn't charge in the middle of the day at high prices. It didn't discharge at night when there was no consumption.

The 73% claim under scrutiny

During my testing period of several weeks I couldn't verify this over a full season. What I can say: the decisions the system made were logical and consistent. The logic is solid -- if you consistently charge when electricity is dirt cheap and discharge when the price peaks, it adds up. Especially with a dynamic contract. But 73% is an average across a population, not a promise for your specific situation.

HEMS 2.0: more than just a battery

Via HEMS 2.0 ZENKI can control up to six appliances per configuration -- including heat pumps via a Shelly relay and Teslas via direct platform integration. In practice this means: your heat pump runs when electricity is cheap, your car charges overnight at the lowest rate, and your battery manages the hours in between.

The CT meter options

You need one of the following for ZENKI to work optimally:

• Zendure Smart Meter 3CT (included option with bundles)
• Shelly Pro 3EM
• HomeWizard P1 (popular in the Netherlands)
• EcoTracker P1

The meter continuously measures your total consumption and production at phase level. Without that data ZENKI lacks context and cannot optimize.

Build quality and safety

Materials and construction

What immediately stands out when you hold the unit: this is not a plastic box. The SolarFlow 2400 has a full metal chassis with a heatsink construction that doubles as the enclosure. That's functional -- the metal conducts heat away from the internal components.

Under the hood are GaN and SiC components (gallium nitride and silicon carbide) for power conversion. These semiconductor materials are more efficient than traditional silicon transistors and generate less heat at the same load. You see that reflected in the efficiency numbers and in the FLIR measurements.

Off-grid functionality

During a power outage the SolarFlow 2400 switches to off-grid mode. In my tests this functioned reliably -- the switchover took about two seconds. Short enough to keep most appliances running, but too long for sensitive electronics that should be on a UPS. The PRO has an advantage here: it can charge directly from solar to battery without the grid, meaning you can replenish your battery during a prolonged outage.

Efficiency tests: the real numbers

This is the core of this article -- the data that doesn't fit in video format.

Test setup and methodology

I tested both models with a ZLCP168008 power analyzer. This device simultaneously measures AC input power and AC output power, allowing you to calculate the actual round-trip efficiency: how much energy goes in, how much comes out.

The test methodology:

Each test consisted of a full charge-discharge cycle. I measured a minimum of five complete cycles per model to verify consistency. Tests were conducted at room temperature (21-23 degrees Celsius) in a well-ventilated space. I measured at three load levels: 500W (light), 1200W (medium), and 2000W (heavy).

What can influence the results: Ambient temperature plays a role -- at lower temperatures LFP cells perform slightly less efficiently. The state-of-charge (SoC) at the start of the test also matters; I consistently started at 10% SoC and charged to 95%. The first and last 5% of the SoC range is inherently less efficient, so by avoiding those you get a more realistic picture of daily use.

Results

These numbers are in line with what Zendure specifies -- no marketing inflation. A round-trip efficiency of 87-88% is competitive for home batteries in this price range. The GaN/SiC components are doing their job here.

For comparison: traditional systems with silicon MOSFETs typically achieve 82-86% round-trip. That 2-5 percentage point difference sounds small, but over thousands of cycles it adds up. With a daily cycle of 5 kWh and an electricity price of 0.25 euro/kWh, 3% efficiency difference saves you roughly 14 euro per year. Over the battery's lifespan that's 100 to 200 euro.

What I also noted: efficiency was highest at medium load (1200W). At 500W the inverter overhead was relatively larger, at 2000W heat generation increased and efficiency dropped marginally. In practice ZENKI runs the system mostly in the mid-range, so the 87-88% is representative of realistic use.

FLIR thermal analysis

With a FLIR thermal camera I mapped the temperature distribution under load. This is information you can't convey in video -- the details in the thermal images tell a story about engineering quality.

Findings at 2000W continuous load after 45 minutes:

What do these temperatures mean?

The 43 degrees at the converter module is an excellent result. LFP cells degrade faster above 45 degrees, and GaN/SiC components can handle up to 150 degrees, so there's enormous margin. For comparison: with a competitor I tested earlier (silicon-based) I measured 58 degrees at the same load. That 15-degree difference translates directly to longer lifespan.

The 32 degrees at the battery cells is even more impressive. The cells barely warm up, which means the thermal decoupling between the inverter section and the battery section is well designed. This is a design choice you only see the payoff of in the long-term degradation figures, but it's precisely the kind of detail that makes the difference between a battery at 80% after eight years and one that dips below 70% after just six.

No hotspots in the thermal images. The heatsink construction of the chassis distributes heat evenly. This is consistent with what I expect from a well-designed passive cooling system.

Comparison with the competition

How does the Zendure stack up against the alternatives? I've lined up the most important plug-in home batteries.

Where Zendure wins: The 2400W AC power is substantial. Most competitors sit at 800W, which is insufficient during peak consumption (cooking, oven). The maximum storage capacity of 16.8 kWh exceeds everything in the plug-in market. And ZENKI 2.0 is currently the most advanced AI optimization in this segment.

Where Zendure loses: The price. Zendure positions itself as premium and that shows. The Anker Solix offers a more attractive entry model for those who need less capacity. And EcoFlow's Home Assistant integration is currently deeper than Zendure's -- a relevant point for the HA community.

PRO or AC+: which one suits you?

After extensive testing with both models my advice is more nuanced than "buy the most expensive one."

Choose the PRO if:

• You want to install new panels that won't go on an existing inverter
• You're planning a ground mount setup (garden, shed, carport)
• You want off-grid backup -- the PRO can charge directly from solar to battery without the grid
• You want maximum flexibility for future expansion of your system

Choose the AC+ if:

• You already have a rooftop installation with its own inverter
• You're renting and can't install panels, but still want to charge cheaply from the grid via dynamic tariffs
• You want simplicity -- plug in, configure the app, done
• You don't want to pay for MC4 hardware you won't use

The price premium for the PRO is justified if you're actually going to use the solar input. If you're not, the AC+ is the smarter choice.

Long-term questions

After several weeks of testing I'm enthusiastic about the hardware and ZENKI's logic, but there are questions I honestly can't answer yet.

ZENKI accuracy across seasons. My test period fell in spring. How does the AI perform in winter, when solar yield is minimal and consumption patterns are completely different? The system learns from historical data, but it only has that data after a full year.

Battery degradation. The specifications promise 3000+ cycles above 80%. I can't verify that in a few weeks. I'll be monitoring this system long-term and will publish updates.

Firmware updates. During my test period I received two firmware updates. Both went smoothly, but the system is entirely dependent on cloud connectivity. What happens if Zendure decides to discontinue the cloud service? This is a risk with all cloud-dependent home batteries, not just Zendure, but it's worth mentioning.

WiFi dependency. ZENKI needs an active internet connection for price data and weather forecasts. During WiFi outages the system falls back to local rules. During my test I simulated a four-hour WiFi outage -- the system continued functioning based on the last known data, but optimization was noticeably less sharp.

Honest verdict

What's excellent

The hardware is top-notch. Full metal enclosure, GaN/SiC components, ZenGuard safety, IP65 -- Zendure delivers a premium product that feels solid and looks professional. During testing I noticed the build quality is consistent: no rattles, no flexible panels, no cheap connectors.

The ZENKI 2.0 AI is the real competitive advantage. No competitor in the plug-in home battery market currently offers a comparable level of AI-driven optimization with support for 840+ energy suppliers. The decisions the system makes are logical, consistent, and pay for themselves.

The efficiency numbers are solid and consistent with the specifications. The FLIR data confirms the internal engineering is sound. This is a product where the spec sheet matches reality.

What could be better

Home Assistant integration is limited compared to some competitors. Mark from IoTDomotica has written extensively about this -- for those who want to run their energy management entirely through HA, the current integration depth is a consideration. Modbus/API capabilities exist but require more configuration effort.

The RJ45 confusion is an unnecessary stumbling block that Zendure could easily have prevented with a different connector type or better labeling.

The cloud dependency concerns me in the long term. A local API or offline mode with full functionality would make the product significantly stronger.

Conclusion

The Zendure SolarFlow 2400 PRO and AC+ are serious home batteries that distinguish themselves on two fronts: build quality and software intelligence. The hardware is solid, the efficiency is competitive, and ZENKI 2.0 does things the competition doesn't offer yet.

Is it perfect? No. The Home Assistant integration could go deeper, the communication around that RJ45 port is confusing, and the cloud dependency is a long-term risk. But as a total package -- hardware, software, expandability -- this is one of the strongest options on the Dutch market for plug-in home batteries.

The full visual review will appear soon on ThuisbatterijNederland -- there you'll see everything you can't capture in text: the installation, the app walkthrough, and the FLIR images in action.

Transparency

This review was created based on our own tests and research. Zendure provided both test products (SolarFlow 2400 PRO and SolarFlow 2400 AC+) free of charge for review. We received no financial compensation. No agreements were made about the content, tone, or conclusions of this review. Zendure has not seen or approved this article before publication.

Test methodology: The efficiency tests were conducted with a ZLCP168008 power analyzer under controlled conditions (21-23 degrees C, ventilated space). A minimum of five full charge-discharge cycles were measured per model. The FLIR measurements were taken with a FLIR IR camera module, also at room temperature, after 45 minutes of continuous load at 2000W. All manufacturer claims were verified with our own measurements where possible.

Potential bias: We received the products for free, which can create an inherent bias. We compensate for this by reporting all positive and negative findings, including comparison with competitors. Our efficiency measurements are independent and reproducible.

Sources

• Zendure SolarFlow 2400 PRO -- official specifications
• Zendure SolarFlow 2400 AC+ -- official specifications
• IoTDomotica.nl -- Zendure SolarFlow 2400 PRO review by Mark
• Rijksoverheid.nl -- net metering ends on January 1, 2027
• Own test data: ZLCP168008 power analyzer, 5+ full cycles per model
• Own test data: FLIR thermal camera, 45 minutes at 2000W load