Fluence Loses Its Charge

The next time you find yourself on EDGAR, consider using my SECSift reader instead!

It turns SEC filings into clean, readable documents. You can search every word, export anything, merge tables, highlight what matters and gray out the rest, and compare against previous filings.

Not familiar with the reader? Here’s a 3-minute interactive demo (viewable on your computer) and another video I made walking through the entire tool.

🌾 Welcome to StableBread’s Newsletter!

I’m down ~47% on Fluence Energy (Nasdaq: FLNC), a top-three global battery storage integrator I first wrote about on July 19, 2026. I then provided an updated analysis on August 10, following earnings.

The position is <2% of my portfolio so the overall downside impact remains minimal.

Regardless, anyone who followed me into this investment lost nearly half within ~60 days.

Market volatility and drawdowns are to be expected if you’re chasing asymmetric setups, especially for small-cap stocks. But that’s not an excuse to hold or buy more when a thesis is clearly broken.

So in this post I hope to accomplish three things:

I believe this will be a lot more of a helpful exercise than an earnings update or proving a thesis worked (so far), like BRUN (+260%), FTEK (+27%), ELMT (+20%), TOST (+24%) and several others I’ve written up in 2026.

Fluence Energy builds the shipping-container-sized battery systems that utilities and data centers plug into the grid.

It doesn’t make the cells, it buys them and wraps its own engineering, controls software, and 20-year service contracts around the hardware, which is where its profit comes from.

Here’s how the July 19 write-up ended:

The thesis was that at $14.50/share you were buying a $2.67B company with a record $5.6B backlog, order intake running at double the prior year’s pace, and roughly breakeven net cash, for ~0.8x FY2026 revenue guidance.

Fluence had two hyperscaler MSAs (framework agreements that pre-qualify its product and pricing so orders can follow quickly), a spot in the Siemens/NVIDIA AI data center reference design, and a 12 GWh data center pipeline the price wasn’t paying for.

Plus, its U.S.-made supply chain was landing just as the prohibited-foreign-entity rules in the One Big Beautiful Bill Act (OBBBA) started pushing Chinese competitors out of the U.S. market.

All the business had to do was hold its usual 11-13% gross margin while the order book compounded.

My sum-of-the-parts (SOTP) valuation applied 0.5-1.0x to the $3.2B of product revenue inside the guide, my range for low-margin electrical hardware, which comes to $1.6-3.2B. The service and software contracts came in at $720M-1.1B on 4-6x their ~$180M of ARR.

Added together, that’s $2.3-4.3B of enterprise value, or $12.70-23.40/share, so from $14.50 you were risking ~$2 to make up to ~$9 (+61%) before the data center option.

Today, the stock's fallen 42.4% below my SOTP range, from $14.50 in July to $7.32 at the September 18 close (down 49.5%).

Demand wasn’t the problem. Fluence booked $1.44B of orders in Q3, which took backlog to a record $6.4B by June 30, and secured ~$850M of data center business through July.

What Fluence couldn’t do was produce at the rate and cost it planned, and that showed up as two guidance cuts six weeks apart.

On August 5, Fluence reported fiscal Q3 (its FY ends September 30). Adjusted gross margin came in at 5.9% against the 11-13% target, and management cut FY2026 revenue guidance to $2.9-3.1B from $3.2-3.6B.

The cut moved ~$400M of deliveries into FY2027, which Fluence blamed on two new contract manufacturing plants:

My August 10 earnings update said I needed Q4 gross margin back toward ~11% before adding, and that another miss would take the margin assumption out from under the whole valuation:

Even after the cut, the midpoint of the $2.9-3.1B guide still implied a record ~$1.41B of Q4 revenue, out of the same plants that had just missed.

Management claimed that close to half of the quarter’s deliveries were already built, with Houston producing on generators and set to connect to the grid “in the coming weeks.”

The September call later put a number on the Q4 production plan: 11 enclosures/day out of Houston through the August-September ramp—the rate ~$1.41B depended on.

CEO Julian Nebreda also said FY2026 would still land around the guided 12% adjusted gross margin excluding one-time costs, and that the products weren’t the problem.

But the deliveries that slipped led to total cash falling to $365M against $400M of convertible notes due 2030, so the ~$13M of net cash Fluence held in March became ~$35M of net debt.

Six weeks later, on September 16, Fluence cut FY2026 revenue to ~$2.4B, or 6% growth over FY2025 instead of the 50% guided in November 2025, and widened the adjusted EBITDA guide from a ~$10M loss to a ~$200M loss.

That’s the fourth guidance cut in 19 months (Feb 2025, May 2025, Aug 2026, and now Sept 2026), on top of the FY2025 miss.

Fluence’s September 16 presentation splits the ~$600M revenue cut three ways:

The adjusted EBITDA guide fell $190M, and the deck ties most of the drop to the missed milestones: ~$65M of gross margin on the revenue that slipped into FY2027 plus that same ~$65M of penalties.

The other $60M is $35M for the rollout of new products and $25M to get Houston’s output up to volume and quality standards.

Per the September 16 call, Houston produced under one unit/day in August because the automated welding line didn’t perform.

The contract manufacturer switched to manual welding and brought in three subcontractors, and output is now ~3 units/day (still well below the 11 units/day Houston was supposed to run).

Nebreda’s explanation was that Fluence underestimated how complex the ramp would be.

Revenue was $940.1M through H1, so the $3.4B guide needed ~$2.46B in H2 ($3.4B − $940.1M), more than Fluence delivered in all of FY2025 ($2.26B).

An H2-weighted year was normal for Fluence, which recognizes revenue as projects are delivered, so deliveries cluster in the summer quarters and FY2025 landed 73% of its revenue in the second half ($1.64B of $2.26B).

So expecting ~$2.46B in H2 wasn’t unreasonable.

But the entire H2 ran through two plants that hadn’t produced at volume, with no slack in the schedule if either slipped.

The ~$2.4B guide implies ~$810M of Q4 revenue ($2.4B − $1.59B through nine months), against the $1.41B the August guide needed and the $1.04B Fluence delivered in Q4 FY2025. So the quarter that was supposed to carry the year is now down ~22% y/y.

For FY2027, Fluence aims to deliver the ~$2.9B of backlog already scheduled, to “right-size” growth so it needs no additional capital, and to run neutral-to-positive operating cash flow.

Detailed FY2027 guidance and an action plan come with FY2026 results in November, and Fluence says it will start publishing U.S. production levels every quarter, against its own forecast (more on this later).

The other September news came from EVE Energy (Shenzhen: 300014), the Chinese battery maker, rather than from Fluence.

On September 18, EVE disclosed a five-year framework its Hubei EVE Power subsidiary signed with Fluence for 206 GWh of energy storage cells from 2027 through 2031.

The 206 GWh is split into 16 GWh of committed deliveries for 2027 and 190 GWh of reserved capacity for 2028-2031, with specifications, quantities, prices, and delivery schedules left to later purchase orders.

For scale, the 16 GWh committed for 2027 is 83% of the 19.2 GWh Fluence had deployed in its entire history as of May. Not to mention, Fluence locked that up days after cutting FY2026, which implies management is still planning for volume.

But this is Fluence buying cells, not customers buying systems, so none of the 206 GWh sits in the $6.4B backlog.

The July thesis also leaned on the OBBBA rule that denies the 48E investment tax credit, the 30% federal credit storage owners claim, to projects built with equipment from a prohibited foreign entity (namely China). Fluence’s U.S.-made systems let the project owner claim that credit, and rivals shipping Chinese equipment couldn’t.

A U.S. project built on EVE’s cells would fail the content limits that phase in through 2030 (75% non-PFE by then), and the cells still face the 28.4% tariff on Chinese batteries.

So the EVE cells are for projects outside the U.S., plus any U.S. buyer willing to give up the credit.

On the September 16 call, CFO Ahmed Pasha said data center developers that need speed to power may be willing to forgo the credits and take non-compliant equipment, which is the one U.S. channel where that could happen.

48E is a U.S. credit, so a project built outside the U.S. never claims it and never faces the content test. Therefore, a U.S. customer buying a Fluence system still gets the same credit it got in July, as long as Fluence doesn’t start putting EVE’s cells into U.S. projects.

What changes is which half of the business is growing. The U.S. half is the half the credit protects, and it runs through Houston (which is behind schedule). The international half runs on cells like EVE’s, and that’s the side management says is on track.

BNP Paribas Exane read the September plan to “right-size” growth as limiting new orders for U.S.-made systems to protect cash. I’d agree, since new U.S. work ties up inventory and supplier advances long before the cash comes back, and outside capital is what management (if you believe them) just said they won’t be needing.

Confidence in management is at an all-time low after its fourth guidance cut in 19 months. Wall Street has adjusted its price targets accordingly:

So what's the plan to start executing and win investors back?

To start, in the September 16 deck, Fluence committed to reporting U.S. production levels every quarter, actual against forecast.

It’s the first time the plan can be checked against output before the revenue shows up, which is a great transparency update.

Management’s also changed who runs operations:

As a reminder, AES co-founded Fluence with Siemens and holds 22.5% of the company through its Class B-1 shares, so the shareholder with the votes to force a change just put its own former COO over everything operational.

This doesn’t fix Houston by itself, but it signals that common shareholders aren’t the only ones who’ve run out of patience.

Management also committed on September 16 to fund FY2027 out of its own cash flow rather than guide to growth again, which is the call I’d want when the alternative is adding debt on top of the $400M convert or issuing equity at $7.32/share.

The cost is that the $6.4B backlog potentially converts more slowly.

Then there’s the disclosure. In August, Fluence booked a ~$15M upfront cost for “a planned agreement for long-term international battery supply” without naming the counterparty.

Every word of that description fits EVE. It was unsigned in August and signed six weeks later, it runs five years, and the supplier’s a Chinese cell maker.

Disclosure is also where the lawyers are looking. On September 17, the plaintiffs’ firm Block & Leviton announced an investigation of the manufacturing-ramp disclosures, with no complaint filed.

Plaintiffs’ firms issue these announcements to recruit clients, so one on its own doesn’t mean much.

But it’s worth mentioning since Fluence is already facing the SEC’s revenue-recognition inquiry open since 2024 (discussed in my July write-up) and a securities class action dismissed without prejudice in March, which left plaintiffs a window to refile.

The July valuation rested on 11-13% gross margins and $3.2B of FY2026 product revenue, and neither number exists anymore, so the $12.70-23.40/share range is gone.

Here’s what’s left to value at $7.32/share:

Here are three scenarios for FY2027 (based on 184.6M shares):

At $7.32/share, the stock sits ~5% under the bottom of the base case. But if you bought around ~$14.50/share, the base case still leaves you down 15-47%, and even the bull case at $15.52 only offers a 7% upside.

None of the three events I flagged in July has gone my way:

On valuation, the base case is $7.67-12.38/share at 0.5-0.8x FY2027’s ~$2.9B of backlog, the bear is $3.93 if FY2027 repeats FY2026’s margin, and the bull is $15.52 if Houston reaches rate and the margin comes back.

Again, from $14.50/share that’s only a 7% upside on the most optimistic case, which relies on an unproven management team getting everything on track.

Back in August, CFO Ahmed Pasha said that FY2027 growth may take another $300-500M of working capital, against ~$515M of cash and drawable credit ($365M + $150M), with FY2026 already guided to a ~$200M adjusted EBITDA loss.

So the low end of that range is fundable and the high end isn’t, and raising $300-500M of equity at $7.32/share means 41-68M new shares, or 22-37% dilution. That means dilution or more debt is still on the table, regardless of what management says.

Overall, demand delivered, but the margin didn’t hold, the U.S. production it depended on didn’t ramp, and instead of decoupling from China, Fluence is doing the opposite.

Here’s what I’m watching:

When an incompetent management team can’t execute in a market where demand for BESS storage outruns the supply of cells, that alone is a reason to sell.

But since this is only <2% of my portfolio, and because the next quarter brings Q4 results, the FY2027 guide, and Houston’s first production numbers, I’m holding until then.

Disclaimer: This write-up is for informational and educational purposes only and is not investment advice. I may hold positions in the securities discussed. Do your own research before buying or selling any security.