FuelCell Energy has won back a bullish Wall Street analyst as its push into data-center power begins producing real customer commitments, but the company’s latest financial results show how much still has to happen before that opportunity turns into profitable growth.
Craig-Hallum initiated coverage of FuelCell with a Buy rating and $22 price target on September 9, citing improving prospects for its carbonate fuel-cell systems as data centers search for reliable power that can be deployed without waiting years for new grid connections.
The call marks a significant change in tone from the firm, which had previously stopped covering FuelCell when its product strategy and commercial backlog were considerably less developed.
The stock has also changed dramatically.
FuelCell closed Friday at $18.11, up 2.3% for the session and nearly 148% this year. It remains more than 50% below the $37.88 high reached in June.
At Friday’s price, Craig-Hallum’s $22 target represents roughly 21% upside—not the much larger gain implied by the $15.50 share price used in the source article.
The investment case now depends on whether FuelCell can turn a surge in AI-related power inquiries into profitable equipment sales.
So far, the demand signals are much stronger than the earnings.
Revenue Fell Even as the Data Center Pipeline Expanded
FuelCell reported fiscal third-quarter revenue of $33 million, down 29% from $46.7 million a year earlier and below Wall Street expectations of roughly $39 million.
Its adjusted loss of $0.64 a share was also wider than analysts expected.
Gross loss widened sharply to $24.5 million from $5.1 million, while adjusted EBITDA deteriorated to negative $36.7 million from negative $16.4 million.
The company did report a much smaller GAAP net loss of $45.3 million compared with $91.9 million a year earlier, but that comparison benefited from impairment and restructuring expenses that hurt the prior-year period.
The operating picture was therefore not nearly as improved as the headline reduction in net loss suggests.
FuelCell’s third-quarter revenue was hurt by fewer module deliveries in South Korea and lower electricity generation from company-owned projects.
The data-center opportunity is largely still ahead of those reported numbers.
The First 30 MW Is Real. The Other 350 MW Is Optional.
FuelCell’s biggest commercial breakthrough came through Fit Energy, which agreed to buy fuel-cell systems intended to power data centers.
The agreement covers as much as 380 megawatts across four phases.
Only the first 30 MW is currently committed.
Fit Energy made an initial deposit, and FuelCell expects deliveries to begin during its fiscal fourth quarter.
The remaining 350 MW is divided into additional 100 MW, 125 MW and 125 MW phases that Fit Energy may elect to pursue.
Until those elections occur, Fit has no payment obligation for those systems.
That distinction matters when reading FuelCell’s backlog.
Committed backlog stood at approximately $1.3 billion at July 31.
FuelCell separately reported $2.35 billion of what it calls Awarded Capacity Backlog, almost entirely representing the estimated product and service value of the optional 350 MW Fit Energy expansion.
Combined, the two categories total roughly $3.65 billion.
But FuelCell explicitly warns that Awarded Capacity Backlog is not firm order backlog, contracted revenue or a guarantee of future sales.
Sites still need to be identified and additional financing, permitting, customer development and project agreements may be required before those optional phases become binding.
The company’s approximately 10-gigawatt sales pipeline deserves the same qualification.
That figure represents commercial discussions ranging from early solution conversations to contract negotiations. It is not 10 GW of signed business.
A Second Data Center Customer Is More Encouraging
FuelCell did add another important data point after the quarter ended.
The company signed a capacity reservation agreement with an unnamed major data-center operator for a planned 75 MW project in Texas.
The customer made an upfront reservation payment in exchange for priority access to FuelCell’s manufacturing capacity.
The planned system would use six 12.5 MW fuel-cell blocks.
That is stronger than an informal sales lead, but it is still short of a final project contract.
FuelCell and the customer are continuing to negotiate definitive agreements while the company begins reserving capacity and ordering long-lead components.
Financial terms were not disclosed.
Together, the Fit Energy contract and Texas reservation provide more evidence that FuelCell is becoming relevant to an industry facing a severe electricity bottleneck.
AI data centers can require hundreds of megawatts or even gigawatts of power, while new transmission lines and utility interconnections can take years.
FuelCell’s systems can generate electricity directly at or near a data-center site.
Unlike intermittent renewable sources, the units can operate continuously, making them potentially useful for workloads that require power around the clock.
The technology is not the unresolved issue.
The cost of producing it is.
FuelCell Is Currently Losing Money on the Initial Data Center Order
FuelCell disclosed something in its quarterly filing that is more important than the size of the prospective pipeline.
Its current manufacturing costs exceed the contractual pricing established for the initial Fit Energy order.
FuelCell operated at an annualized production rate of only about 37 MW during the third quarter.
At that low utilization level, fixed factory expenses are spread across relatively few systems, pushing unit costs above what the company can charge on large commercial orders.
The Fit Energy agreement contributed to $17 million of charges during the quarter related to inventory and purchase commitments associated with the first 30 MW phase.
Management says those charges should be limited to identified inventory and commitments for the initial phase and do not represent the economics expected across the entire agreement.
The solution is scale.
FuelCell expects its annualized production rate to reach 100 MW in October.
It is simultaneously expanding its Torrington, Connecticut, factory toward 500 MW of annual production capacity, with completion targeted for June 2028.
The company believes higher factory utilization will lower unit costs enough to make large data-center contracts economically attractive.
Management now targets positive adjusted EBITDA in the fourth quarter of fiscal 2027.
Even that target carries conditions.
FuelCell says it depends on optional backlog converting into firm contracts, customers maintaining their delivery schedules and the company successfully executing additional cost reductions.
The balance between orders and factory economics is therefore the most important part of the new FuelCell story.
A 380 MW agreement means considerably less if producing the equipment at competitive prices remains unprofitable.
FuelCell Has Cash, but Shareholders Paid for Much of It
FuelCell’s financial position has improved substantially.
Unrestricted cash reached $658 million at the end of July, up from $278 million at the beginning of the fiscal year. Including restricted cash, the total was $737 million.
That gives the company substantial resources to increase production and pursue large projects.
Much of the improvement came from issuing stock.
FuelCell raised approximately $454 million of net proceeds from public offerings and other common-stock sales during the first nine months of fiscal 2026.
Shares outstanding increased to nearly 80 million at July 31 from about 46 million at the end of October 2025.
The company sold approximately 21.3 million shares through its at-the-market program alone during that nine-month period at an average price of $10.
Operating activities consumed $73.4 million of cash over the same period.
FuelCell is therefore better capitalized, but existing shareholders absorbed significant dilution to create that stronger balance sheet.
Future capital requirements will matter if the company reaches the scale envisioned in its data-center pipeline.
The Analyst Debate Is Really About Conversion
Craig-Hallum is not alone in becoming more constructive.
UBS upgraded FuelCell to Buy in July with a $27 target. B. Riley moved to Buy with a $32 target, while Canaccord recently maintained a Buy rating with a $30 target.
Jefferies remains positive but lowered its target to $20 following the latest earnings report.
There is still significant disagreement.
Citi initiated coverage this month at Neutral with a $19 target, and Wells Fargo maintains an Underweight rating with an $8 target.
That unusually wide range reflects the same uncertainty visible in FuelCell’s own financial statements.
The company has moved beyond simply claiming that data centers represent a future opportunity.
It now has a funded 30 MW order, a 75 MW reservation agreement, a much larger optional Fit Energy expansion and a sales pipeline measured in gigawatts.
What it does not yet have is evidence that large-scale data-center sales produce attractive margins.
FuelCell’s next milestones will therefore matter more than another increase in its theoretical pipeline.
The initial Fit Energy systems are scheduled to begin shipping during the current quarter, while production is supposed to reach a 100 MW annualized rate in October.
Investors will need to see whether higher manufacturing volume begins narrowing gross losses, whether the unnamed Texas customer converts its reservation into a definitive contract, and whether Fit Energy commits to any portion of the remaining 350 MW.
If those things happen, Craig-Hallum’s renewed optimism will have considerably stronger financial support.
If the pipeline grows while gross margins remain deeply negative and optional orders remain optional, FuelCell will still have an AI power story—but not yet the profitable business the stock’s 2026 rally increasingly assumes.