CLNE — Q2 2026 Financial Report Analysis

Clean Energy Fuels' second-quarter net loss narrowed to $14.9 million ($0.07 a share) from $20.2 million ($0.09) a year earlier, on revenue up 3.7% to $106.4 million. But almost all of that improvement came from a smaller non-cash accounting charge tied to Amazon's share warrant, not from the fuel business earning more. Strip that charge out and operating profit fell, and the company's own adjusted EBITDA measure slipped 8.6% to $16.0 million even though it sold 7.2% more fuel.

At a glance

• 81.8 million gallons of fuel sold, up 7.2% (renewable natural gas, or RNG, up 2.9% to 63.2 million). Customers are buying more; the problem is what each gallon earns.
• Fuel sales fell to $61.1 million from $67.9 million despite the higher volume, which the company puts down to lower pricing and lower natural gas costs being passed through.
• Environmental credits were $14.2 million, 13.4% of revenue (11.6% a year ago). Those government credits grew while fuel sales shrank, so more of the business now rests on credit prices.

The quarter in numbers

Operating margin is the operating loss divided by revenue: the share of each sales dollar left (here, lost) after running the business, before interest and tax. A "gasoline gallon equivalent" is the amount of natural gas holding the same energy as a gallon of gasoline; the company reports all its volumes that way.

Why a business selling more fuel earned less from it

Clean Energy sells natural gas, increasingly the renewable kind captured from landfills and dairy farms, to truck, bus and refuse fleets through more than 570 US stations. Its revenue has three moving parts: the fuel itself, the environmental credits that fuel generates, and building stations for customers.

The fuel. Volume rose 7.2%, but the 10-Q says fuel sales fell $6.7 million "due to lower pricing and lower underlying natural gas commodity costs," partly offset by a smaller Amazon warrant charge and more gallons. Much of Clean Energy's pricing passes the cost of gas through to customers, so cheaper gas lowers revenue and cost together. Product cost of sales fell only $0.9 million, to $64.3 million, because station construction costs rose from $6.4 million to $14.1 million at the same time.

The credits. When RNG goes into a vehicle's fuel tank it earns government credits the company then sells:

• RIN credits (Renewable Identification Numbers) under the federal Renewable Fuel Standard, which requires refiners to blend renewable fuel or buy credits from those who do.
• LCFS credits under the Low Carbon Fuel Standard programs in California, Oregon, Washington and New Mexico, which reward fuels with a lower carbon footprint.

Their prices are set by trading markets and can move sharply in either direction, so for this company they can swing results more than fuel sales do. RIN revenue rose $1.0 million to $10.3 million and LCFS revenue $1.3 million to $3.9 million. The company credits higher prices and volumes, $1.6 million of credits from its own consolidated dairy RNG project, and a larger share of LCFS value. It also says RIN prices in the first half averaged about 3.2% above a year earlier. On RINs the company kept a lower share of the credit value, which ate into the gain.

Construction. Station construction revenue doubled to $16.0 million on more building activity. It is low-margin work: it cost $14.1 million to deliver, leaving about $1.9 million.

The Amazon warrant: why GAAP and "adjusted" figures disagree

Amazon holds a warrant (a right to buy Clean Energy shares at a set price) that vests as Amazon buys fuel from the company. Accounting rules treat the value of that warrant as a discount given to a customer, so it is subtracted from revenue as a "contra-revenue" charge. No cash changes hands.

That charge was $9.6 million this quarter against $17.4 million a year ago. Of this year's figure, $6.5 million reduced fuel sales and $3.1 million reduced service revenue (a facility use fee that is new in 2026). The $7.8 million smaller charge is more than the entire $5.4 million improvement in the net loss.

What the headline numbers hide

• The underlying business went backwards. If you add the warrant charge back, the operating result went from roughly +$8.2 million in Q2 2025 to +$4.5 million this quarter (our calculation from the income statement). The company's own adjusted EBITDA tells the same story: $16.0 million against $17.5 million. Adjusted EBITDA is earnings before interest, tax, depreciation and amortization, with the warrant charge, stock pay and some investment items also added back. Non-GAAP EPS slipped from $0.00 to a $0.01 loss.
• Fuel earns less per gallon. Adding the warrant charge back, fuel sales were about $85.3 million in Q2 2025 and $67.6 million this quarter, down 21% while gallons rose 7.2%. Volume-related revenue net of the fuel-related cost (excluding construction costs) fell from about $37.9 million to about $31.4 million on the same basis. These are our calculations from the filing's revenue and cost tables. Some of the revenue drop is lower gas costs passed through, but the margin left after costs also shrank.
• More reliance on credits. RIN and LCFS revenue was 13.4% of total revenue, and about 17% of volume-related revenue once the warrant charge is added back. A sustained fall in credit prices would come straight off the thin margin above. Prices rose this year, which helped.
• Lower interest cost on the income statement, higher cash interest. Interest expense fell to $5.5 million from $7.7 million after the company repaid $65 million of debt early in 2025. In cash terms, though, interest paid in the first half rose to $8.9 million from $2.7 million. A year ago it paid $10 million of interest "in kind" (adding it to the loan instead of paying cash); none was paid in kind this year. Interest income also halved to $1.6 million on lower rates and smaller cash balances.
• Cash flow weakened. Operating cash flow for the first half was $20.4 million against $59.3 million a year earlier. The company cites working-capital timing; deferred revenue alone was a $9.0 million outflow. After $10.5 million of equipment purchases, $1.2 million for dairy projects and $26.8 million invested in other entities ($24.0 million of it capital calls on the Maas dairy RNG projects), cash and short-term investments fell to $138.0 million from $156.1 million at year-end.
• The debt. Clean Energy owes $250 million on a Stonepeak term loan at a fixed 9.5% interest rate. The full balance is due in one payment in December 2029, and repaying it early carries a premium that guarantees lenders at least an 11.5% return. The company says it was in compliance with all loan covenants. Net of cash and investments, debt is about $112 million.
• No buyback help. Share count was essentially flat at 220.3 million, so the EPS change reflects earnings, not fewer shares.
• First-half comparisons are flattered by 2025 one-offs. H1 2025 included a $64.3 million goodwill write-off and $50.7 million of accelerated depreciation on station equipment removed from Pilot sites. That is why the six-month net loss shrank from $155.2 million to $27.3 million. The second quarter compared above is not affected by either.

Takeaway: The narrower GAAP loss is an accounting effect. The Amazon warrant charge fell by $7.8 million, more than the whole improvement, while adjusted EBITDA fell 8.6% even as fuel volumes rose 7.2%. Clean Energy is selling more gas but keeping less per gallon, and credit revenue is carrying a growing share of the load.

Outlook

Management left its 2026 guidance unchanged from May:

The guidance assumes no unrealized gains or losses on its truck-financing fuel contracts and excludes acquisitions and other unusual events. Hitting even the low end of the EBITDA range needs about $18.7 million a quarter in the second half, against $16.6 million in Q1 and $16.0 million in Q2. The larger GAAP loss implied for the second half is mostly the bigger warrant charge management expects. The company plans about $25 million of capital spending in 2026 and up to $42 million of equity into the Maas dairy RNG projects ($24 million already contributed). It plans no new spending on its own dairy projects.

Our read: the second half has to be noticeably stronger than the first for the guidance to hold, at a time when fuel margins are shrinking. Watch three things: how credit prices hold up; whether the six bp joint-venture dairy projects, now all producing, lift upstream RNG output (1.4 million gallons in Q2); and the Treasury's final rules on the Section 45Z clean fuel production tax credit, which the company is still evaluating. Leadership is also new. Clay Corbus became CEO in April, a new chief operating officer was named in June, and on October 6 longtime CFO Robert Vreeland left, replaced by corporate controller Jason Armstrong. Q3 results should arrive in early November if the company keeps its usual timing.

Source: Clean Energy Fuels Corp. Form 10-Q for the quarter ended June 30, 2026, filed August 6, 2026; 2026 guidance and adjusted figures from the company's August 6, 2026 earnings release (Form 8-K, Exhibit 99.1).