AVAV: Funded Backlog Still Needs a Funded Factory

The 60-second read

Five things to know

1. The profitable engine grew. AxS added $60.6M of sales and $9.5M of adjusted EBITDA. Its margin slipped slightly.
2. The second segment reversed the gain. SCDE adjusted EBITDA fell $12.7M, leaving group EBITDA lower.
3. Adjusted EPS is a different story. $0.59 versus $0.32 reflects financing changes and lower stock compensation; amortization is excluded in both periods.
4. A funded order does not prepay the factory. Inventory and unbilled work consumed cash while equipment spending accelerated.
5. Our valuation gives growth credit. The DCF assumes sustained expansion and much better margins, yet remains below the price.

Three scenarios, one probability-weighted number

Twelve-month valuation cases using FY2027 EBITDA bases; assumptions and probabilities are ours.

Signposts: what would change our mind

These are our monitoring thresholds, not management promises. The next packet will grade each. Exact future earnings dates are unannounced; December 2026 is a calendar expectation.

The tape

The central question is what an investor is paying for future owner cash, not whether autonomous defense systems have strategic importance. AeroVironment is a larger and more complicated company after BlueHalo. Its products can win significant orders while the economics of factory expansion, contract milestones and integration remain unresolved. That combination supports a useful operating business and still allows the stock to be expensive. The SELL call is a valuation judgment with moderate conviction, not a prediction of immediate price direction or a claim that customer demand is weak.

Our fixed valuation price is $140.80, the September 9 regular-session close. Results were released after that close. FMP recorded a 5.36% decline during the regular session; that move preceded the report and must not be presented as the earnings reaction. A later FMP snapshot at 6:13:55 p.m. Eastern showed a $144.02 bid and $144.29 ask. Those are indicative quotes, not a verified last trade. The next full-session earnings reaction was unavailable at publication preparation. Freezing the price makes the model reproducible even while the after-hours market moves.

The chart uses closing prices, not intraday highs or a synthetic total-return series. The quote’s 52-week range is $135.20–417.86. A large drawdown can make a stock look inexpensive compared with its old high; it does not establish fair value. The relevant comparison is the capitalized value of the cash that can reach shareholders after the investment program. Our chosen central value is below the price even though the model allows substantial growth and a much stronger mature margin.

The print

Revenue of $480.490 million exceeded the year-ago quarter by $25.814 million. Product revenue increased and service revenue increased at the consolidated level, but the segment mix tells a more discriminating story than those two broad categories. The release’s headline record is accurate; the analytical question is why a larger sales base generated less adjusted EBITDA. It was not a uniform margin collapse across the portfolio. One segment produced healthy growth, while the other lost revenue and profit. The appropriate operating comparison begins there.

FMP recorded revenue consensus of $451.948 million, making the sales beat 6.3% on our calculation. Its adjusted EPS estimate was $0.2226, against the company’s $0.59 result. This is a material positive surprise against that snapshot, but consensus is a moving dataset and the EPS definition differs from GAAP. We do not multiply the percentage EPS beat into a valuation target. The useful follow-up is whether analysts raise sustainable operating and cash estimates after reviewing the detail.

Management maintained revenue guidance of $2.125–2.225 billion and adjusted EBITDA of $305–325 million. Midpoints are $2.175 billion and $315 million. Maintaining guidance after a beat leaves room for conservatism, but does not itself prove another beat. The remainder needs $1,694.510 million of revenue and $261.611 million of EBITDA to reach those midpoints: a 15.44% remaining-year margin, compared with 11.11% in Q1. The presentation places approximately two-thirds of annual EBITDA in the second half, so the burden of proof is back-loaded.

Segments

Autonomous Systems, or AxS, supplied $345.969 million of revenue, up from $285.324 million. Its adjusted EBITDA increased to $62.285 million from $52.760 million. That is the healthy part of the quarter: the larger sales base delivered an additional $9.525 million of operating profit on the segment’s preferred measure. However, margin declined from 18.49% to 18.00%. Strong revenue growth should not be described as margin expansion. The business needs delivery discipline while investing in capacity and absorbing the broader portfolio.

Space, Cyber and Directed Energy, or SCDE, supplied $134.521 million of revenue, down from $169.352 million. Adjusted EBITDA moved from a $3.796 million profit to an $8.896 million loss. The $12.692 million deterioration exceeded AxS’s improvement, explaining the entire $3.167 million consolidated decline. SCDE’s margin was negative 6.61%. A recovery to simple break-even would help, but is not assured. The segment shows how an attractive collection of technologies can be difficult to convert into predictable financial performance.

Before BlueHalo, the company reported UxS, LMS and MacCready Works, with adjusted gross margin as segment profit. Fiscal 2026 introduced AxS and SCDE and adjusted EBITDA. Our tables retain each original reporting basis. We do not splice adjusted gross margin into a continuous EBITDA line or label acquisition-driven revenue organic. ESAero, acquired in March 2026, adds another portfolio change. Readers looking for organic growth should require a disclosed constant-perimeter bridge rather than infer one from consolidated totals.

Bookings were approximately $0.7 billion, book-to-bill was 1.4 and funded backlog was approximately $1.5 billion. Funded means appropriations support firm orders; it does not mean customers have advanced all cash needed to build the products. Contract milestones, inventory and supplier commitments can require working capital first. The presentation’s broader revenue visibility measure includes categories beyond funded backlog. We use the order book as evidence of demand while testing its value through deliveries, margins and cash collections.

Profit bridge

The GAAP operating loss narrowed from $69.272 million to $10.871 million. The $58.401 million improvement reconciles to $29.481 million of additional gross profit, $19.768 million lower selling, general and administrative expense and $9.152 million lower research and development expense. Acquisition accounting was a major contributor: intangible amortization and related purchase-accounting costs declined from approximately $79.7 million to $43.4 million. That matters for comparability, but is not evidence of an equivalent increase in cash earning power.

Net interest moved from $17.415 million of expense to $4.136 million of income after the acquisition financing was refinanced with equity and zero-coupon convertible notes. Other expense was $0.595 million. The resulting pretax loss was $7.330 million; a $0.397 million tax benefit and $1.867 million equity-method income brought net loss to $5.066 million. Adjusted EPS can improve even when EBITDA falls because financing and adjustments changed. Every below-the-line item matters to the comparison.

The adjusted EPS reconciliation adds $0.69 of purchase-accounting amortization and $0.04 of acquisition costs and removes $0.04 of investment activity from the $0.10 GAAP loss, producing $0.59. EBITDA additionally removes stock compensation and cloud implementation amortization. These are different measures with different purposes. We retain stock compensation as an owner cost in valuation and include broad capital spending. Removing acquired intangible amortization does not erase the capital originally committed to acquisitions.

Earnings quality

The earnings-quality scorecard separates accounting noise from recurring economic claims on cash. Low stock compensation this quarter is welcome, but a single quarter should not replace the $40 million full-year guidance assumption. Likewise, noncash acquisition amortization does not directly consume current operating cash, but repeated acquisitions require capital and can indicate that reported earnings omit part of the cost of maintaining growth. We avoid both extremes: treating every adjustment as fraudulent or treating every adjustment as economically free.

Operating cash flow improved sharply to $13.496 million from a $123.726 million outflow. Collections of billed receivables supplied $128.346 million, while unbilled receivables used $68.041 million and inventories used $100.310 million. This is a mixed conversion result. Total billed plus unbilled receivables declined from the fiscal year end, even though unbilled work rose. Describing all receivables as increasing would be wrong. The question is whether new work progresses through billing and collection quickly enough to finance delivery.

Our receivable and inventory days are 91-day endpoint-balance proxies, not management disclosures or exact collection periods. Fiscal quarters vary in length; average balances would better match the quarter’s revenue and cost flows, but endpoint measures are consistently available. Changes in acquisition perimeter also disrupt comparability. Beneish and Altman scores are not published: the acquisition reset, goodwill correction and changing segment perimeter undermine a clean comparable input series and interpretation. This is an applicability limitation, not a passing risk signal. Use the chart to identify a question, not to accuse the company of premature revenue recognition. The annual report’s accounting control weaknesses are a separate, documented concern requiring remediation evidence.

Balance sheet and cash

The balance sheet provides liquidity but not a free cash windfall. Cash was $278.390 million, short-term investments $301.837 million and long-term investments $94.777 million. Their total is $675.004 million. The convertible notes have $747.5 million principal, producing $72.496 million of face-value net debt. Carrying debt is lower at $730.057 million because of accounting discounts and issuance costs. We use principal for equity valuation: the contractual repayment obligation does not disappear because book value is lower.

The notes carry no regular coupon and mature July 15, 2030. Their initial conversion price is approximately $322.40. Settlement must be in cash at least up to principal; the excess can involve cash or shares. At the current price we model debt rather than mechanically add all conversion shares and simultaneously subtract the debt, which would count the same claim twice. Future scenario shares incorporate separate dilution assumptions. Operating leases remain an operating cost in the cash model and are not inconsistently added to debt while leaving rental expense in profit.

The quarter generated $13.496 million of operating cash and spent $44.033 million on property and equipment, producing negative $30.537 million of conventional FCF. Capitalized software consumed another $5.417 million, taking the broader figure to negative $35.954 million. These figures are computed definitions, not a new company guidance promise. The investor presentation’s full-year capital plan is broader still: 12–14% of revenue, including cloud implementation and software capitalization. At the revenue midpoint, that is $261–304.5 million. Using a small maintenance-only capital budget would overstate near-term cash.

Goodwill and intangibles represent a large portion of reported assets. They are not liquid reserves available to pay suppliers. The June 22 Q3 amendment increased the goodwill impairment from the original report, and the subsequent annual filing is the corrected historical authority. We retain the amendment as an identifiable control event, not a fictitious cluster of repeated restatements. The proper response is a higher required return and scrutiny of remediation, acquisition discipline and segment recovery. It is not to subtract every intangible again from a cash-flow valuation that already values operating cash directly.

Valuation

Three methods frame value rather than manufacture a false consensus. The cash-flow route is $95.32 per share, the normalized adjusted-EPS route is $95.40 and the enterprise-value route is $122.53. Our $105 central judgment sits within that spread. The $85–125 working range is an editorial range for operating and valuation uncertainty, not a statistical confidence interval or the minimum and maximum of the DCF sensitivity. The wider downside sensitivity is visible below. At $140.80 there is no margin of safety against our central case.

The EPS method applies 30 times the $3.18 guidance midpoint. That premium multiple recognizes strategically relevant products and sustained growth opportunities, while acknowledging that adjusted earnings omit acquisition amortization. It is not supported by a claimed peer median: 30 times is our judgment. The enterprise-value method applies 20 times $315 million of adjusted EBITDA, then subtracts $72.496 million of principal net debt and divides by 50.822963 million shares. EBITDA can flatter a capital-intensive expansion, so this method receives less authority when cash conversion remains weak.

The DCF starts from the August 1 financial balance sheet. It values the remaining fiscal-2027 cash-flow stub and later fiscal years, rather than deducting the already-spent first-quarter investment twice. Full-year FY27 modeled cash flow is reduced by an analytical proxy for Q1 owner cash: reported operating cash less PPE and software, less stock compensation, less net interest income after a normalized 24% tax. Cloud implementation cash already classified in operating cash is not subtracted again. This proxy is not reported FCFF; it approximates tax, cash-versus-accrual and financing timing. Unreported flows between August 1 and the September 9 stock price are not known.

Fiscal-2027 modeled revenue is $2.175 billion and adjusted EBITDA $315 million. We deduct the $40 million stock-compensation guide and $10 million acquisition-expense guide as economic costs. Normal depreciation is 3.4% of sales and tax is 24% of economic operating income. Broad capital spending is 13% of sales. Additional working capital consumes 20% of incremental revenue, an explicit assumption rather than a release forecast. Later revenue growth is 20%, 18%, 15% and 12%; adjusted EBITDA margin rises to 18%, 20%, 21% and 22%. No recurring acquisition expense after FY27 is assumed, which is favorable.

Capital intensity falls to 9%, 6.5%, 5% and 4.5%. Stock compensation stays at the guidance-implied share of sales. Those assumptions require successful capacity utilization and SCDE recovery. The terminal calculation uses 3.5% perpetual growth, 22% adjusted EBITDA margin, continuing stock compensation, 3.4% depreciation, 4.5% investment and working capital on incremental terminal sales. The discount rate is 10%. These are generous growth assumptions, not a liquidation case. A higher required return is reasonable given integration risk, contract dependence and control weaknesses.

Terminal value contributes 89.5% of enterprise value. That makes the output sensitive to assumptions far beyond current guidance. The grid changes the discount rate and perpetual growth while leaving operating assumptions fixed; it is not a probability distribution. Our reverse DCF keeps the same margins and investment rules and solves for the growth path needed to justify $140.80. The required multiplier on the base growth path is 1.79 times. That is one possible implied case, not a unique forecast hidden in the price. Better margins, cheaper capital or lower reinvestment could substitute for growth, but each needs evidence.

High uncertainty means we require room for error before a BUY. A lower quote alone would not repair deteriorating business evidence, and a higher quote would not prove our operating analysis wrong. Our decisive evidence is conversion: profitable delivery, stabilizing inventories, collections of unbilled work and an investment program that eventually produces distributable cash. The scenario range deliberately allows both a strong recovery and a significant funding disappointment. Scenario probabilities are subjective and sum to 100%; the weighted outcome is published even though it differs from the central call.

Versus Wall Street

The Street is more optimistic than our valuation. FMP’s ratings snapshot contains 16 buys and 12 holds, with no sells. This is a provider aggregation, not our own vote or evidence that all analysts share an identical financial model. The latest target item in the retrieved news series is from July 16, before this earnings release. Older price targets can be informative about the debate, but they must not be labeled fresh reactions to September results. We will revisit revisions after analysts incorporate the new segment and investment detail.

Consensus revenue, earnings and EBITDA also require definition discipline. FMP’s annual EBITDA series is not established as identical to the company’s adjusted EBITDA, which removes stock compensation, cloud amortization and other items. A numerical gap between those two datasets is not necessarily an earnings miss. Our table reports revenue and EPS with analyst counts and keeps the provider EBITDA field labeled separately. Long-range estimates with fewer contributors deserve less weight than a well-supported near-term consensus. They remain estimates, not signed contracts.

Management scorecard

Management’s stated priority is capacity and execution. Wahid Nawabi described expanding manufacturing and strengthening the supply chain, consistent with the large capital plan. That is a concrete strategic direction, but investors still need evidence that investments produce profitable throughput rather than only higher installed capacity. The first-quarter AxS revenue result supports demand and delivery capability. SCDE’s losses, the back-loaded profit outlook and the previous accounting correction require the management scorecard to remain mixed.

Capital allocation has changed the shareholder proposition. BlueHalo substantially increased the operating perimeter, debt and share count; ESAero adds integration work. The annual filing identifies Arlington entities with approximately 26.3% of outstanding common shares. The company is a US common-stock issuer, not an ADS or China VIE structure. There is no reason to import an unrelated foreign-listing discount. There is reason to consider acquisition discipline, shareholder influence and whether future strategic opportunities are financed with cash that could otherwise reach existing owners.

1. What milestones turn SCDE’s negative EBITDA into sustained positive cash contribution?
2. How much of the capital plan is temporary expansion versus recurring maintenance and software?
3. What share of unbilled receivables converts to billed receivables and cash over the next two quarters?
4. What utilization and delivery metrics should investors use to judge the new capacity?
5. What completed control-remediation evidence would demonstrate the goodwill accounting process is reliable?

Ranked risks

The strongest bull case is that demand arrives ahead of capacity, which makes today’s cash outflow an investment in a larger earnings base. AxS growth and funded orders support that possibility. If SCDE returns to profit while factories reach efficient utilization, operating leverage can be substantial. The bear case is that capacity, staffing and integration absorb cash for longer than expected while government awards convert unevenly. The same operating leverage then works in reverse. Our model credits a substantial recovery but refuses to value all future spending as automatically productive.

Government concentration is central: approximately 85% of fiscal-2026 revenue came from the US government directly or indirectly, including foreign military sales, and approximately 63% came from the Department of Defense. Funding authorization helps but does not remove procurement timing, program changes, acceptance conditions or execution risk. Fixed-price and development work can also expose costs before billing catches up. These are operating sensitivities rather than predictions about a particular political outcome. Risk ranking reflects our judgment of financial consequence and plausible frequency.

Catalysts

A catalyst is useful only if it changes the evidence. The next earnings release can test SCDE break-even, the cash conversion of accumulated work and whether the annual profit guide still requires a large second-half acceleration. Contract announcements are helpful when they specify funding, delivery periods and expected scope. A headline contract ceiling with optional future awards should not be treated as cash revenue already secured. The annual capital plan is another checkpoint: a revision needs to be considered alongside the sales and utilization it is supposed to enable.

Appendix and sources

Source hierarchy is deliberate. Current quarter figures come from the September 9 Form 8-K earnings release and presentation. The nine-quarter history uses each release, with Q3 fiscal-2026 GAAP figures overridden by the June 22 amended Form 10-Q. Three-year cash flow comes from the fiscal-2026 comparative cash-flow statement, and structural information comes from annual filings. All raw statement and reconciliation rows are retained in the dataset. Original March reconciliation rows are historical evidence only; their old net loss and goodwill impairment are superseded, not mixed into the corrected series.

There is a small unresolved comparative distinction: the current guidance table presents fiscal-2026 adjusted EBITDA as $287 million, while the June annual release reports approximately $286.1 million and the four exact segment-quarter totals sum to $286.056 million. We do not pretend that difference is ordinary rounding or invent a reconciliation. The current table implies about 10% growth at guidance midpoint. Historical charts retain the source quarter totals. This discrepancy does not alter the current year-over-year decline, which uses exact $53.389 million versus $56.556 million.

FCF means operating cash less cash purchases of property and equipment. Broader FCF also deducts capitalized software. Fiscal-2025’s original annual release bundled those two investment categories; the later comparative separates $19.547 million of PPE and $3.269 million of software, and later nine-month comparatives identify no software cash spending through Q3. Quarterly cash-flow amounts are cumulative differences and reconcile to fiscal-year totals. EBITDA means earnings before interest, tax, depreciation and amortization; the adjusted measure makes further exclusions. These definitions must travel with every number.

No prior packet with these dated signposts was available for grading in this production record. This packet establishes the next-quarter baseline; it does not claim a backtested prediction record. No source has been used to manufacture an earnings call quotation or undisclosed organic growth rate. Values are rounded for reading, while calculations retain source precision. A reader should distinguish reported fact, computed ratio and our forecast assumption. Those three categories are the foundation of a repeatable investment discussion rather than a confidence score masquerading as certainty.