Drones: when do $ONDS, $RCAT, $AVAV and $KTOS turn orders into cash?

01The missing link is the status of the order

A drone can be strategically useful, technologically impressive and commercially unprofitable at the same time. Those statements are compatible because they describe different tests. The first concerns a customer’s need, the second the product’s performance, and the third what the supplier retains after building, delivering and supporting it. For shareholders in Ondas, Red Cat, AeroVironment and Kratos, the useful comparison connects all three rather than treating every defense announcement as the same kind of revenue event.

The immediate news hook is Ondas’ September 23 acquisition of three defense technology businesses. Adding sensing, communications and navigation capabilities can expand what a group is able to offer. It does not establish how quickly those capabilities will be integrated, which customer will order the combined product or how much cash the acquired activities will generate. Those are separate questions addressed in the company analysis below. Ondas acquisition announcement, September 23, 2026.

The distinction also responds to a visible reader question. In a sampled September 17 Reddit discussion about Red Cat, investors debated opportunities that had not yet become orders or revenue, competitive pressure and the timing of execution. The comments identify a useful information gap. They do not establish the company’s contract status, validate claims about individual managers or provide a representative sentiment survey. This article uses the discussion to frame the question and primary documents to evaluate the businesses. The original discussion.

Four terms need to remain separate throughout: opportunity, obligation, revenue and cash. An opportunity can be economically important before it becomes an enforceable customer commitment. A commitment can exist before the supplier recognizes revenue. Recognized revenue can precede collection, while customer advances can bring collection forward. The chronology depends on the actual agreement and the company’s accounting policy. The purpose of the comparison is to locate each disclosed milestone on that chronology, not to impose one accounting timetable on every drone manufacturer.

There is also a product problem hidden inside the sector label. Small reconnaissance aircraft, counter-drone systems, larger unmanned aircraft, propulsion, space systems and software do not share one unit price or one procurement cycle. A group’s total revenue may include several of these activities. Calling the entire number “drone revenue” can create a false comparison even when the reported financial figure itself is accurate. The four company sections retain the segment boundaries and dates disclosed by each issuer.

02Five states that an announcement can describe

Selection is the first state to examine. A company can be selected for a competition, demonstration, qualification process, contract vehicle or production award. Each selection has a different consequence. The decisive noun often appears later in the announcement: selected for what, with what scope and under which instrument? A demonstration can reduce technical uncertainty without giving the manufacturer an order for serial production. Conversely, a production agreement should not be dismissed as a mere test because an earlier stage involved prototypes.

A ceiling is a second state. The Federal Acquisition Regulation describes an indefinite-quantity contract as a framework with stated limits and individual orders for specific requirements. The government commits to a stated minimum; the maximum is not automatically the amount it must purchase. This is why an IDIQ headline needs its minimum, maximum, duration and ordering structure. FAR 16.504.

In a multiple-award framework, several suppliers may compete for subsequent orders. Dividing the total ceiling by the number of selected vendors is not a substitute for the actual award terms. Nor should the whole framework’s maximum be assigned to every participant. A supplier’s opportunity may be substantial, but its eventual share depends on orders, performance and the applicable allocation process. FAR’s ordering provisions require individual orders to identify the relevant supplies or services and remain within the framework’s scope. FAR 16.505.

These FAR provisions illustrate a U.S. federal procurement route; they are not a universal template for every transaction in this article. Other Transaction Agreements, international defense purchases and commercial contracts have their own terms. A production OTA can represent a real production arrangement. An export order may involve an agency, an integrator or an end customer whose role must be identified. The instrument should be read before deciding whether a headline describes a framework, an order, a development activity or a delivery commitment.

Funding is the third state. An obligated amount identifies a spending commitment in federal reporting; an outlay records money paid. The Congressional Research Service explains that distinction and also warns that federal award data can be incomplete or delayed. A search result is therefore evidence to reconcile with the actual award, not an automatic statement of the supplier’s revenue or bank balance. CRS guide to federal award data.

Delivery and acceptance form the fourth state. A manufacturer can order components, assemble units and conduct tests before a customer accepts the relevant performance. Depending on the contract and accounting policy, revenue may be recognized over time or at a particular transfer point. Physical shipment is an important operating event, but it is not a universal accounting rule. A clear disclosure connects the production milestone with the revenue-recognition policy and any remaining acceptance conditions.

Collection is the fifth state. The supplier ultimately needs cash to pay employees, suppliers, interest and new equipment. Progress payments or advances can improve that timing, but leave work to perform. A delayed invoice can consume working capital even when accounting revenue is growing. The SEC’s financial-statement guide explains why the income statement and cash-flow statement need to be read together: profit and cash generation answer related but different questions. SEC financial-statement guide.

03Why a funded order still needs an economic examination

The price structure determines who absorbs a cost surprise. Under the FAR’s firm-fixed-price description, the contractor bears responsibility for costs and the resulting profit or loss rather than automatically passing overruns to the customer. Cost-reimbursement contracts instead provide for allowable costs under the agreement’s limits. These are different risk allocations, and neither label removes the need to inspect the actual terms. FAR fixed-price description, FAR cost-reimbursement description.

For a manufacturer ramping a new product, this can matter as much as the nominal order size. More labor than expected, component substitutions, rework or lower production yield can reduce the margin on a fixed-price sale. A larger facility can improve throughput while initially increasing rent, staffing and depreciation. The useful disclosure connects accepted units and revenue with manufacturing costs and the capacity actually used. A planned production rate is an engineering or management target until the operating record supports it.

Backlog needs a boundary too. One company may distinguish funded and unfunded work; another may include particular purchase commitments under its own definition. Acquisitions can add another business’s backlog without representing new demand won organically during the quarter. A pro forma figure can be useful for describing an enlarged group, provided it is not added again to the reported figure or presented as the historical balance at a different date. The company sections apply each issuer’s definition instead of assuming that every backlog dollar has identical quality.

The same caution applies to book-to-bill. The ratio compares new bookings with revenue for the stated period. A strong ratio can indicate replenishment of future work, but it does not reveal when deliveries will occur, how much funding is committed or what margin the work will earn. A single large award can move the ratio sharply. Investors need the timing, scope and concentration of bookings alongside the ratio, especially when the group has recently changed through acquisitions.

Finally, the shareholder owns a claim on the whole capital structure. Paying for acquisitions in stock can conserve cash while issuing additional ownership claims. Paying cash can reduce available resources even when the acquired business brings revenue. Earnouts, employee awards and future construction spending can sit outside a transaction’s headline purchase price. They should be identified according to their terms rather than combined into an invented maximum dilution figure. The relevant objective is to understand the obligations that accompany the expanded opportunity.

The following map keeps the comparison on a consistent question while preserving the companies’ different reporting dates. It is a guide to the detailed evidence in the company sections, not a ranking of the shares.

These are reporting periods, not claims that the June balances are current September cash. The later announcements are examined separately below.

04ONDS: September’s acquisitions add capabilities, costs and a changing share count

Ondas’ September 23 announcement concerns completed acquisitions of Insignito, Ottopia Defense and Caribou Labs. The disclosed aggregate consideration is $56 million, payable in cash or common stock, with up to another $32 million contingent on performance milestones through 2028. Separately, the company approved approximately 2.98 million restricted stock units and options over 80,000 shares for incoming employees. Those employment awards have vesting conditions; they should neither disappear from the dilution analysis nor be treated as shares all immediately issued. Ondas, September 23, 2026.

Ondas’ announcement is the sole primary source identified for these detailed purchase terms. The industrial logic is to fill different gaps in an autonomous system. Insignito contributes passive acoustic sensing; Ottopia Defense brings remote-operation and connectivity capabilities; Caribou adds navigation and communications for environments where conventional satellite positioning is unavailable. For investors, the attraction is the possibility of selling a more complete solution to customers already buying another part of the portfolio. The challenge is turning technical compatibility into a supported product that a customer actually orders.

That integration has an economic test. Additional sensing or connectivity can increase the value of a system, but also its installation, validation and support requirements. Acquiring technology can shorten development time without guaranteeing that the resulting product wins procurement. The useful evidence will be orders identifying the integrated offering, repeat deliveries and margin contribution after the engineers, software and customer support needed to maintain it.

September also brought the larger GATE Technologies and Bron transaction. Ondas announced $205 million of consideration, comprising $105 million in cash and $100 million in stock, plus adjustments and potential performance-based earnouts of up to $185 million. These amounts belong to a separate acquisition and cannot be confused with the three businesses announced on September 23. Ondas, September 14, 2026.

The pattern makes acquisition discipline central to the ONDS thesis. Announced revenue multiples based on a target’s expected future sales are forecasts, not purchase prices divided by audited historical sales. Earnouts can align sellers with performance, but successful achievement may also require additional cash or shares. A business can deliver its operating targets while the parent distributes part of that success to the sellers under the acquisition agreement. Common shareholders therefore need both the operating result and the final economic cost.

05ONDS: Mistral shows why a contract ceiling is different from an executable order book

Mistral gives Ondas access to an existing U.S. government contracting business. The acquisition closed in April 2026. Its largest headline number requires particular care: the government announced a $982 million Lethal Unmanned Systems contract in September 2025, explicitly stating that funding and work locations would be determined with each order. The official contract number is W91CRB-25-D-A009. That announcement establishes a contracting framework; it does not establish that $982 million was immediately funded for delivery. U.S. Department of War, September 29, 2025, Mistral acquisition closing, April 24, 2026.

A more specific commercial milestone followed. On August 5, Ondas announced an additional Army order exceeding $50 million and said cumulative awards under that program exceeded $240 million, including an initial $190.8 million award. The company said systems were being produced for deliveries beginning in the third quarter. The order amounts and delivery schedule are issuer disclosures; the government’s original ceiling notice does not independently confirm each subsequent funding obligation. Ondas Army order announcement, August 5, 2026.

This provides a concrete chain to monitor: the specific award, the production schedule, contractual performance, billing and collection. The same program can appear in several releases as it advances through those stages. Adding the initial award, a later cumulative total and the ceiling would count the opportunity repeatedly. Nor should a delivery timetable be promoted to completed customer acceptance simply because the quarter has started.

The wider backlog has another complication: acquisitions change its perimeter. Ondas reported approximately $613 million of backlog at June 30, and approximately $757 million pro forma, including DZYNE and Cyberhawk, which closed on July 2 and August 10. These are two views of an overlapping portfolio. They are not amounts to add together, and their difference is not evidence of organically generated new orders. Ondas second-quarter results, August 13, 2026.

The investor’s next question is the delivery profile and profitability within that backlog. A large integration program may convert over a different period from a shipment of equipment. Acquired orders may also require working capital or additional engineering before they generate cash. The most informative progress would be a reconciliation showing which programs converted, what new orders replaced them and how much cash the expanded operating base required in the process.

06ONDS: rapid sales growth has not yet produced operating profitability

Ondas reported second-quarter revenue of $83.8 million, against $6.3 million a year earlier. Management also presented 85% growth on a pro forma organic basis, assuming the businesses owned in the 2026 quarter had also been owned in the comparison period. That separate calculation matters: the much larger reported increase combines operating progress with the acquisition of businesses that were absent from the earlier consolidated group. Ondas results, August 13, 2026.

The expanded portfolio includes industrial inspection as well as defense products. Operating labels such as Ondas Autonomous Systems and Ondas Sentinel help describe the strategy, but they should not be used to invent standalone margins that have not been disclosed. The June filing reports one operating segment. Ondas Networks was deconsolidated in January, further complicating a simple comparison with the historical group. Ondas Form 10-Q, June 30, 2026.

Revenue timing also differs across the portfolio. In the quarter, $49.6 million was recognized at a point in time and $34.2 million over time. The latter is roughly 41% of quarterly sales. Some qualifying work can therefore produce recognized revenue before the final equipment delivery or an unconditional right to invoice. The filing distinguishes contract assets from receivables on that basis. June 2026 revenue and contract-asset disclosures.

For ONDS, “wait for deliveries” is consequently an incomplete accounting rule. Investors need to understand what performance has been completed, what estimate supports recognition, and when invoicing becomes permitted. Revenue accompanied by a contract asset can represent valid progress while still leaving a financing requirement until a contractual billing milestone is reached. That is a timing issue to investigate, rather than an automatic accusation that the revenue is artificial.

The earnings statement shows that growth and profitability remain separate. Second-quarter gross margin was 43.1%, but the operating loss was $162.9 million and adjusted EBITDA was negative $50.6 million. Acquisition accounting, stock compensation and other items explain part of the gap between those measures; the adjusted loss nevertheless shows that the growing business was not yet covering its adjusted operating cost base. Ondas earnings release and reconciliation.

Our analytical focus would therefore be conversion quality: whether existing products generate more gross profit as volumes rise, whether integration spending becomes more efficient, and whether recognized revenue approaches cash collection. A rising backlog supports visibility. It cannot, by itself, demonstrate that the cost of operating the larger portfolio is scaling at a sustainable rate.

07ONDS: the cash reserve is substantial, but June is not September

At June 30, Ondas held $657.9 million in cash and equivalents, $726.6 million in short-term investments and $8.5 million of restricted cash. The frequently cited approximately $1.39 billion total includes all three categories. The first-half cash-flow statement reported $137.4 million used in operations and $959.1 million of net proceeds from sales of common stock and warrants. The liquidity reserve therefore reflects access to capital markets, rather than cash generated by the operating business. Ondas June 2026 Form 10-Q.

The filing also records substantial subsequent acquisition spending. DZYNE and Cyberhawk together involved approximately $322.3 million of cash consideration after quarter-end. The DZYNE transaction included roughly 40 million shares at closing and approximately 45 million additional shares scheduled for issuance on January 4, 2027. These obligations are relevant to a shareholder evaluating the business after June, even though the acquired operations were not consolidated in the second-quarter revenue. Subsequent-events note, June 2026 Form 10-Q.

It would still be wrong to calculate “current cash” merely by subtracting those purchases from June liquidity. September acquisitions, ordinary operations, collections and other payments also matter. The proper conclusion is narrower: the historical reserve was large, and publicly disclosed transactions subsequently used part of it. An actual September balance requires a new company disclosure or later financial statement.

Equally, a fair-value warrant liability should not be read as a conventional bank loan repayable at its marked accounting value. Changes in that valuation can move reported earnings sharply without producing a corresponding operating cash movement. The same principle works in both directions: a favorable remeasurement can improve profit without improving the economics of building and supporting drones. Cash-flow analysis is necessary precisely because accounting earnings contain these other effects.

For ONDS, a stronger investment case would combine repeat orders, conversion into economically profitable work and a slowing need for additional capital per unit of growth. Further acquisitions could accelerate that process or make it harder to judge. The decisive comparison is not simply whether the consolidated company becomes larger, but whether the acquired and original operations together create more durable cash generation relative to the expanded share count and remaining acquisition obligations.

08RCAT: the Army relationship is real, while procurement remains competitive

Red Cat’s Army opportunity should be grounded in both customer disclosures and actual deliveries. The Army’s August 2025 update identified Teal Drones’ Black Widow as one of two vendors for the second Short Range Reconnaissance tranche. Red Cat subsequently reported that its July 2025 low-rate production contract had been expanded to approximately $35 million. That is a dated contract value, not the entire SRR budget or a statement of today’s undelivered backlog. U.S. Army, August 26, 2025, Red Cat third-quarter 2025 update.

More recent government evidence shows the competition continuing. On August 10, 2026, the Army described a further SRR evaluation and identified systems from Parrot, Skydio, Skycutter, Teal and Vantage Robotics for manufacturing-readiness assessment. Advancing to that stage confirms relevance to the procurement process. It does not award Red Cat all future platoon-level demand or establish the dollar value of a new production order. U.S. Army, August 10, 2026.

The commercial evidence is stronger when it reaches the accounts. Red Cat attributed second-quarter revenue growth primarily to expanding Army deliveries and the start of deliveries to the Japan Ground Self-Defense Force. In April, it had identified a Japanese order for 173 Black Widow systems, funded under Japan’s fiscal 2026 arrangements. “Systems” should retain that unit: it cannot automatically be translated into the same number of individual aircraft or a contract value that was not disclosed. Red Cat Japan announcement, April 30, 2026, June 2026 Form 10-Q.

Drone Dominance is a separate opportunity. The official page already publishes Gauntlet II’s Top 5 performers in each of two missions, Deep Strike and Close Quarters Battle. It lists Teal Drones among the Close Quarters Battle participants that did not place in the Top 5 for either mission, while retaining eligibility for future competitions. Advancing to Gauntlet II therefore should not be presented as winning one of those finalist positions. The page says finalists’ order quantities will be finalized upon contract award and cautions that awards are not guaranteed. These remaining award conditions must be distinguished from the competition results already published. Official Drone Dominance leaderboard, Gauntlet II tab checked September 25, 2026.

This is where the distinction between strategic relevance and an attributable revenue forecast becomes useful. Red Cat has evidence of real military customers and deliveries. Future competitions may extend that business. The financial model should add their revenue only when the award, scope and expected performance schedule provide a defensible basis, rather than applying the size of a government initiative to one selected participant.

09RCAT: Blue Ops and acquired technology widen the opportunity—and the execution task

The latest September update concerns Blue Ops, Red Cat’s maritime division. A September 24 announcement described progress at its 155,000-square-foot Valdosta facility and the ramp of Variant 7 uncrewed surface vessels toward full-rate production. It also referred to plans for $30 million of investment and more than 200 local jobs. The event demonstrated a developing manufacturing operation; the release did not establish an additional quantity of accepted customer vessels or a new revenue award. Red Cat, September 24, 2026.

Factory space and manufacturing equipment are enabling assets. Their return depends on the production schedule, the product mix and customers taking delivery. For Blue Ops, a test pool or a vessel shown to visitors demonstrates a different milestone from a completed contractual acceptance. Both can be useful information, provided they remain attached to the question they actually answer.

The group has also expanded through acquired technology. Apium entered consolidation on March 27, 2026, and Quaze on May 19. Their autonomy software and wireless-power capabilities broaden the possible combinations within Red Cat’s portfolio. They also mean that year-over-year financial comparisons increasingly describe a changing group, rather than an unchanged Black Widow manufacturing operation. Red Cat business-combination and consolidation disclosures.

A broader offering may help a customer operate and sustain several unmanned platforms with fewer disconnected suppliers. But ownership of the components is only the starting point. Integration requires engineering responsibility, compatible interfaces, support arrangements and evidence that the customer needs the combined product. The economics improve when the additional capability creates a billable requirement or reduces the cost of delivering one. A successful demonstration alone does not disclose that result.

Red Cat reports one operating segment. Investors should therefore resist inventing separate maritime, software or charging margins from the group’s overall figures. The absence of that breakdown is a limit on precision: it does not show that any particular activity is unsuccessful, but it prevents confident claims that one new division already offsets another’s development spending.

The most useful future disclosure would connect these investments to orders, delivery schedules and recurring support work. Blue Ops output, software deployment and wireless-power adoption need not develop at the same pace. Treating them as distinct commercial paths helps explain why group spending can rise before all parts of the expanded portfolio contribute meaningful revenue.

10RCAT: growth is strong, but the annual target requires a much larger second half

Second-quarter revenue was $20.2 million, up 527% from the unusually small prior-year quarter. Gross profit reached $3.3 million and gross margin was 16.1%. The release’s reference to a 39% year-over-year margin improvement describes a relative increase: the corresponding reported margin was 11.6% a year earlier, a difference of approximately 4.5 percentage points. It does not mean a 39-point improvement or a 39% gross margin. Red Cat results, August 6, 2026, June 2026 Form 10-Q.

The improvement indicates that more sales and manufacturing efficiency can produce a better gross result. It does not yet show that the group is profitable after research, selling and administrative costs. The quarter still produced a $35.3 million net loss. For an expanding supplier, the question is whether successive delivery batches absorb fixed costs and improve yield fast enough to cover the larger organization.

Management reaffirmed a $150–180 million full-year revenue target in August. With first-half revenue of $35.66 million, reaching that range would require $114.34–144.34 million in the second half. That is an average of $57.17–72.17 million per remaining quarter, compared with $20.19 million in the second quarter. This is our arithmetic on the disclosed target, not a separate forecast or evidence that those sales have already been booked. August outlook.

The gap makes execution timing unusually consequential. A production increase can support the target, but it must be matched by the contractually required transfer of products or services. A late shipment, an incomplete requirement or a different delivery mix can move revenue between periods. Conversely, inventory assembled ahead of orders can allow a faster ramp when customer schedules become firm. The evidence should be the resulting deliveries and recognized revenue, rather than an assumption that more manufacturing space ensures the target.

Red Cat’s accounting policy generally recognizes product and prototype revenue on shipment unless a purchase order or contract specifies otherwise; engineering, training and support have their own performance requirements. That makes the actual contractual terms more useful than applying one universal acceptance rule. Revenue-recognition policy, June 2026 filing. A sound quarterly assessment should connect the delivery mix to margin and the reported sales to subsequent collections.

11RCAT: the cash build came from equity while inventory absorbed operating funds

Red Cat ended June with $325.6 million of cash, against $167.9 million at year-end. That increase accompanied an operating cash outflow of $78.7 million during the first half and $12.6 million of equipment and property purchases. Financing brought in $249.4 million net. The company therefore strengthened its ability to fund growth without yet demonstrating an operating business that replenishes the cash it spends. Red Cat cash-flow statement, June 30, 2026.

The May share offering was the main source. Including the exercised additional allocation, gross proceeds were approximately $258.75 million, with $245.24 million remaining after disclosed offering costs. Shares outstanding increased from approximately 120.07 million at December 31 to 152.69 million at June 30, including financing, acquisitions and other issuances. Those dates matter: neither count is a live September total. Red Cat equity disclosures.

Inventory and prepaid inventory reached $84.8 million, up from $30.4 million at year-end. For the shareholder, that is capital committed before the related cash collection. It can be preparation for a larger delivery program, but its usefulness depends on having the right components and finished products for the orders that actually materialize. The increase should not automatically be labeled either evidence of guaranteed future sales or evidence of obsolete stock. Red Cat balance-sheet update, August 6, 2026.

The operational test is the movement from that inventory into accepted or otherwise contractually completed deliveries, then into receivables and collected cash. If the manufacturing ramp works, previously purchased material can support faster revenue conversion. If schedules slip, the same material keeps cash tied up while payroll and engineering expenditure continue.

The larger cash reserve gives Red Cat time to execute; the larger share count raises the standard for measuring the result per share. The most persuasive progress would combine higher delivery volumes, improving gross economics and less cash absorbed for each additional dollar of sales. Another contract headline can support that path. It cannot replace the evidence that the company has converted the work into cash at an economic cost.

12AeroVironment: the business after BlueHalo is broader than drones

AeroVironment’s latest quarter covers the three months ended August 1, 2026, its first quarter of fiscal 2027. Revenue was $480.490 million: $346.0 million from Autonomous Systems and $134.5 million from Space, Cyber and Directed Energy. The consolidated increase was approximately 6% against the comparable quarter. These figures describe a diversified defense technology company, not a single drone factory. September 9 results.

BlueHalo closed on May 1, 2025, adding activities across several domains. Crucially, that acquisition was already included in the prior-year quarter used in the latest comparison. The present 6% increase should therefore not be confused with the much larger acquisition-driven expansion recorded when BlueHalo first entered the accounts. Nor should it automatically be called organic growth: the perimeter also changed through other transactions, including ESAero in March 2026. BlueHalo completion announcement, September 10 quarterly filing.

This wider perimeter changes how a drone-related headline should be interpreted. Switchblade procurement, laser systems, satellite communications and engineering services do not share one production schedule or one margin profile. A strong order for one product can coexist with reduced work elsewhere. Consolidated growth provides the total outcome; the segment bridge explains how that outcome was assembled.

BlueHalo also changed the capital invested in the business. The fiscal 2026 annual report records 17,425,849 shares issued as acquisition consideration and approximately $3.485 billion of consideration net of acquired cash. That figure includes the settlement of acquired debt and transaction expenses; it is not simply cash paid to the seller. The same report records a $240.708 million goodwill impairment in the Space reporting unit after the SCAR program encountered a stop-work order and was subsequently terminated for convenience in March 2026. Fiscal 2026 Form 10-K, acquisition and impairment notes.

An impairment is not a second cash payment for the acquisition. It does, however, reduce the accounting value assigned to expected future economic benefits. That distinction matters: excluding a noncash charge can clarify current operating performance, but cannot make the underlying deterioration in a program’s prospects irrelevant. Conversely, an impairment in one reporting unit does not prove that every acquired product has lost value.

For the latest quarter, GAAP gross margin improved to approximately 26% from 21%, partly because acquisition accounting charges were lower. Adjusted EBITDA nevertheless declined to $53.4 million from $56.6 million. The two measures are answering different questions. A cleaner comparison follows product mix, operating expenses and cash generation alongside the accounting adjustments, rather than choosing whichever measure tells the most attractive story. September 9 financial release and reconciliation.

13AVAV’s September orders: production, delivery orders and different customers

The September 2 E-HEL announcement is a concrete change in commercial status. AeroVironment reported a $464.8 million award for the Army’s Enduring-High Energy Laser program, covering dozens of LOCUST X3 systems over several years and associated support and training. The Army independently described the award as a production Other Transaction Agreement. It should not be reduced to a demonstration merely because the instrument is an OTA. AeroVironment announcement, Army confirmation, September 2.

The reverse shortcut is equally unhelpful. An announced production agreement does not mean the entire contract value has already been delivered, recognized as revenue or collected. The public releases do not provide a complete schedule of funded installments and payments. The defensible description is the disclosed award value and production scope, with multiyear execution still ahead. It is not appropriate to invent a funded percentage or replace the company’s description with an unsupported IDIQ ceiling label.

On September 8, AeroVironment announced a separate international direct commercial sale order worth more than $50 million for an initial delivery of LOCUST systems and support. The buyer was not named. This provides evidence of another commercial channel, but does not disclose a countrywide procurement program, a recurring annual order or a guaranteed follow-on. First international LOCUST order, September 8.

Switchblade illustrates a different contract structure. On August 26, the company announced a $51 million delivery order under the existing five-year, $990 million Lethal Unmanned Systems IDIQ. It covers Switchblade 600 Block 2 and Block 1 systems, including a foreign military sale. The delivery order belongs inside the framework; adding $51 million to $990 million as if the two were independent new awards would double-count part of the opportunity. Switchblade delivery order, August 26.

These orders also sell different economic products. A loitering munition purchase, an installed counter-drone system and a support package have different combinations of manufacturing, acceptance and service obligations. Contract value alone does not disclose the margin of each. For a reusable system, later maintenance may add revenue, but the installed base must exist and the service arrangement must actually be purchased. An analyst cannot manufacture an annuity by attaching a generic maintenance percentage to the headline award.

Industrial follow-through is consequently the next test: component availability, manufacturing yield, customer acceptance and the timing of payment. More demand can require additional inventory before it produces cash. An award strengthens visibility only to the extent that the factory and the supplier network can execute it at the price and timetable agreed.

14AVAV’s backlog: the date and funding category matter

At August 1, AeroVironment reported $1.457822 billion of funded remaining performance obligations. It expected to recognize approximately 78% as revenue in fiscal 2027, with the rest in fiscal 2028 or later. This is a dated expectation for converting existing obligations, not a promise that every dollar will arrive in the bank during the same period. Form 10-Q, performance obligations.

Management separately discussed approximately $1.4 billion of unfunded backlog and roughly $2.8 billion of total backlog on the September 9 call. Those are rounded figures and should not be recombined to create a supposedly exact total. The call also explains that residual capacity under certain sole-source IDIQ contracts sits outside the reported unfunded backlog. A headline containing the word backlog therefore requires its own definition, even within the same company. Official September 9 call transcript.

There is a separate timing trap. August and September awards occurred after the August 1 balance-sheet date. Their existence can improve subsequent visibility, but it does not make the quarter-end number stale in the accounting sense: that number still measures exactly what existed at its stated date. Producing a new total by adding every later press release would overlook deliveries, funding changes, modifications and other movements that the company has not yet reconciled publicly.

Revenue itself is not restricted to the moment a completed aircraft leaves a loading dock. AeroVironment reports that 67% of the quarter’s revenue was recognized over time. Customer-funded development can therefore generate recognized sales while the underlying product is still being developed. That is legitimate contract revenue, but it should not be described as evidence of equivalent serial hardware shipments. Form 10-Q, revenue recognition categories.

For the reader, a useful distinction is between the economic engine and the accounting milestone. A development contract pays for work performed; a manufacturing contract pays for a specified product or performance obligation. Both can contribute revenue. Their repeatability, cost uncertainty and future investment needs can differ sharply. The correct question is what work generated the revenue and what must happen before that work becomes cash.

15AVAV’s cash bridge: inventory, factory investment and acquisition financing

At August 1, cash was $278.390 million, short-term investments $301.837 million and long-term investments $94.777 million. These components explain management’s approximately $675 million cash-and-investments figure. Calling the whole amount bank cash would obscure the distinction. During the quarter, operating cash flow was positive $13.496 million, while property and equipment purchases were $44.033 million and capitalized software purchases were $5.417 million. The simple subtraction produces an approximately $35.954 million cash shortfall after those investments, consistent with management’s rounded negative $36 million free cash flow. Quarterly financial statements, September 9 call.

The balance-sheet movements explain why the growth phase needs liquidity. Inventories rose from $312.856 million at April 30 to $410.773 million at August 1; unbilled receivables and retentions increased from $570.408 million to $637.832 million. Ordinary receivables moved the other way, providing a substantial cash release. Reading only the net operating cash figure would conceal these offsetting movements. Quarterly balance sheet and cash-flow statement.

Inventory can be a sensible preparation for contracted demand. Its quality still depends on whether the components match the customer’s final configuration, remain usable and enter accepted products on time. Unbilled work presents a different question: what contractual step must be completed before the supplier can issue an invoice? Neither balance is automatically a problem; sustained growth without corresponding billing and collection deserves closer examination.

Factory announcements belong in the same analysis. The March Albuquerque plan committed more than $30 million of company investment, alongside a $6 million state-and-city incentive package tied to performance and employment milestones. The public incentive is not a customer order. New floor space creates the possibility of higher throughput, but staffing, equipment installation and qualified suppliers determine when that possibility becomes usable production. Albuquerque expansion, March 3.

Financing adds another obligation. AeroVironment had $747.5 million principal of zero-coupon convertible notes outstanding, due July 2030. Under the disclosed conversion terms, principal is payable in cash, while value above principal may be settled in cash or shares. Zero coupon therefore does not mean zero refinancing risk or zero dilution potential. Form 10-Q, debt note.

The decisive operating sequence is increasingly specific: convert the existing funded work, deliver the new production awards, collect without a persistent buildup of unbilled balances, and obtain sufficient margin to support the enlarged asset base. A production contract is meaningful progress. The financial outcome depends on this sequence continuing beyond the press release.

16Kratos: unmanned aircraft are one part of the earnings base

Kratos reported $458.8 million of revenue for the quarter ended June 28, 2026. Government Solutions contributed $379.7 million; Unmanned Systems contributed $79.1 million. The latter was approximately 17.2% of the total, calculated from the reported segment figures. Describing every dollar of Kratos revenue as drone revenue would materially misrepresent the company. Second-quarter results, August 4.

The segment distinction does not imply that Government Solutions is unrelated to autonomous systems. Propulsion, communications and electronics can support broader defense platforms. It means that economic exposure must follow the business that receives the work. A satellite gateway order has its own customer and delivery schedule; it cannot establish the number of tactical aircraft ordered. Likewise, enthusiasm for a particular aircraft need not describe the quarter’s strongest revenue contributor.

Consolidated revenue increased 30.5%, while the company calculated organic growth of 19.1%. Government Solutions included the acquired Nomad and Orbit businesses. Unmanned Systems grew 8.1%, with increased Valkyrie activity identified as a driver. These are different comparisons: reported growth includes a larger corporate perimeter, whereas organic growth attempts to isolate the existing business. August 4 earnings release.

Even within Unmanned Systems, maturity varies. On February 25, Kratos announced a $61.068 million Navy contract modification exercising an option for 70 BQM-177A aerial targets and associated equipment. This is an exercised production option for a defined lot. It offers a different kind of evidence from a successful experimental flight. The additional equipment also means that dividing the total by 70 would not produce a clean standalone aircraft selling price. BQM-177A production announcement.

A target aircraft used for testing and training and a collaborative tactical aircraft address different buying decisions. Existing target production demonstrates manufacturing experience; it does not automatically establish the future sales volume of Valkyrie. Shared engineering capabilities can help, but demand, qualification and program funding must still be demonstrated separately.

That separation is also necessary for margins. A company may earn revenue from development, early production and mature repeat lots at the same time. Initial engineering effort and later manufacturing efficiencies can pull profitability in different directions. The useful comparison follows similar program stages over time. A single quarter’s consolidated margin should not be treated as the permanent margin of a future fleet that has not yet been ordered.

17KTOS backlog: funded work, future appropriations and Solar Shield

Kratos ended June with approximately $2.084 billion of total backlog, including $1.572 billion funded and $512.7 million unfunded. Unmanned Systems accounted for $374.6 million of total backlog, with the much larger balance in Government Solutions. The company’s separate $15 billion bid-and-proposal pipeline is a pursuit measure, not another layer of signed backlog. August 4 backlog disclosure.

Its accounting definition provides important discipline: for IDIQ arrangements, only awarded or funded task orders enter backlog. Unexercised options are generally excluded unless they constitute a material contractual right. Kratos expected approximately 35% of its June total backlog to become revenue in fiscal 2026, another 35% in 2027 and the rest thereafter. Those percentages apply to the reported total, not just the funded portion. June 28 Form 10-Q, backlog discussion.

The July 21 Solar Shield announcement demonstrates why the terminology matters. Kratos reported an approximately $156 million sole-source, single-award IDIQ for mobile counter-UAS platforms for the Department of Energy’s National Nuclear Security Administration. The release said work was expected to begin immediately, but did not disclose a complete task-order schedule or an amount funded at signing. The headline value cannot therefore be assumed to represent $156 million of immediate revenue or cash. Solar Shield announcement, July 21.

Single award identifies the supplier structure. It does not, by itself, explain how much the customer will order in each year. Sole source describes the procurement route; it does not remove execution obligations or payment conditions. The practical follow-up is to identify subsequent task orders, funded balances and reported delivery progress where the customer permits disclosure.

Solar Shield also concerns an integrated counter-drone platform, rather than a purchase of Kratos tactical drones. It is an example of the same industry theme creating different revenue streams within one group. The distinction matters when investors compare an aircraft manufacturer, an integrator and a component supplier: each can benefit from demand without receiving the same contract or retaining the same share of value.

Contract visibility is strongest when several records connect: the award establishes scope, the backlog shows remaining work, the revenue statement shows performance and the cash-flow statement shows collection. A missing public detail should remain missing. Confidentiality is not a license to estimate undisclosed volumes, and a large bid pipeline should not be used to fill a gap in awarded business.

18Kratos milestones: Valkyrie tests and the GEK800 are different programs

The January 8 MUX TACAIR announcement identifies Northrop Grumman as the competitively selected contractor, combining its mission capabilities with Kratos’ Valkyrie platform. This is a partnership with a specific division of industrial roles. The prime contractor’s entire program should not automatically be booked as Kratos revenue, nor should a team selection be converted into an undisclosed aircraft quantity. January 8 program announcement.

On August 18, Kratos reported a successful USMC Valkyrie test involving communications and mission-system integration. The event itself took place in April 2026. Keeping both dates avoids manufacturing a new August flight from the publication date. Economically, the result supports a technical development milestone; the announcement does not specify a new production order. August 18 test announcement.

A demonstration can still matter before it produces additional sales. It can reduce uncertainty about integration, support customer evaluation and reveal work needed before acceptance. But successful performance in a defined test is narrower than production readiness across an entire program. The next commercial evidence would be a funded scope of work, delivery commitments or another disclosed milestone with an identifiable customer obligation.

The September 21 GEK800 news belongs to propulsion. Kratos and GE Aerospace confirmed successful ignition of the first engine, starting a new test campaign. GE’s own release independently confirms that milestone. Neither release supplies a serial-production quantity or newly recognized production revenue. The engine’s association with cruise-missile programs also means it should not be casually relabeled a Valkyrie order. Kratos announcement, GE Aerospace confirmation, September 21.

An engine developer can earn money during development while still needing later testing, qualification and production commitments. That is why technical and financial milestones should be tracked together rather than collapsed into one label. An early test is neither worthless nor proof that the future production economics have already been secured.

Manufacturing expansion introduces another clock. Kratos announced more than 100,000 square feet of additional Oklahoma City production space on July 6 for Valkyrie, Firejet and other jet systems. Floor area measures industrial infrastructure, not accepted aircraft output. The economic test is whether equipment, staffing, suppliers and actual customer schedules use that space productively. Oklahoma City expansion.

19KTOS liquidity: equity provides capacity, but execution must earn the return

At June 28, Kratos held $1.4376 billion of cash. Over the first six months, operating activities used $38.4 million, capital expenditure consumed $37.1 million and acquisitions required $346.8 million net of acquired cash. The balance was strengthened by $1.3484 billion of net equity proceeds. This is a well-funded expansion position, but the increase in cash does not demonstrate equivalent cash generation from customer contracts. Quarterly cash-flow statement.

The offering completed on March 2 after the underwriters exercised their option in full. It issued 16,428,571 shares at $84 per share. That dated transaction price explains the financing; it is not a current market-price reference. Existing holders received a better-funded company with a larger share base, so future operating progress must be assessed against both the new assets and the added shares. March 2 Form 8-K.

Working capital is central to the use of that funding. Unbilled receivables were $405.4 million and inventoried costs $235.9 million at June 28. The company also had $139.9 million of finance-lease liabilities. Its revolving credit facility was undrawn, but describing the business as having no financial obligations would ignore those leases. June 28 balance sheet and lease note.

The strategic tradeoff is straightforward. Buying components early can protect delivery schedules and improve purchasing economics. It also puts capital at risk before the final customer payment. A product redesign, a delayed order or a slower acceptance cycle can extend the time that money remains tied up. Cash reserves provide resilience; they do not remove the need to measure the return on the inventory and facilities built with them.

Kratos’ second-quarter free-cash-flow presentation requires particular care. The company reported $11.0 million of operating cash use and $17.2 million of capital expenditure, then included $9.3 million received from selling company-owned Valkyries to arrive at $18.9 million of free cash use. Readers comparing companies should preserve that adjustment. It is not identical to the simple operating-cash-flow-minus-capex calculation. August 4 results and non-GAAP explanation.

The most informative future evidence is therefore a combination of program conversion and capital efficiency: funded orders replacing pursuits, repeat lots following development, revenue becoming billable, and factory investment supporting economical deliveries. None requires assuming that every demonstration becomes a fleet or that every government priority becomes a Kratos award. The company already has substantial commercial activity; the unresolved question is how efficiently the next layer of investment turns into additional, repeatable business.

20A worked example: a large ceiling, a profitable sale and a cash deficit

Consider a fictional procurement framework with a shared maximum of $500 million and five eligible suppliers. Assume one supplier subsequently receives a $30 million funded order, including any applicable minimum commitment. The framework does not give that supplier $500 million of sales, and dividing the ceiling by five does not establish a $100 million allocation. The $30 million order is the starting point for evaluating that supplier’s actual work. None of these figures describes ONDS, RCAT, AVAV or KTOS.

Now assume the company recognizes $12 million of revenue from the order during a reporting period, in accordance with the hypothetical contract’s accounting conditions. If the entire $30 million was originally included in backlog and there are no cancellations, adjustments or additional orders, $18 million remains in that simplified backlog calculation. The $12 million already recognized should not be counted again as future work. An exercised option or scope modification would require a separate update to the bridge.

Assume further that the company has billed the full $12 million but collected $8 million, with no customer advances or other adjustments. The simplified receivable is $4 million. If some recognized revenue had not yet become billable, the accounting classification could differ; that is why the billing condition is explicit. The $8 million collection also should not be added to $12 million of revenue as though the customer had purchased $20 million of goods. It is the settlement of part of the same economic activity.

Suppose the cost associated with that revenue is $9 million, all paid in cash during the same period. Gross profit is $3 million, while collections minus those production payments are negative $1 million. The example excludes overhead, taxes, capital expenditure and financing. A profitable sale can therefore require funding before the customer pays the remaining balance. That does not make the order fictitious; it identifies the amount and timing of the financing bridge under the stated assumptions.

There are several ways the outcome could change. Earlier customer payments would reduce the bridge. A supplier-payment extension would move cash outflow to a later period. Rework could raise cost and reduce gross profit. A delay in customer acceptance could affect recognition or billing, depending on the agreement. These are distinct mechanisms, so they should not be compressed into the single phrase “execution risk.” The next filing is useful when it shows which mechanism actually changed.

The example also explains why comparing a small manufacturer with a diversified defense group requires care. The same order can represent a large share of one company’s annual activity and a modest share of another’s. Existing facilities, supplier terms and access to capital can change the funding burden. That does not establish which stock is more attractive; it shows why identical headline contract values can have different operating and financial consequences.

21The evidence that would make the next quarter more informative

For Ondas, the key reconciliation runs from the acquired perimeter to comparable operating results. The next useful figures would show what the acquired companies contributed after their closing dates, which customer commitments belong to the enlarged group and how acquisition payments affected liquidity. Added technologies can broaden the proposition, but the economic evidence comes from accepted products, repeat customers and the cost of integrating the businesses. A larger portfolio should eventually become a more understandable financial statement, not just a longer list of subsidiaries.

For Red Cat, the useful distinction is between actual deliveries and the opportunities opened by program participation. Evidence of customer acceptance, recurring orders and manufacturing performance can reduce uncertainty in a way that a selection announcement alone cannot. Where a competition includes multiple vendors or further readiness assessments, that context belongs beside the opportunity. Investors can recognize the value of technical progress while preserving the steps still required for serial commercial execution.

For AeroVironment, funded backlog and a larger operating base provide a different starting point. The analysis still needs segment boundaries, program terms and the cost of capacity expansion. A counter-drone award does not turn all group revenue into sales of small aircraft. For Kratos, the same principle separates unmanned systems from other defense activities and separates a propulsion test from series production. The commercial significance of each milestone depends on what it unlocks next, not just on how advanced the technology sounds.

Across the four, a particularly useful next disclosure would connect orders, delivery, revenue and cash within the same reporting perimeter. Some gaps will remain because customers or program details are not public. That absence should be acknowledged rather than filled with a presumed contract value or customer identity. Primary records establish what is known; scenario analysis can explore what might follow, provided the assumptions remain visible. That separation makes the drone theme easier to evaluate through both strong headlines and disappointing quarters.