AXON: A $130 Million Delay Meets a $1.2 Billion Drone Bet

The 60-second read

Five things to know

1. The guidance cut is concentrated. Three Federal Dedrone deployments, totaling approximately $130M, leave the 2026 outlook. The filing supports a timing problem; it does not establish cancellation or a guaranteed recovery date.
2. Software keeps gaining weight. Software & Services reached $437.0M, up 43.2%, contributing a computed two-thirds of the company’s incremental year-over-year revenue.
3. Hardware growth carries a margin cost. Connected Devices gross margin fell to 43.0% from 49.9%; counter-drone mix, memory costs and new-product scaling matter alongside demand.
4. Skydio expands the commitment before cash conversion is proven. Approximately $127.5M cash was paid at the November 5 primary closing; further tender and secondary investments remain conditional.
5. The prior bearish operating tests largely triggered, yet valuation changed. Negative quarterly FCF and inventory above $550M confirm the old cautions. A much lower share price brings the updated $310 reference close to the market.

Three scenarios, one probability-weighted number

Illustrative 12-month valuation states; assumptions, not price forecasts

Signposts: what would change our mind

These green/red levels are our monitoring rules, except explicitly labeled issuer Q4 revenue and FY FCF guide floors. They are not company commitments. A lower entry price only helps while the operating thesis remains intact.

The price reset

The November 6 close was $313, down 18.18% from the preceding $382.56 close. That is the first regular-session close after the November 5 release. The direction comes from the saved daily prices and quote, not the ranker’s absolute-move field. A large decline can reduce valuation risk without repairing the business. It also does not make the former high a useful estimate of intrinsic value.

Our preceding Q2 episode carried SELL, conviction four out of five, a $310 base value and a $609.49 reference price. Today’s $310 estimate retains the prior rounded value under an updated model. The change to HOLD primarily reflects the price approaching the modeled value. This is not a claim that the prior thesis has been disproved or that a fall automatically creates a bargain. The previous cash-conversion conditions are evaluated explicitly in the management section.

The quoted market capitalization is $25.43B. It uses outstanding shares, whereas the valuation uses a more conservative 85.3M diluted share assumption. These are different denominators for different purposes. We do not silently substitute one for the other. The modeled share count includes a buffer for conditional Skydio stock consideration, but future employee awards and convertible settlement choices can still change per-share outcomes.

Source: FMP quote, historical prices, profile and float, retrieved November 8. Monetary tables use US dollars; calculated ratios are marked as computed.

The print and the guide

Axon reported $908.554M of revenue, compared with $710.641M a year earlier and $904.389M in the preceding quarter. The year-over-year comparison is strong; sequential growth is nearly flat. Both observations matter. A business can expand its installed base rapidly while individual large deployments create a lumpy quarterly result. Treating the slowdown as either complete demand failure or a harmless bookkeeping detail would outrun the evidence.

The saved FMP estimate snapshot contains $934.329M of revenue and $1.89 of adjusted EPS expectations, versus reported $908.554M and $1.72. The computed shortfalls are 2.76% and 8.99%. These are retrieved provider expectations, not an independently archived pre-announcement analyst survey. GAAP EPS was $0.68 and must not be compared directly with the adjusted EPS estimate.

The central change is the removal of approximately $130M of Federal Dedrone deployments from the 2026 outlook. Approximately $30M had been expected in Q3 and $100M in Q4. Management explicitly states full-year revenue of approximately $3.61B–$3.63B, about 30% growth versus the prior 32%–34%, and adjusted EBITDA of $924M–$929M at a 25.6% margin versus 25.5% previously. The release describes both full-year ranges as implied by Q4 guidance. Our exact nine-month-plus-Q4 arithmetic gives revenue $3,610.288M–$3,630.288M and EBITDA $924.314M–$929.314M; these extra decimal places are calculations, not additional issuer precision.

The issuer also cut FY free-cash-flow guidance to $350M–$400M from $450M because of accelerated inventory investment. Full-year inventory cash use is expected at $225M, moderating after Q1 2027. SBC guidance fell to $570M–$600M from $590M–$620M, including approximately $240M of Employee XSP and CEO performance awards. Capex fell to $130M–$150M from $160M–$190M, excluding new headquarters investment costs. The reduced cash target raises the burden on year-end collections.

Restoring the excluded deployments mechanically would produce approximately 34.9% growth, compared with 30.2% on the midpoint as issued. That counterfactual helps explain management’s argument. It is not booked revenue, a receivable or a promise that the same dollars will appear next year. Customer milestones and eventual delivery must validate recovery.

Source: Q3 shareholder letter and attached statements. Monetary tables use US dollars; calculated ratios are marked as computed.

Software strength, hardware strain

Software & Services revenue rose to $437.030M, up 43.2%. Connected Devices grew 16.3% to $471.524M. Software contributed a computed 66.6% of incremental company revenue and reached 48.1% of total sales. That is the strongest part of the operating case: growth is broadening beyond a device shipment, and customers can adopt additional services across evidence, real-time operations, records and AI.

Annual recurring revenue reached $1.750B, net revenue retention was 126% and future contracted bookings reached $15.8B. These measures answer different questions. ARR annualizes a recurring revenue base; retention describes expansion in a defined customer cohort; bookings capture future contractual activity. None equals cash in the bank. The 10-Q’s remaining performance obligations were approximately $10.2B, with 20%–25% expected over the next twelve months. The difference from bookings reflects contract definitions and scope, not a missing $5.6B of recognized revenue.

Connected Devices gross margin declined to 43.0%, from 49.9% a year earlier and 51.9% in Q2. Platform Solutions grew 106% to $125.0M, but the company attributes pressure to counter-drone equipment mix, memory costs and the scaling of new categories. Volume alone is therefore an incomplete success metric. Software adjusted gross margin also fell, to 75.0% from 76.8%, even while software-only gross margin remained above 80%. Professional services and the launch mix matter.

The drone partnership could improve the economic mix because Axon describes itself as a channel partner recognizing distribution revenue net in Software & Services. That is a revenue-accounting statement, not evidence that every investment dollar earns a high return. Investors should resist comparing net channel revenue with gross equipment sales as if the two have identical economic denominators. We assign no extra acquisition premium merely because the distribution model sounds attractive.

Source: Latest segment commentary and operating metrics. Monetary tables use US dollars; calculated ratios are marked as computed.

What reaches the income statement

GAAP operating income was $42.080M, a 4.6% margin. Net income of $57.006M was higher than operating income because the below-the-line block includes $59.520M of other income. Interest income contributed $5.657M, interest expense consumed $28.337M and the tax provision was $21.914M. Removing other income from pretax profit leaves $19.4M before tax. That calculation is a diagnostic, not an alternative reported earnings measure.

Adjusted EBITDA was $235.660M, or 25.9% of revenue, up 33.1%. The measure is useful for comparing operating scale, but its distance from GAAP operating profit is large. Total stock compensation was $156.365M. The regular SBC adjustment in the non-GAAP reconciliation was $151.312M; the remainder is included within separately classified non-recurring compensation costs. Mixing those two SBC definitions would double-count part of the bridge.

Adjusted net income was $143.538M. The $86.532M gap over GAAP net income is not simply stock compensation: investment income is removed, amortization and other adjustments are added, and associated tax effects are subtracted. The complete source reconciliation is retained in the dataset. A strong adjusted growth rate is not proof that GAAP earnings are irrelevant; equally, an investment mark can make GAAP net income a noisy proxy for the operating franchise.

For the nine months, other income of $255.722M substantially affected pretax earnings of $311.825M. The year-to-date perspective strengthens the need to distinguish operations from portfolio marks. We do not capitalize those marks as a recurring operating cash stream. Strategic holdings enter valuation separately at a discounted carrying value, with explicit uncertainty and no assumed cash realization.

Source: GAAP statements and non-GAAP reconciliation. Monetary tables use US dollars; calculated ratios are marked as computed.

Nine checks on earnings quality

The cash warning is larger than a single inventory headline. Quarterly operating cash flow was $19.687M and property-and-equipment purchases were $38.588M, leaving negative $18.901M FCF. Approximately $100M of strategic inventory investment explains an important portion of the cash use. It does not explain away receivable growth, contractual billing schedules, employee compensation or the need to fund expansion. One cannot simply add all inventory spending back and treat the result as sustainable owner cash.

Nine-month FCF of negative $74.515M means that the $350M–$400M annual guide requires computed Q4 FCF of $424.515M–$474.515M. That is the concrete cash hurdle. Management says conversion excluding strategic inventory would exceed its 60% long-term target on adjusted EBITDA; that hypothetical adjustment is not cash already available. Even meeting the lower annual guide requires a large seasonal collection rebound.

Receivables were $838.294M, and current plus long-term contract assets totaled $1.166B. Contract assets grew 66.3% year over year, faster than revenue. The issuer’s multi-year customer bundles naturally separate delivery, billing and collection; this is not by itself evidence of impropriety. It does raise the burden of proof for cash conversion. Our end-balance day ratios are deliberately labeled simple proxies and are not presented as average-balance working-capital cycles.

Stock compensation consumed 17.2% of revenue. The reduced annual SBC range of $570M–$600M stays below our prior bearish trigger above $620M; that condition did not trigger. The lower range is favorable evidence even though the absolute economic cost remains substantial. Subtracting it from adjusted EBITDA leaves about $79.3M, still before several costs and reinvestment requirements; that is not a new GAAP metric. We use it to make the economic scale of the compensation visible. Diluted GAAP shares rose from 78.416M to 83.673M, but last year’s loss excluded anti-dilutive securities. The non-GAAP year-ago diluted denominator was 83.045M. Treating the entire GAAP denominator difference as new issuance would be misleading.

The 2024 debt-classification restatement is a resolved historical issue. The filed amendment moved convertible notes between current and noncurrent presentation; it did not change total debt, total liabilities, net income or cash flow. The current Q3 10-Q reports effective disclosure controls and no material quarterly ICFR changes. We retain the history, use amended affected classifications and do not turn a resolved event into a permanent exclusion.

Source: Q3 10-Q, working capital and controls; annual history from FY2025 10-K. Monetary tables use US dollars; calculated ratios are marked as computed.

Funding the drone strategy

At September 30, cash and equivalents were $1.459B and short-term investments were $181.6M. Notes principal totaled $2.900B. Our net-debt definition is therefore $1.260B, excluding $13.211M of restricted cash and $14.751M of marketable securities. The company’s supplemental table includes restricted cash and reports $1.246B net debt. Both definitions are shown; our valuation uses the more conservative one consistently. Carrying debt of $2.868B is lower because of issuance-cost accounting and is not substituted for principal.

The subsequent Skydio agreements call for incremental investments of up to approximately $1.2B, with up to approximately $720M cash and the remainder in Axon shares. The primary investment of approximately $254.9M closed November 5, half cash and half stock; about $127.5M cash was paid, and a similar stock delivery commitment was recorded. Further tender and secondary purchases, valued at up to approximately $968.6M in the note, are expected through Q4 2026 and Q1 2027. These detailed amounts are approximate and need not sum to an exact $1.200B headline.

A mechanical use of the full $720M against September liquidity would leave about $920M of cash and short-term investments and increase net debt to about $1.980B, before intervening cash flows. That is a scenario, not a current balance sheet. It excludes cash generated or consumed after September, transaction fees, acquisitions and any changed final mix. The already-paid $127.5M is part of the full cash commitment, not an additional deduction on top.

The investment is expected to take fully diluted Skydio ownership to 37%, yet the filing states Axon lacks significant influence and uses the measurement alternative for nonmarketable equity. Ownership percentage alone does not determine accounting control. The valuation must separate distribution economics, investment asset value, cash funding and stock dilution. Counting the entire drone opportunity as new operating revenue and also adding the stake at a premium would risk paying for the same optimism twice.

Source: Note 15 subsequent events and liquidity discussion. Monetary tables use US dollars; calculated ratios are marked as computed.

What $313 already requires

Our base value is $310 per share, with high uncertainty and three-out-of-five conviction. The three routes produce approximately $299, $319 and $310. The revised cash ramp lowers the DCF after including the verified annual cash and compensation outlook; the base EBITDA multiple is reduced from 24x to 23x. The rounded $310 base sits near the middle of these cross-checks. The rounded conclusion is not the output of averaging away uncertainty. All three methods depend on persistent software growth, better cash conversion and a continued premium valuation. Agreement among related models cannot protect against the same mistaken operating assumption.

The economic DCF begins with computed FY2026 revenue of $3.620B. Revenue growth assumptions for 2027–2031 are 28%, 25%, 23%, 20% and 18%; unlevered economic free-cash-flow margins after recurring SBC cost are 4%, 8%, 12%, 17% and 21%. These are demanding analyst assumptions, not management guidance. FY2026 FCF midpoint of $375M less the $585M SBC midpoint is negative $210M. That levered diagnostic is not directly interchangeable with unlevered economic FCF, but it exposes the recovery required: the revised 2027 model still assumes about $185M of economic FCF. Inventory normalization, operating leverage and financing/tax differences need validation; no inventory add-back is automatically capitalized. Cash flows are discounted at 10%, with 4% terminal growth. The resulting enterprise value is $26.304B. Deduct $1.260B net debt, add a 50% credit on $924.934M existing strategic investments and divide by 85.3M diluted shares: approximately $299.

The second route values assumed 2027 Software & Services revenue of $2.2B at ten times and Connected Devices revenue of $2.4B at two and a half times. That yields $28B enterprise value before the same bridge and approximately $319 per share. Multiples are analyst choices and are not an observed peer transaction appraisal. The third route uses $4.55B revenue, a 26% adjusted EBITDA margin and a 23-times multiple, yielding about $310. Because EBITDA excludes recurring equity compensation, this route deserves less confidence than a properly specified cash model.

Our enterprise quote arithmetic is market capitalization plus net debt, or $26.687B: about 7.37 times computed 2026 revenue and 28.79 times computed 2026 adjusted EBITDA. Even after the selloff, a substantial future improvement is priced. In the DCF, matching $313 requires margins approximately 1.045 times the base margin path, with every other assumption fixed. The price is roughly on our demanding cash case, not below a conservative liquidation anchor.

Terminal value contributes about 86.7% of DCF enterprise value. That concentration is the reason for the high uncertainty rating and a chosen 25% margin of safety. Our $232.50 entry reference is conditional on unchanged fundamentals. For future Skydio transactions, cash-funded acquired value is assumed at cost and offset by the cash consideration. No separate credit is added for stock-funded acquired value, while the diluted share buffer remains. That is an intentionally conservative reserve using the September numerator, not a complete transaction pro forma. Actual closing terms require a fresh asset, cash and share reconciliation.

Wall Street versus our model

The saved FMP rating distribution has eighteen buys, four holds and no sells. Its target average is $612.56, median $615, low $500 and high $715. The apparent 95.7% upside to the average is a provider snapshot, not a promise and not a uniformly refreshed post-results model. Individual dated revisions are more informative than treating that average as a current consensus conclusion.

November 6 target news records Northland at $540, RBC Capital at $615 and Piper Sandler at $540. Those are post-release observations, unlike older target records in the same feed. We attribute the targets to FMP’s dated news records and do not claim to have read proprietary analyst notes. The snapshot preserves the distinction between a firm’s rating, its target and our own valuation.

The provider’s FY2026 revenue average is $3.711B, above the company-guidance-derived $3.620B midpoint. FY2027 averages $4.806B, with fifteen revenue analysts, and FY2028 $6.165B, with fourteen. The difference can reflect incomplete refresh, a different growth view or definitional differences. We therefore use our own disclosed revenue assumptions in the valuation and show the provider path as context. A number retrieved after earnings is not automatically an estimate revised after earnings.

Our caution is not disagreement with the installed-base story. We demand a stronger bridge from recurring revenue to cash available per diluted share, and we assign no automatic premium to the new capital commitment. The Street may be right that AI adoption and drone distribution accelerate enough to justify the higher values. The signposts specify what evidence would make that outcome more credible.

Source: Saved price-target news, grades and annual estimate responses. Monetary tables use US dollars; calculated ratios are marked as computed.

Grade the old tests

The prior episode’s bullish conditions were FCF above $100M, contract assets growing more slowly than revenue, or Software & Services adjusted gross margin recovering above 78%. The current quarter fails all three tests: FCF is negative $18.9M, combined contract assets rose 66.3% versus 27.8% revenue growth, and software adjusted gross margin is 75.0%. Two explicit bearish conditions also triggered: another negative-cash quarter and inventory above $550M. Inventory reached $579.2M. The third bearish condition, annual SBC guidance above $620M, did not trigger: management guided to $570M–$600M.

This is a useful example of separating a business judgment from a stock judgment. Operating concerns have not disappeared. Nevertheless, the former $609.49 price was almost twice the old $310 reference. The current $313 price is close to the new $310 base. Moving from SELL to HOLD acknowledges a changed price-to-value relationship; it does not erase the failed operating checkpoints or retroactively rewrite the old call.

Management deserves credit for software growth, continued customer expansion and willingness to disclose the concentration of the federal guide reduction. It also faces a capital-allocation test: cash generation remains weak while acquisitions, strategic investments and equity compensation expand the claim on future owner cash. The 2025 annual report records no history of cash dividends and a small remaining buyback authorization; this is a reinvestment proposition, not an income distribution thesis.

The questions we would ask are specific. What contractual milestones move each delayed federal deployment into revenue? How much of the inventory build is required for already contracted demand? When do contract assets become billable and collected? What return threshold governs Skydio’s cash and stock funding? How much recurring equity compensation can scale below revenue while retaining engineering talent? Answers should provide dates, amounts and definitions that can be checked next quarter.

Source: Current evidence; historical thresholds retained from prior episode metadata. Monetary tables use US dollars; calculated ratios are marked as computed.

Ranked risks and counterarguments

The strongest bullish argument is that software adoption can outrun the operational friction. Retention above 120%, an expanding installed base, new AI tools and distribution relationships can improve lifetime economics. Hardware deployments can create access to recurring services. Inventory purchased ahead of demand may protect availability, and federal timing can reverse without a fundamental loss of customer need. Under that path, a depressed valuation could reward patience.

The strongest bearish argument is that investors keep capitalizing adjusted growth while the cash conversion and funding burden worsen. Device mix may remain lower margin; memory costs may persist; government approvals may be slow; and the strategic investment portfolio can absorb capital whose returns are difficult to observe. A premium stock multiple magnifies the impact of a merely good outcome replacing an exceptional one.

Our middle case accepts substantial software growth and an improving economic cash margin while refusing to assume all delayed deployments immediately return. It also recognizes that Axon operates in sensitive public-safety markets. Product performance, liability, privacy, procurement restrictions, AI reliability and public acceptance can influence adoption. These risks are not captured by a revenue spreadsheet alone, and the financial consequences may arrive discontinuously.

The scenario range is intentionally wide. The $200 bear value does not require the business to fail; weaker margins combined with a lower multiple can get there. The $480 bull value requires both operating delivery and a premium market valuation. Probabilities summarize judgment at this date, rather than a statistically estimated distribution. Investors should update them when evidence changes, particularly the cash and funding disclosures.

Dated evidence to watch

The next earnings release should clarify whether Q4 met the $990M–$1.01B revenue guide and whether Q4 FCF delivered the calculated $424.515M–$474.515M needed to achieve the reduced annual $350M–$400M guide. We use February 28, 2027 as a review deadline, not a confirmed company release date. A provider calendar estimate is not substituted for an issuer announcement. The same discipline applies to the Q1 review deadline: it is a research checkpoint.

Skydio tender and secondary transactions are expected across the fourth quarter and first quarter. Subsequent filings should reconcile cash paid, shares delivered, acquired carrying value and the final ownership position. The current investment announcement is not a completed pro forma balance sheet. A verified change in terms should flow into both the valuation numerator and the diluted share denominator.

Federal deployment updates need more than an expression of confidence. Quantified customer milestones, shipment acceptance, deployment timing and recognized revenue would support a recovery thesis. If there is no material update, the lost 2026 dollars should not automatically become an assumed 2027 beat. We will keep the uncertainty visible until evidence resolves it.

Product launches such as Body 5, TASER Ion and expanded AI services create a wider adoption opportunity. Revenue, retention, professional-services mix and gross margin together will show whether that opportunity is becoming attractive economics. The most useful catalyst is not a product announcement by itself, but a measurable improvement in the connection between customer adoption, billing and cash after compensation and reinvestment.

Data, definitions and sources

The quarterly history below covers nine consecutive periods from Q3 2024 through Q3 2026. Financial amounts come from saved SEC earnings exhibits, with complete income, reconciliation, cash-flow, operating-metric and balance-sheet tables retained in the research dataset. Historical reportable segments changed; the continuous product/service revenue categories should not be mistaken for unchanged old segment definitions. A quarter and a nine-month period are kept separate throughout. Starting Q1 2026, adjusted net income/EPS excludes acquired-intangible amortization and the issuer recasts comparative periods, including taxes. Our current comparison uses the latest recast Q3 2025 adjusted NI of $102.746M and EPS $1.24; the retained quarterly history keeps its original reporting basis. Adjusted EBITDA changed in Q1 and adjusted EBITDA/gross margin changed for XSP taxes in Q2; prior periods were not recast because the issuer calls the impact de minimis.

The 2024 affected current/noncurrent debt classifications use the filed May 2025 amendment. Original release documents remain in the source archive as historical evidence, but their affected classification rows are not used as current analytical inputs. Total debt and total equity were unchanged by that correction. Current controls are evaluated from the latest 10-Q rather than inferred from the age of the historical issue.

FCF means operating cash flow less property-and-equipment purchases unless an issuer-specific definition is explicitly noted. The 2023 issuer measure also deducted intangible purchases; our annual PP&E-only proxy is therefore $0.635M higher. SBC is an economic cost even though it is noncash in the operating cash-flow reconciliation. The DCF’s assumed economic FCF margins include that cost; the adjusted EBITDA cross-check does not become an owner-cash measure merely because it produces a similar value.

All valuation assumptions are reproducible in the episode’s calculation source. Probability-weighted value is an arithmetic expectation across three illustrative states, not a target guaranteed to be reached. Fair-value ranges are not statistical confidence intervals. Sources were checked as of November 8, and market prices are dated November 6. The analysis should be revised when customer deployment, funding, margins or dilution evidence changes. No personal portfolio circumstances are assumed.

• Q3 2026 shareholder letter: actuals, guide, segment and cash reconciliations
• Q3 2026 10-Q: balance sheet, controls, debt and Skydio subsequent events
• FY2025 10-K: three-year cash flow, structure, audit and capital policy
• FY2024 10-K/A: resolved debt-classification correction
• FY2023 10-K: employee count and historical structure
• Q3 2024 earnings filing
• Q4 2024 earnings filing
• Q1 2025 earnings filing
• Q2 2025 earnings filing
• Q3 2025 earnings filing
• Q4 2025 earnings filing
• Q1 2026 earnings filing
• Q2 2026 earnings filing
• FMP saved quote, daily prices, float, earnings estimates, rating distribution, targets and annual consensus; accessed November 8, 2026.