Autonomous Last-Mile Delivery — Sector Report

Executive Summary

Autonomous last-mile delivery has crossed from pilot theater into early commercial scale, and 2026 is the sector's breakout year: Walmart passed 1 million cumulative drone deliveries — 40% of them in FY27 Q1 alone [13] — Zipline's valuation doubled to $7.6B [48], and disclosed sector funding in the first eight months of 2026 (~$1.3B) exceeds full-year 2024 and 2025 combined (Section 4). Yet no scaled operator has been acquired outright, and the FAA's Part 108 BVLOS rule — the sector's single largest repricing catalyst — remains months from finalization [45].

Key judgments:

1. The economics now work — but only on constrained, dense routes. Sidewalk robots already deliver at $1.50-3.00 per trip versus a fully loaded human courier cost of $8-12 [166, 174]; drones are economic for small, urgent orders at network density — not as general delivery replacement. Utilization, not technology, is the binding constraint: it is what killed DroneUp at ~$30/delivery [28] and what makes marginal independent assets accretive inside Walmart's 4,600-store network [29].
2. Drone control positions are repricing fast; ground assets remain accessible. Zipline is at $7.6B and Wing is captive to Alphabet [48]; by contrast, scaled sidewalk fleets (Starship, Coco) trade at venture-scale marks, the enabling hardware/UTM layer is modestly valued, and Gatik — already integrated into Walmart's middle-mile — just reset its ownership base with QIA/Koch capital [73, 82, 128].
3. Part 108 finalization is the clock. The final BVLOS rule entered OIRA review in July 2026 [45]; publication would standardize what today requires bespoke waivers and likely trigger a step-change in drone-operator and UTM valuations. Positions taken before finalization capture the repricing.
4. Competitors are assembling multi-modal networks now. Amazon acquired RIVR and is scaling Prime Air toward ~500 locales [54, 156]; Uber holds stakes across Zipline, Flytrex, Nuro, and Avride [105]; DoorDash backed Also and built its own robot [143, 145]. The current partnership-based equilibrium is hardening into closed networks.

1. Market Overview & Size

Autonomous last-mile delivery has moved from pilot theater to early commercial scale. The sector spans three delivery modalities — aerial drones, sidewalk robots, and autonomous road vehicles — plus an enabling layer of drone hardware and uncrewed traffic management (UTM) platforms. Published market sizings vary widely by definitional scope, and the leading segment estimates overlap; they should be read as indicative, not additive.

Best available estimates place drone delivery (hardware, autonomy software, docking infrastructure, and integration services) at approximately $1.9B in 2025, growing at a 21.6% CAGR [1]; sidewalk/ground delivery robots at approximately $1.1B, growing ~19.7% [2]; and the broader autonomous ground last-mile vehicle segment at approximately $1.6B, growing ~24.8% (derived from a $1.96B total autonomous last-mile market with an ~82% ground-vehicle share) [3]. The drone hardware & UTM layer lacks a cleanly published standalone size; we estimate it at roughly $0.6B, treating enabling hardware, docking, and airspace software as approximately one-third of the drone delivery market's defined scope [1] (see Analyst's Notes). McKinsey and Deloitte both identify autonomous ground vehicles, drones, and delivery droids as the defining last-mile technologies, with Deloitte reporting that 80% of surveyed companies were investing or planning to invest in them [5, 6].

Trajectory to 2030 and Global Benchmarks

The sector's growth story is steep but front-loaded in the drone segment. Grand View Research projects the total autonomous last-mile delivery market growing from $1.96B (2025) to $5.93B by 2030, a 24.8% CAGR [3]. Within that, drone delivery is expected to nearly triple its share — from roughly $2.9B in 2025 to $8.4B by 2030 on broader scoping that includes hardware and services [179] — while sidewalk delivery robots grow to roughly $2.8B [180]. Applying the current ~82% ground-modality share to the 2030 total implies an autonomous ground last-mile segment of roughly $4.9B (derivation in Analyst's Notes) [3].

China is the operating benchmark, not the market forecast. Meituan has completed 900,000+ cumulative drone orders across 500+ route sites, with per-site daily capacity scaling from ~10 to ~400 deliveries as route density matured [210, 211] — yet drone volume still represents roughly 0.01% of Meituan's total order flow, a sober reminder of how long the tail of scaling is even in a permissive regulatory environment [210]. JD.com operates 3,200+ autonomous delivery units and is deploying 200-kg-payload drones for rural logistics [212], and Chinese forecasts anticipate the national autonomous delivery fleet growing from ~47,000 vehicles today to 1.5M by 2030 [216]. The lesson for U.S. strategics: the technology scales operationally where regulation permits, but autonomous share of total delivery volume remains small for years even at the frontier — network economics, not novelty, determine the ramp.

2. Growth Drivers

Fulfillment-from-store economics. Walmart's ~4,605 U.S. stores — substantially all offering same-day pickup and delivery — function as a distributed forward-inventory network, the structural precondition for autonomous last-mile economics [29]. U.S. e-commerce reached ~$79.3B in FY25 [29], store-fulfilled delivery grew ~50% in FY26 and ~45% in FY27 Q1, and expedited sub-three-hour deliveries now represent ~36% of store-fulfilled orders [35, 36]. Net delivery cost per order fell 20-40% across FY25 quarters on ~20% better batch density [31, 33, 34]. Autonomy is the next cost-curve lever: it targets the driver — the largest remaining variable cost per drop.

Labor cost pressure defines the target economics. DroneUp's experience is the sector's cautionary benchmark: ~$30 cost per drone delivery against a sub-$7 viability threshold, which drove Walmart to end that partnership at end-2024 and consolidate around operators with stronger automation and utilization [24, 28]. Every credible player's roadmap is a march from labor-intensive pilots toward high-utilization, low-touch networks.

FAA BVLOS (Part 108) rule progress. The FAA published its proposed beyond-visual-line-of-sight rule in August 2025 [38]; after two comment periods (closing October 2025 and February 2026), the draft final rule reportedly entered OIRA review in July 2026 but is not yet final as of August 2026 [39, 42, 45]. Current operations run on Part 107 waivers, Section 44807 exemptions, and Part 135 certifications — workable but bespoke. Finalization would replace operator-specific approvals with standardized performance requirements, materially lowering the cost and time to scale drone networks.

Walmart's own drone expansion is the sector's largest demand catalyst. Walmart crossed 1 million cumulative drone deliveries in May 2026 — 40% of them in FY27 Q1 alone — at a 23-minute average delivery time, from ~66-70 stores [13, 14]. Wing plans to reach 270+ Walmart stores and ~40 million Americans by end-2027, with Zipline expanding regionally across DFW [10, 13, 18].

Unit Economics and Willingness to Pay

Cost per delivery is the sector's single most important number — and the spread across operators today is enormous. Amazon's internal 2025 projection reportedly put Prime Air's cost at ~$63 per delivery [168]; Zipline's 2023 cost was independently estimated at ~$13.50 [162]. Against a fully loaded human courier cost of $8–12 per drop [166], autonomy only wins at scale: analyst estimates put scaled drone operations at $2.70–8.40 per delivery [166], sidewalk robots at $1.50–3.00 today [174], and operators including Zipline and Serve target ~$1–2 at maturity [164, 175].

The gap between the $63 and $1.50 rows is not primarily technology — it is utilization. Fixed costs (aircraft, nests, remote operators, compliance) dominate, so cost per delivery falls almost linearly with deliveries per asset per day. This is precisely why Walmart's fulfillment-from-store density is a structural advantage: 4,600+ stores within 10 minutes of 90% of the U.S. population creates the route density that drives utilization, which no standalone operator can replicate.

Consumer willingness to pay bounds the revenue side. McKinsey finds 76% of global consumers open to switching to autonomous delivery and 58% willing to pay a premium — but only 37% of U.S. respondents [217]. Auterion's U.S. survey found 41% unwilling to pay any drone-delivery fee, 41% willing to pay $1–10, and 18% more than $10 [218]; a Portland-area study measured mean willingness to pay at just $2.25 per delivery [222]. The most encouraging datapoint: ~80% of consumers are open to a $3–5 fee when it buys 30-minute delivery [220].

3. Technology Segments

3a. Drone Delivery Operators

Zipline leads on cumulative scale — 2M+ commercial deliveries, 100M+ autonomous miles — and private-market value, at a $7.6B valuation after raising ~$800M in 2026 [46, 48, 112]. Wing (Alphabet) has passed 1M deliveries and holds the strongest retail distribution channel via its 270+-store Walmart commitment [10, 12]. Amazon Prime Air lags operationally (~16,000 reported deliveries from 11 locations) but plans expansion toward ~500 locales at a $4.99 price point [53, 54]. Manna (300K+ deliveries; $110M raised) and Flytrex are regional specialists, both integrated with DoorDash — which itself obtained FAA Part 135 certification in July 2026, shifting from aggregator to operator [55, 56, 57, 59].

3b. Drone Hardware & UTM Platforms

Matternet — the only FAA type-certified drone-delivery platform — went public via reverse merger in May 2026, raising ~$33M [62, 63]. A2Z Drone Delivery supplies winches, delivery drones, and automated docking infrastructure to operators [64]. On the airspace layer, ANRA Technologies enabled 55,000+ commercial drone operations per month by June 2026 (users include Amazon Prime Air, DoorDash, Manna, and Matternet) [72], and Airspace Link holds FAA approval as a UTM provider for BVLOS strategic deconfliction [69, 70]. This layer monetizes regulatory infrastructure rather than deliveries — a distinct, capital-light value pool that grows with total flight volume regardless of which operator wins.

3c. Sidewalk Delivery Robots

Starship Technologies has the deepest operating record: 9M+ deliveries across seven countries with ~2,700 robots, funded by a $50M Series C (October 2025; $280M+ total) [73, 74]. Serve Robotics, the public pure-play comparable, reached ~2M cumulative deliveries; Q2 2026 revenue was $3.2M (+404% YoY), but a guidance cut from ~$26M to $9-10M for 2026 — driven by a major partner's volume shortfall — illustrates the segment's utilization and customer-concentration risk [77, 79, 80]. Coco Robotics (500K+ deliveries; $80M raised June 2025, backed by Sam Altman) targets a 10,000-robot fleet [81, 82]. Avride secured up to $375M in strategic commitments from Uber and Nebius and delivers for Uber Eats in four markets [91, 92].

3d. Autonomous Last-Mile Vehicles

Gatik is the most commercially proven — fixed-route, fully driverless middle-mile trucking with Walmart as its first major customer, plus Kroger and Tyson; it raised a $200M Series D in August 2026 led by Qatar Investment Authority and Koch (~$500M cumulative) [99, 100, 128]. Nuro ($6B valuation; $2.3B raised) has pivoted from owned delivery fleets to licensing its autonomy stack, anchored by an Uber/Lucid robotaxi program approaching $500M in Uber commitments [96, 97]. Waymo's delivery activity remains ancillary to robotaxis (Uber Eats/DoorDash in Phoenix) [103, 104]. Also, the Rivian spinout, raised a $200M Series C at a ~$1B valuation with DoorDash as investor and vehicle-development partner [143]. Europe's Clevon was acquired by indiGO Tech in 2025 — an early consolidation marker [107].

4. Recent M&A and Funding Activity

Capital formation has accelerated sharply since early 2025, concentrated in a handful of category leaders; outright M&A remains limited and small, dominated by capability tuck-ins and distressed asset deals — a signal that consolidation is early and strategic acquirers still have room to move before assets reprice.

Capital formation has accelerated sharply in 2026: disclosed equity funding across the tracked companies has more than doubled versus each of the prior two full years, led by Zipline's $800M round [48], Gatik's $200M, and Also's $200M spin-out raise. The 2026 figure excludes undisclosed strategic investments (e.g., Uber's investment in Zipline) and Avride's up-to-$375M in conditional commitments — actual capital committed to the sector is materially higher (methodology in Analyst's Notes).

Two structural patterns stand out. First, demand platforms are buying optionality through minority stakes, not control: Uber has invested in Zipline, Flytrex, Nuro, and Avride, while DoorDash backed Also and built its own "Dot" robot in-house [105, 141, 143, 145]. Second, the only completed acquisitions have been small capability tuck-ins (Serve's Vayu, Diligent, and teleoperation deals; Amazon–RIVR; Kiwibot–AUTO; Kite Aero's purchase of distressed Swoop Aero assets) [147, 153, 156, 159] — no scaled operator has yet been acquired outright.

5. Key Risks & Watch Items

6. Outlook for Strategic Acquirers

The window for control transactions is narrowing in drones and still open on the ground. Zipline's valuation doubled to $7.6B in under two years and Wing is captive to Alphabet [48], making aerial-operator control transactions expensive or unavailable. Ground segments remain accessible: Starship (~$280M raised) and Coco offer scaled sidewalk fleets at venture-scale valuations [73, 82], and Serve trades publicly with a depressed 2026 outlook following its guidance cut and Uber's full exit from its stake [79, 139].

Three considerations should frame corporate-development posture. First, partner-versus-own economics favor owning where utilization is captive. Walmart's store density means a dedicated autonomous fleet (aerial or ground) can achieve the utilization that killed third-party models like DroneUp [24, 28]; assets that are marginal for an independent operator can be accretive inside Walmart's network. Second, Part 108 finalization is the repricing event. A final BVLOS rule — plausibly within 12 months given OIRA review began July 2026 [45] — would de-risk drone operations at scale and likely trigger a step-change in drone-operator and UTM valuations; positions taken before finalization capture that repricing. Third, competitors are assembling multi-modal networks now. Amazon is buying robotics capability (RIVR) and scaling Prime Air toward ~500 locales [54, 156]; Uber and DoorDash are stitching together drone, sidewalk, and AV coverage through investments and partnerships [105, 143]. Walmart's structural advantages — store density, basket economics, and proven delivery demand — are strongest if paired with ownership or exclusivity in at least one modality before the sector's current partnership-based equilibrium hardens into closed networks.

Catalyst Timeline

Indicative Target Screen

Analyst's Notes

Derivation of segment estimates (Section 1 chart). The autonomous ground last-mile figure (~$1.6B, 2025) is derived: $1.963B total autonomous last-mile market × ~82% ground-vehicle share, per Grand View Research's published scope [3]; 2026 extension applies the reported 24.8% CAGR. The drone hardware & UTM estimate (~$0.6B) is analyst-derived as roughly one-third of the $1.94B delivery-drone market, whose published scope explicitly includes hardware, payload systems, docking/charging infrastructure, and integration services alongside operations [1]; no reputable standalone sizing of the hardware/UTM layer was found. Segment figures overlap (sidewalk robots are a subset of ground vehicles) and derive from commercial research publishers whose figures should be treated as directional; McKinsey and Deloitte corroborate direction but publish no standardized market value [5, 6]. Private-company valuations cited throughout are funding-round marks, not audited values.

Derivation of the 2030 ground-segment estimate (§1): Grand View's 2030 total-market projection of $5.93B [3] multiplied by the current ~82% ground-modality share [3] ≈ $4.9B. This assumes the modality mix holds; if drone delivery grows faster (as [179] implies), the ground share — and this figure — would be lower.

Funding-by-year chart methodology (§4): Sums of disclosed equity rounds among tracked companies only. 2024 ≈ $596M (Zipline $350M, Starship $90M, Serve Robotics $126M, Gatik $30M). 2025 ≈ $513M (Nuro $203M, Serve $180M, Coco $80M, Starship $50M). 2026 YTD (through August) ≈ $1,283M (Zipline $800M, Gatik $200M, Also $200M, Manna $50M, Matternet $33M). Excluded: undisclosed strategic investments (e.g., Uber into Zipline), Avride's up-to-$375M conditional commitments, and debt facilities. These are floor figures for capital formation, not totals.

Unit-economics chart caveats (§2): Figures mix vintages and estimation quality — Amazon's ~$63 is a reported internal projection [168]; Zipline's ~$13.50 is a 2023 third-party estimate [162]; scaled-operation ranges [166, 174] are analyst modeling; and the $1–2 row reflects operator-stated targets [164, 175], not achieved costs. Range midpoints are charted. Treat the chart as directional ordering, not a like-for-like comparison.