DJI used two weeks to anoint the Pocket 4P as a cinema product at Cannes rather than a consumer product at its own event, absorb a fourth US-block casualty with Osmo Mobile 8P, watch the FCC fight enter its most adversarial phase, and collect NAB hardware that validates the pivot. Here's what happened — in priority order, with our take on each.
DJI made more news in seven days than most companies make in a quarter. A new product launched (and got blocked from US sale, again). A major event got teased. The FCC fight escalated to a $1.56 billion court filing. And DJI's home market joined the regulatory pile-on. Here's what happened, in priority order, with our take on each.
DJI confirmed a global launch event for May 7, 2026 at 12:00 UTC. Two products are on deck: the Osmo Mobile 8P global rollout (already on sale in China since April 21) and an expected Pocket 4 Pro reveal with dual-lens optics. Here's the complete preview, what we know about each product, and what US buyers should expect given the ongoing FCC ban fight.
This is the week the FCC fight stopped being only about DJI. Autel walked into the docket and publicly broke ranks — going on the record with encryption disclosures, sworn declarations, and a Fifth Amendment due-process attack citing the 2014 Ralls Corp. precedent. Michigan's 15-bill SHIELD drone package collapsed in committee, passing only the two procurement bans that just mirror the federal default. The Air 4 surfaced inside a Chinese flight-filing app in a sub-250g classification that would let DJI sidestep the entire FCC pathway. And the Pocket 4P pricing silence — one of last roundup's "watching next week" items — is now ten days deep. Six stories, in priority order, with our take on each.
On May 19, 2026, Autel Robotics filed its formal reply at the FCC challenging the agency's December 2025 decision to add Autel to the Covered List. The filing does something Autel had not done publicly before: it goes on the record about how Autel actually handles flight data, and it explicitly separates Autel's case from DJI's. The substance lands in three places. Encryption and storage: sworn declarations describe flight data as stored locally by default, with cloud backup disabled unless the user opts in, US customer data on US servers, and AES-128 or AES-256 encryption protecting communications. Government access: the company asserts under oath it has never received a data request from the Chinese government or any other adversarial nation. Procedural attack: Autel argues the FCC's reliance on classified intelligence the company was never allowed to see violates Fifth Amendment due process, citing the 2014 Ralls Corp. v. CFIUS decision — the case where the Federal Circuit held that companies subject to national-security adverse action are entitled to notice of the unclassified portions of the evidence against them.
The filing's most cutting line, lifted directly from the brief: "A company-specific federal designation cannot be justified by importing one company's alleged technical defects into the case of another." That is Autel saying, in legal English, stop putting us in DJI's file.
Our take: This is the most credible legal attack the entire proceeding has seen, and it changes the political shape of the fight. DJI structurally cannot run the "we're different from the other guy" argument — DJI is the other guy. Autel can. By going on the record about encryption, storage, and government access in a way DJI has been more cautious about, Autel signals to the FCC commissioners that there's a procedural off-ramp — soften the Covered List on Autel without conceding anything on DJI — that lets the agency claim a win on national security while reducing the political cost of the carveout. The Ralls citation matters too: that's the case future drone-ban challenges will be built on, because Ralls established that even in national-security cases the affected party gets notice of unclassified evidence. The Pentagon's classified opposition (covered in our prior roundup) was DJI's ceiling; Ralls is Autel's floor. For DJI, Autel's separation strategy is double-edged — if Autel walks while DJI doesn't, the ban looks less like a foreign-drone policy and more like a DJI-specific policy, which is exactly the framing DJI has been trying to avoid. Track the full FCC picture on the US availability hub.
DJI revealed the Osmo Pocket 4P at Cannes on May 14, 2026. The press release confirmed the dual-camera hardware — 20mm f/2.0 1-inch main + 60mm-equivalent (3×) telephoto on a 1/1.5-inch sensor, 14-stop dynamic range, 10-bit D-Log2, 4K/240fps, ActiveTrack 7.0 — but deliberately held back final specs, configurations, pricing, and US availability, promising more "at a later date." As of May 24, ten days after the reveal, that later date has not arrived. The pre-launch chatter clusters around $700 base / ~$733 USD-equivalent from Chinese retail estimates, with shipping expected to begin in June 2026. None of that is official.
Our take: The silence is the strategy. In a normal consumer-electronics launch, pricing drops with the reveal — withholding it costs retail momentum, depresses pre-order velocity, and gives competitors a free week to position. DJI did it on purpose. Two reasons fit the pattern. First, framing the Pocket 4P as a cinema-tier product (Cannes stage, "Wonders" cadence, festival press) means the launch doesn't trade in the consumer pricing cycle — it trades in the cinema-product cycle, where reveal-then-ship-then-pricing is normal (think RED, ARRI, Blackmagic). Second, withholding pricing buys DJI time to read Autel's FCC filing, the post-comment FCC response, and the gray-market signal on Mic Mini 2 / Mobile 8P before committing to a configuration ladder. If Autel-style separation buys DJI any softening, the Pocket 4P pricing might come in at a US-friendly tier. If it doesn't, DJI prices it as a global-minus-US product without burning the announcement window. The Pocket 4P stays in "official pricing TBD" status on our Pocket 4 hub until DJI actually says a number.
On May 14, 2026 (reported in roundup coverage on May 22), the Michigan House passed two bills out of state Rep. William Bruck's 15-bill SHIELD drone package: HB 5329 and HB 5331, both prohibiting public entities from contracting to buy drones containing software developed by companies on the NDAA Section 1260H list. The remaining thirteen bills — including critical-infrastructure overflight penalties, drone-interception authority for state and local police, a mandatory state-run geofencing app, and broader "no drone zones" provisions — stalled in committee. Industry observers reading the procedural posture characterize the stall as a meaningful win against state-level overreach.
Our take: What passed is just the federal default. NDAA Section 1260H procurement bans on Chinese-made drones for federal entities have been the law since 2024; Michigan adding state-agency procurement to the same list is symbolic, not operational. What didn't pass is where the stakes actually were. State-mandated geofencing apps would have fragmented airspace authority — a precedent every other state legislature was waiting to copy. Drone-interception authority for state and local police would have created a patchwork of jurisdictional kill rights that the FAA has spent years preventing. Critical-infrastructure overflight penalties would have applied to commercial drone operators flying for utility inspections, infrastructure surveys, and journalism. The 13-bill stall isn't a permanent kill — these things come back next session — but the precedent that the most ambitious version of state-level drone restriction couldn't survive committee in a state actively trying to pass it is the political ceiling the next attempt will measure against. For drone operators worried about state-by-state regulatory fragmentation on top of the federal FCC fight, this is the first piece of news in a long time that lands on the right side of the ledger.
Earlier in May, drone leaker Igor Bogdanov (Quadro_News) posted a screenshot of a Chinese drone flight-reporting app showing the DJI Air 4 listed in the aircraft-selection dropdown. The categorization placed the model in the sub-250g class — the same regulatory category as the Mini series, well below the current Air 3S's ~724g all-up weight. Earlier prototype leaks had shown a redesigned upper sensor array (three upper obstacle avoidance sensors plus an integrated LED beacon) suggesting Mavic 4-style obstacle handling. Coverage carried into the May 18–24 window as the sub-250g implications got picked up by trade press. No FCC equipment authorization filing has yet appeared.
Our take: If the Air 4 ships sub-250g, DJI is engineering around the regulatory pathway, not just the FCC ban. Three things happen at sub-250g that don't happen above it: (1) US Part 107 recreational pilots get a friction-free experience — no registration, no Remote ID controversy in the form that's been litigated, simpler insurance posture. (2) The European registration regime drops from "A1/A2/A3 with documentation" to "A1 open category with minimal paperwork." (3) The aerial-systems framing the FCC has been using to justify Covered-List expansion becomes harder to apply — Covered-List arguments have leaned heavily on threat models that scale with airframe capability, and sub-250g toy-class framing is exactly the lane the Mini series has been wriggling out of for two consecutive product cycles. Ground-truth caveat: a sub-250g Air-line drone is engineering-hard. The current Air 3S carries a 1-inch sensor at 724g; getting that to sub-250g requires either dropping the sensor class (against the Air-line positioning) or unconventional materials and battery chemistry. If DJI pulls it off, the Air 4 is the most important product of 2026 — a flagship-tier consumer drone the FCC pathway can't easily touch. For the rolling picture, see our Air 4 hub and the 2026 product tracker.
On May 19, 2026, DJI released DJI Terra version 5.2.5. The headline change: automatic Cloud PPK retrieval of original base station data during LiDAR reconstruction tasks, which removes a manual import step that surveying operators have been griping about for two release cycles. Secondary additions: an optimized Multiple Time Around (MTA) algorithm for long-distance LiDAR point cloud reconstruction, plus bug fixes covering CSV coordinate accuracy after seven-parameter transformations, missing fields in LiDAR quality reports, and incomplete point cloud colorization when Smooth Point Cloud was enabled.
Our take: The enterprise hospice posture from the prior roundup continues unbroken. DJI can't sell new consumer product into the US, but it can keep the installed base of Terra-driven surveying and mapping workflows compounding — and enterprise customers, especially utilities, surveying firms, and public-sector users, are the hardest to displace because their workflows live downstream of DJI's specific point-cloud and reconstruction outputs. Skydio's military-adjacent procurement push (covered in trade press throughout the quarter) is the obvious threat, but Skydio's enterprise software stack is still meaningfully behind Terra on photogrammetry and LiDAR depth. Every Terra release that adds workflow automation widens that gap, and every quarter that DJI holds the enterprise install base is a quarter the FCC fight stays politically expensive because public-safety and utility commenters keep showing up in the docket. The firmware-waiver extension to January 2029 (covered in the prior roundup) only works if DJI is actually shipping updates worth caring about. Terra 5.2.5 is that.
In a mid-week May deployment that surfaced in May 22 trade coverage, Saskatchewan's Corman Park Police Service used a DJI Mavic 3 Enterprise thermal drone to locate an intoxicated man inside a roadside dumpster in –20°C weather. Officers detected the man's heat signature and rescued him before hypothermia became critical.
Our take: This is exactly the operator-story shape the FCC public-comment record has been accumulating for months. With 460+ comments now on the docket (closed May 11) and Autel's reply newly filed, public-safety rescue narratives are the political ammunition that softens a future commissioner ruling. The pattern is consistent: every week a search-and-rescue agency, utility, fire department, or local police force publicly attributes a saved life or averted disaster to a DJI Enterprise platform, the cost of the ban gets re-priced in the public record. DJI doesn't have to surface these — public-safety agencies do it for them, in their own jurisdictions, with their own credibility. The cumulative effect on the docket is the kind of soft pressure that doesn't move a commissioner who's already decided but gives political cover to a commissioner who hasn't. Saskatchewan is north of the border, so this one doesn't go in the US docket directly, but the trade-press coverage circulates among the same audiences that read the FCC commentary.
Four things on the radar for the week of May 25–31:
We'll be back next week. For real-time tracking between roundups, the 2026 product tracker and US availability hub stay current.