Upcoming Milestones
1107 announced milestones across every company we track - launches, certifications, first flights, deliveries - sorted by target date. The soonest 100 are below, and every one is listed by year.
All milestones by year/ 2026
Execute SDA satellite programs after passing TRKT3 SRR
Build and deliver the 18-spacecraft Tracking Layer Tranche 3 constellation for the Space Development Agency under Rocket Lab's $1.3B SDA prime work. The May 27, 2026 System Requirements Review remains the last publicly confirmed milestone; as of September 13, 2026 no Preliminary or Critical Design Review has been announced, with those the next gates ahead of spacecraft delivery. Mynaric CONDOR Mk3 laser links are produced in-house.
Sustain Electron cadence toward the 100th launch
With 96 Electron missions flown through the September 19, 2026 'Owl By The Dozen' dedicated Synspective StriX-12 launch (Rocket Lab's 17th launch of 2026), sustain Electron's record 2026 tempo across commercial, civil, and defense customers toward the company's 100th launch milestone. Remaining constellation bookings include iQPS (9 dedicated missions remaining on its total 18-launch Electron bookings, running through 2030) and Synspective (further missions booked to complete that constellation by decade's end). A HawkEye 360 Electron mission is listed no earlier than September 30, 2026.
Convert backlog into higher launch cadence
Turn a record $2.36B of backlog (as of Q2 2026) into higher Electron and HASTE flight tempo across constellation, agency, national-security, and hypersonic missions; Rocket Lab guided to another record revenue quarter in Q3 2026 and reported more than $1B in additional new contracts signed since the quarter closed. The new GHOST containerized launch system, unveiled August 10, 2026 with its first site at Kodiak, Alaska, is aimed directly at expanding cadence for responsive and national-security missions.
Build & launch the first GEO satellites for U.S. Space Force
Design, build, integrate, and operate two geostationary satellites hosting the Heimdall space domain awareness payload under a $90M Space Systems Command contract - Rocket Lab's first GEO satellite production program. No public schedule update (PDR/CDR/delivery date) has been announced since the May 21, 2026 award as of September 13, 2026.
Blue Ring first operational mission
Launch the first full Blue Ring mission, expected to become the first fully commercial GEO space-domain-awareness mission, carrying Optimum Technologies' Caracal sensor with Scout Space's Owl payload and internal payloads. Blue Origin has integrated the flight vehicle's primary structure, internal harnessing, and core propulsion module, but launch timing remains gated by New Glenn's return to flight after the LC-36 explosion; reverified September 12, 2026, no firmer date has been reported, though NASA's Sept 1 award of a Blue Ring-based Mars Telecommunications Orbiter (delivery by Dec 31, 2028) adds a second Blue Ring-derived flight in the same window.
Scale SciTec Golden Dome SBI work
Execute SciTec's Space-Based Interceptor OTA under U.S. Space Force Space Systems Command, building software, sensing, and AI for the Golden Dome space-based missile defense layer, with capability integrated into the architecture targeted by 2028. SciTec's other Space Force work (CBC2 data fusion and the $93.7M GBARD radar-digitization OTA) remains in execution; Firefly has posted no new Golden Dome contract update since the May 4, 2026 SBI announcement.
Ramp Blue Ghost and Elytra on the dual ISO 8 Cedar Park line
With the new ISO Class 8 cleanroom complete as of September 22, 2026, Firefly said the two Cedar Park cleanrooms can support up to 12 Blue Ghost landers and Elytra orbital vehicles at once. Watch whether that capacity shows up as a faster Blue Ghost Mission 2 integration cadence and more simultaneous CLPS and DIU spacecraft in work.
Return Vulcan to flight and fly the first Vulcan Amazon Leo mission
Return Vulcan to flight - grounded since the February 12, 2026 USSF-87 GEM 63XL solid-rocket-booster nozzle anomaly and the resulting U.S. Space Force pause on NSSL Vulcan launches. As of September 19, 2026 ULA had not posted a return-to-flight date; a spokesperson told Ars Technica on September 11 that the investigation is ongoing and that ULA will share the Vulcan Leo 1 mission launch date only upon completion of that investigation and coordination with Amazon. Independent trackers still list LV-01 as no earlier than late September 2026 (NET September 30, status TBD) from SLC-41 on a six-booster VC6L, carrying 45 satellites on the LEO-optimized Centaur V, which completed a wet dress rehearsal on August 6 and was stacked in VIF-G on August 15. A senior Space Force acquisition official confirmed on September 16, 2026 that Space Systems Command will review data from that Amazon Leo flight separately before restoring Vulcan to the national security manifest, so returning to flight commercially and regaining NSSL clearance are two distinct milestones.
Clear the Space Systems Command review and resume NSSL launches
Even after Vulcan flies its Amazon Leo return-to-flight mission, Space Systems Command must review data from that commercial launch, and be satisfied with the GEM 63XL root-cause investigation and corrective actions, before Vulcan carries national security payloads again. Until it does, SpaceX remains the only provider actively flying NSSL missions and ULA's roughly $5.3 billion Phase 3 Lane 2 backlog stays stalled. Ars Technica reported September 11 that the Space Force has already reassigned several Vulcan NSSL missions to SpaceX.
Bring SLC-3 at Vandenberg online for Vulcan
Finish the Vandenberg pad modifications (more than 83% complete as of the last public construction update on June 3, 2026, which targeted an operational pad in Q3 2026) and fly the first West Coast Vulcan mission - SDA's Tranche 1 Tracking Layer B. Re-verified September 10, 2026: no newer completion figure has been published, no West Coast Vulcan launch has flown, and the Q3 2026 pad-ready window has not yet fully passed. First West Coast flight still depends on Vulcan's return to flight.
Rebuild launch cadence once Vulcan returns to flight
ULA entered 2026 aiming for 18-22 launches (16-18 Vulcans plus 2-4 Atlas Vs), but the Vulcan grounding left it with just five launches through September 19, 2026 (one Vulcan and four Atlas V Amazon Leo flights), putting the full-year goal out of reach. Ars Technica reported September 11 that Vulcan prices have risen to more than $200 million per mission and that several NSSL missions have already been reassigned to SpaceX. The near-term focus is posting a return-to-flight date and rebuilding cadence into 2027 through the 38-launch Amazon Vulcan backlog and the NSSL Phase 3 manifest. ULA said it still has more than 80 launches on contract.
Complete stage hot-fire test campaigns
With all Flight 1 engine acceptance testing complete and the stage-one structural qualification campaign finished, Relativity's next test milestone is still stage-level hot-fire: the fully engined first-stage thrust structure must be mated to the rest of Stage 1 and shipped to NASA Stennis, while the second stage (already through cryoproof and COPV proof testing on Stennis's A-2 stand) has its Aeon V engine installed and returns to the stand. As of September 11, 2026 Relativity's public updates page has posted no campaign after the July 2026 company update, and no Stennis hot-fire has been announced.
Bring LC-16 fully online for Terran R
Finish outfitting the Horizontal Integration Facility, launch mount, and transporter-erector so Cape Canaveral operations are ready for first flight. The July 2026 company update said the HIF had been energized, a 270,000-pound assembly-room crane had completed its first proof-load test, and the transporter-erector upper strongback had arrived on site. As of September 11, 2026 Relativity has not posted a later pad-complete milestone.
First Haolong cargo-spaceplane resupply mission to Tiangong
Launch AVIC/Chengdu Aircraft Design Institute's reusable Haolong cargo spaceplane to the Tiangong space station on a Zhuque-3, one of two low-cost-cargo concepts China Manned Space Engineering Office selected in October 2024 (the other pairs CAS Space with a Microsat spacecraft). LandSpace has said it expects to begin Tiangong cargo missions in 2026, but no specific launch date has been officially announced as of September 20, 2026. The company newsroom's latest items remain the September 15 Zhuque-2E Y7 Qianfan flight and the August 19 Zhuque-3 landing recap, with no new Haolong schedule.
Convert the Shanghai STAR Market filing into a completed public listing, raising roughly 7.5 billion yuan. LandSpace refiled an updated prospectus on June 29, 2026, returning the review to 'inquiry' status, projecting profitability by 2029 for the first time, and disclosing full-year 2025 results (52.10 million yuan revenue, a 1.71 billion yuan net loss, and an accumulated deficit near 4.84 billion yuan). The share-issuance terms imply a listing valuation near 75 billion yuan (~$10.4B), well above the 20.7 billion yuan (~$2.9B) private valuation set in an April 2025 capital increase. Payload reported on September 4, 2026 that LandSpace is the farthest along of China's commercial rocket firms in the STAR Market queue after the August 18 Zhuque-3 landing, still racing CAS Space to become China's first publicly listed commercial-space company, with proceeds funding Zhuque-3 production and reusability work. No listing date has been set.
Scale Zhuque-2E commercial cadence
Sustain the Zhuque-2E launch tempo after Y7 on September 15, 2026 placed 10 Qianfan Polar Orbit Group 19 satellites into a roughly 900 km orbit - the ninth Zhuque-2-family flight and the first large-scale satellite-internet constellation deployment by a Chinese private rocket, ending a three-month gap after Y6 (June 9). LandSpace said Zhuque-2E is now in regular commercial service for constellation work while Zhuque-3 engineering reuse verification continues. As of September 20, 2026 the company newsroom still lists Y7 as the latest launch, with no Y8 or later Qianfan/Guowang batch announced.
Close the 'zeroing' fault investigation into the April 3, 2026 engine-bay failure and fly a second Tianlong-3. As of early September 2026 no public close-out or dated attempt has been issued; a second vehicle is still only reported as being prepared for a possible try before the end of 2026.
Advance the Tiangong Kaiwu space-resources program
Carry out feasibility studies and stand up an integrated, ground-supported experimental system for space resource development under the newly unveiled 'Tiangong Kaiwu' initiative, announced January 29, 2026 as a major project of China's 15th Five-Year Plan (2026-2030). The program targets small-celestial-body prospecting, intelligent autonomous mining, low-cost in-space transportation, and in-orbit processing, with a longer-term roadmap toward a lunar and near-Earth-asteroid resource system by 2035 and Mars/main-belt-asteroid development by 2050.
Having concluded on June 17, 2026 that the January 2026 maiden failure was caused by a design flaw in the first stage's nozzle throat liner, fix the deficiency and re-fly Ceres-2 to bring its larger solid-lift capacity online for constellation deployment. No return-to-flight date has been announced as of September 10, 2026, and the 2026 window is tightening.
Develop Pallas-2 heavy-lift and scale cadence
Advance the larger Pallas-2 (single-stick and CBC/tri-core configurations using the new CQ-90 engine, which completed a full-system hot-fire test in January 2026) while restoring Ceres-1/1S cadence after an eight-month pause. As of September 10, 2026, Galactic Energy has not publicly confirmed that Pallas-2 assembly and testing are complete; the maiden-flight target of before the end of 2026 still stands in the last company comments but looks increasingly tight with no new hardware updates since the Pallas-1 debut.
Demonstrate first-stage recovery and reuse
Move beyond reaching orbit to recovering and reflying the Hyperbola-3 first stage at sea - the core of iSpace's low-cost launch strategy - having completed full-profile ground verification of the landing system and June 2026 offshore sea trials of the Qinglan recovery drone ship. The debut mission itself is planned as the first sea-recovery attempt; a successful reflown booster would follow in 2027 if the late-2026 landing works.
Bring Chengdu production base online; advance Hyperbola-9
Put the record 5.037 billion yuan (~$729.6M) Series D++ closed in February 2026, plus the first tranche of a new Series E round (nearly 1 billion yuan, ~$140M, out of a round reported at ~7 billion yuan / ~$980M total) closed in August 2026, to work - completing the 3.3B yuan Chengdu production base (Chengdu reporting in April 2026 still targeted completion by end of 2026, equipment debugging late 2026, and official production in 2027, sized for 20 rockets/year), expanding Hyperbola-3 final assembly and integrated-test capacity, building out Focus-2 methalox engine and sea-recovery infrastructure, raising Hyperbola-1's reliability, and starting early R&D on the heavy-lift Hyperbola-9 and its 300-tonne-class Focus-3 engine.
Scale Kinetica-2 and deploy constellations
Ramp Kinetica-2 production at the new Zhejiang super factory (up to 12-20 vehicles per year) and fly its planned 2026-2027 manifest as the line enters serial production across Guangzhou and Shaoxing. The second flight (Y2), carrying 18 Qianfan (Thousand Sails) mega-constellation satellites, remains targeted for October 2026 after the vehicle departed the Guangzhou factory for Jiuquan on August 26, 2026 - LandSpace already flew its 10-satellite Qianfan demonstration batch on September 15. A third flight (another 18 Qianfan satellites) is targeted for December 2026 and a fourth mission for January 2027; the first operational Qingzhou cargo flight to the Tiangong station remains expected around Q4 2026.
Demonstrate Kinetica-2 first-stage reusability via Kinecore-2 engine
The in-house Kinecore-2 engine (110 tf) completed a 620-second cumulative long-duration qualification test - including a 400-second single burn, a new stability record - on June 30, 2026, en route to replacing the off-the-shelf YF-102 on Kinetica-2's first stage; no further Kinecore-2 test milestones have been reported as of Sept 20, 2026. CAS Space's phased reuse plan adds a full grid-fin system on Kinetica-2's third flight for aerodynamic reentry control, integrates Kinecore-2 capabilities (multiple engine firings, thrust control, cluster recovery) across flights four through seven, and targets complete first-stage recovery by the rocket's tenth flight around 2028. A roughly 100 km recovery demonstration of the Lihong-2 vehicle is now planned for Q1 2027 rather than later in 2026.
Hit monthly Kinetica-1 cadence, begin sea launches, and clear final STAR Market IPO registration
The July 24, 2026 monthly Kinetica-1 cadence commitment for the second half of 2026 has missed its August slot and, as of September 20, 2026, its September slot as well. Y17 completed a first sea-launch shipboard rehearsal on August 25 and remains in countdown for a three-satellite mission from waters near Guangdong; Y18 passed factory review July 31 for a nine-satellite Dongfeng land flight and is also unflown. Company news through September 17 covers a space-medicine partnership rather than a launch, and the homepage still lists 15 successful flights. A second sea launch is still described as targeting October. On the IPO, CAS Space's $607M STAR Market application still has no CSRC registration, pricing, or trading date as of September 20, 2026.
Develop Nebula-2 heavy-lift rocket
Advance the larger 5 m-diameter Nebula-2, designed to carry more than 25,000 kg to LEO, to broaden Deep Blue's reach into heavy constellation deployment. Its Thunder-RS first-stage engine passed a first 100-second full-duration static fire (1,297.5 kN / ~132.3 t) on August 21, 2026, and the company engine page now rates it at 1,300 kN sea-level / 1,500 kN vacuum. 2025 funding coverage said the first stage would cluster 11 engines. No maiden-flight date has been confirmed.
Resume Kuaizhou-1A cadence after a lengthening gap
Kuaizhou-1A/1A Pro has not flown since its 31st mission (the VDES maritime-navigation pair) on December 5, 2025 - now more than nine months of quiet for a vehicle marketed on hours-long, high-tempo launch readiness. The latest flown vehicle was a Kuaizhou-11. Watch for the next 1A mission to confirm the cadence claim still holds.
Fly the next Kuaizhou-11 after Y3
Kuaizhou-11 Y3 succeeded from Jiuquan on September 17, 2026 with Tianyi-51 and Tianyi-52. No follow-on Kuaizhou-11 mission, payload, or window has been posted as of September 18, 2026. A next flight would show whether the post-rebrand 11 cadence can hold after three 2026 missions (March, June, September).
Bring larger Kuaizhou-21 / -31 online
Advance the much larger solid Kuaizhou-21 (~20 t to LEO) and Kuaizhou-31 (~70 t to LEO) to move into medium- and heavy-lift constellation deployment. An earlier informal 2025 debut target for Kuaizhou-21 has passed with no confirmed maiden-flight date yet published. Wikipedia still lists a 2025 projected first flight that has not happened as of September 18, 2026.
Respond to reusable-rocket competition
Counter the mounting cost pressure from reusable rivals as CASIC's Mingfeng liquid-engine line and reusable-launch concepts remain earlier-stage, with no orbital reusable vehicle or maiden-flight date yet named. Industry reporting still flags Mingfeng-1's lack of a maiden flight as a potential IPO-review weakness. Kuaizhou Aerospace's all-solid, expendable lineup has not announced a comparable recovery demonstration.
Bring the Hangzhou plant online
Finish the 5.2 billion yuan (~$740M) Hangzhou Qiantang sea-recovery rocket factory, which broke ground January 7, 2026 and is designed for up to 25 medium-to-large reusable rockets per year and a launch-cost target below 20,000 yuan/kg. The plant was planned to begin 'Qiantang-hao' final assembly by the end of August 2026; that start has not been publicly confirmed.
Demonstrate tower-catch recovery
Validate the 'chopstick' tower-catch and intelligent-control recovery architecture at scale - the technical centerpiece that distinguishes Cosmoleap from China's leg-landing reusable rivals - building on completed drop tests onto the catch arms, auxiliary-propulsion thruster hot fires, the June 2026 first-stage propellant-precooling test, and the August 2026 second-stage structural static test. As of early September 2026 the company has not published a later catch-tower demonstration.
Complete Qingyu-11 engine development
Mature the in-house 150-ton-class Qingyu-11 methalox engine (30-105% throttle, up to 50 reuses) to replace the externally sourced YF-209 on later vehicles. Core-component tests (torch ignition, low-temperature and liquid-flow testing) finished in June 2026; full-system test firing is still the next sourced gate and has not been reported complete as of early September 2026.
Stand up the Thoothukudi integration facility
The August 13, 2026 MoU with Tamil Nadu commits Rs 250 crore (~$26 million) and about 500 jobs to a storage, assembly, integration and testing site at Thoothukudi, next to the Kulasekarapattinam spaceport. Chandana said the plant is to support high-inclination southward flights from that port. It is still an MoU; no construction start or opening date has been published.
Reach first-module manufacturing and fuel readiness
NuScale says its main 2026 priority is ensuring full readiness to manufacture the first NuScale Power Module while scaling fuel and supply-chain support with partners such as Framatome. Components for 12 power modules are already in production; in June 2026 NuScale contracted Paragon Energy Solutions to complete final design of the module's safety-related instrumentation-and-control systems. On September 1, 2026 NuScale and MillenniTEK announced they had fabricated first-of-a-kind boron-oxide pellets for the module's passive emergency core cooling system, a concrete manufacturing-readiness step. The first physical module is still not complete.
Ramp Groves to full power and start isotope production
With first criticality achieved on August 5, 2026, Oklo's next step is low-power testing and ascension to full power at the Groves Isotope Test Reactor in Lockhart, Texas, ramping toward R&D-scale isotope production over the following year. On its August 7, 2026 earnings call, management said it now expects Oklo's first isotope revenue in early 2027, and that it is more likely to come from the NRC-licensed Idaho Radiochemistry Laboratory than from Groves. Oklo's newsroom and isotopes page still describe Groves as having reached criticality in under a year, with no later public update on power ascension as of September 19, 2026.
Carry Aurora-INL through final safety review and construction
With the DOE's Preliminary Documented Safety Analysis approved on June 11, 2026 - the third of four major Reactor Pilot Program authorization steps - and a November 2025 binding Siemens Energy contract closing out major long-lead power-conversion equipment procurement, Oklo's near-term focus is securing the final Documented Safety Analysis, the last safety-basis document before DOE authorizes operation, while completing construction at Idaho National Laboratory ahead of fuel loading and first power. As of September 19, 2026, the final Documented Safety Analysis has not yet been approved; on its August 7, 2026 earnings call Oklo raised its full-year 2026 capital-expenditure guidance to $400-500 million to accelerate Aurora-INL procurement and grid-interconnection work while reaffirming a 2028 commercial start-up target.
Produce HALEU TRISO fuel for Hermes 1
Using DOE-allocated HALEU and Los Alamos National Laboratory's LEFFF capability to manufacture the first fuel pebbles needed for Hermes 1 startup and operation. As of late August 2026 Kairos has finished installing process equipment at LEFFF and is commissioning the pilot line with depleted uranium before introducing HALEU; no fuel pebbles for Hermes 1 have been reported produced yet. Company updates through September 10, 2026 (ETU 2 modular integration) did not announce a first HALEU pebble; the June 24 pebble campaign was surrogate (non-nuclear) fuel for ETU 2.
Complete ETU 2 integration and hot argon testing
Finish module installation and piping/electrical connections on Engineering Test Unit 2's more than 30 factory-built skids in Albuquerque, then run hot argon testing (valves, inert gas, temperatures and pressures) before heating the Flibe salt system. Kairos described ETU 2 as the world's largest planned Flibe molten-salt test once operations start later in 2026, and said the modular path will feed Hermes 2 construction.
Convert pipeline and supply-chain reservations into projects
Having locked in a binding 16-unit Doosan Enerbility reservation, an expanded SGL Carbon graphite supply deal, an IHI Corporation MOU, and a definitive LEU/HALEU enrichment agreement with Centrus Energy (deliveries starting 2030), X-energy still needs to turn its 11.5 GW (144-unit) pipeline - including gigawatt-scale evaluations with Talen Energy and LG&E/KU - into named, financeable deployments. Q2 2026 remains the latest reported quarter ($54.6M revenue and grant income, $1.9B liquidity, up to $1B more ARDP). On the August 13 earnings call CEO Clay Sell said a roughly 1 GW investor-owned-utility agreement was in its final stages; as of September 18, 2026 that counterparty has not been named. Energy Northwest is still aiming to file a Cascade construction-permit application in the first half of 2027.
Palisades nuclear plant restart
Holtec has closed out the last of Palisades' major restart construction projects and, on August 30, 2026, began loading fuel into the reactor core - 204 assemblies of new and partially used fuel - marking the most visible physical step yet toward restart. Engineering News-Record reported on September 18, 2026 that the 805 MW plant was still in Mode 6 (refueling) with no announced grid-reconnect date. An August 31 NRC inspection report then cited two green, non-cited radiation-protection violations from December 2025-March 2026 work, including a February airborne-radioactivity event in which required notifications and dose evaluations were not completed; the NRC assigned very low safety significance and issued no penalties. A federal court dismissed the original anti-restart lawsuit on July 3, 2026, but the environmental coalition (Beyond Nuclear, Don't Waste Michigan and allies) appealed that dismissal to two federal circuit courts of appeals in late August 2026, keeping a legal challenge alive without blocking restart activity so far. Extensive NRC-coordinated tests and inspections remain, after which the unit must be returned to criticality and reconnected to the grid. Completion has slipped from late 2025 and then from February 2026; CEO Kris Singh still expects a 2026 restart, ahead of a March 2027 power-supply contract, but Holtec has declined to set a firm date. Success would make Palisades the first U.S. nuclear plant to return to service after entering decommissioning.
Convert the US-Saudi nuclear pact into an AP1000 order
Following the July 23, 2026 US-Saudi peaceful nuclear cooperation agreement, the deal remains in diplomatic limbo: one day after the signing, President Trump conditioned approval on Saudi Arabia joining the Abraham Accords to normalize relations with Israel, and Saudi Arabia has so far resisted signing, insisting on progress toward Palestinian statehood first. Congressional review was still pending as of late August 2026. A potential multi-decade, multi-billion-dollar Saudi AP1000 program with Westinghouse as likely builder cannot proceed until these preconditions are resolved. Re-verified September 2026, no change.
Aalo is reportedly negotiating a Series C of roughly $500M, building on an existing SAFE with a $3B valuation cap, expected to price around a $4-5B valuation; proceeds are earmarked chiefly for the Gigawatt Factory. Still not closed as of September 15, 2026 - no company announcement, the official site continues to cite $300M+ raised, and Nasdaq Private Market in late July still described Aalo as in the process of raising that round.
Stand up mass-manufacturing reactor factories and in-house fuel production
Valar says its $1B Series B is earmarked to build reactor-manufacturing factories that scale output from single units toward 'tens, then hundreds, then thousands' of reactors per year, and to stand up in-house TRISO fuel production tied to its DOE Fuel Line Pilot Program selection - vertically integrating the hardware supply chain from deployment through long-term operations and fuel. The first piece of that fuel buildout is already under construction: a co-located Valarin Fuel Fabrication Facility next to Ward250 at the Utah site, intended to manufacture TRISO-coated particle fuel in graphite compacts using German HOBEG-derived technology. On September 18, 2026 Utah Trust Lands approved a Carbon County Special Use Lease covering up to 640 acres (starting at 480 acres for 50 years near Wellington and Price) as Valar's first named expansion site beyond the Emery County test lab; construction still requires development plans and local, state, and federal permits, and no factory opening date has been announced.
Complete SPARC assembly and commissioning
Close the external cryostat, integrate fuel and diagnostic lines, and complete final commissioning of the integrated tokamak, cryogenic, RF heating, and magnet systems at the Devens campus. The cryogenic plant cleared its own commissioning bar on September 16, 2026, running the full system at the 8 kelvin operating temperature the TF magnets require.
Run SPARC's dry dress rehearsal
Operate the entire SPARC plant - cryogenics, power, RF heating and vacuum systems - under full pulse loads as if a fusion pulse were happening, but without the tokamak itself, to validate the support infrastructure before first plasma. CFS described the September 16, 2026 8 kelvin cryogenic result as commissioning work leading into this integrated test; no date has been published for the rehearsal itself.
Move Fall Line Fusion Power Station through PJM interconnection study
Local land-use approval is already secured (Chesterfield County granted final zoning/conditional-use approval in September 2025). The remaining long-lead item is PJM Interconnection's multi-year (roughly 4-6 year) engineering study and approval process for the April 2026 grid-interconnection request, alongside locking down the final ARC plant design ahead of construction later this decade.
Scale the HTS magnet supply business
Execute the April 2026 Realta Fusion long-term HTS magnet supply agreement and turn the Type One Energy licensing deal, WHAM magnet supply, and Realta partnership into a recurring magnet business that funds the fusion roadmap before SPARC and ARC reach revenue.
Raise Polaris repetition rate toward 1 Hz
Push Polaris from one pulse every ~10 minutes toward sustained 1 Hz operation - the rate Helion needs to convert fusion energy into recapturable electricity in commercial operation. A separate pulsed-power testbed, Vela, already operated at 1 Hz and moved more than 11 GJ through a simulated Polaris compression magnet, but that 1 Hz result has not been demonstrated on Polaris itself as of September 16, 2026.
Finalise the OpenAI gigawatt-scale supply deal
Move the OpenAI negotiations from term sheet to a binding agreement - reportedly 5 GW by 2030 scaling to 50 GW by 2035 - and identify the sites and Orion-class plants needed to deliver that capacity. Talks remained ongoing and unsigned as of September 16, 2026, with site selection among the unresolved terms. Helion's newsroom still lists Microsoft and Nucor as its only announced customers. The arrangement has drawn Congressional and courtroom scrutiny over Sam Altman's personal stake in Helion and his recusal from the talks.
Pick the first plant site and start construction
Convert the April 7, 2026 completion of the Alabama/Ohio/Texas site evaluation tour into a final site choice and start construction of the initial ~50 MWe Da Vinci facility. TAE's August 6, 2026 update said key designs were complete and construction was still planned later in 2026, contingent on the TMTG merger close (still targeted by end of 2026) and approvals. Re-verified September 18, 2026: no final site had been announced, and TAE's news page still lists the April tour as the latest siting update.
Upgrade Norm to 100 million °C
Push the current experimental platform from 70+ million °C - already the highest temperature recorded on a steady-state FRC - toward 100 million °C to validate operating modes and hardware for TAE's next-generation reactors. Re-verified September 18, 2026 on TAE's FAQ: the upgrade remains described as planned/in progress, with no announcement that 100 million °C has been reached.
Deliver the first STEP magnet work packages
Execute the eight magnet work packages under the £70M STEP contract with UK Fusion Energy, covering magnets, tokamak systems and plasma integration, and use ST40 as the test bed for magnetic configurations feeding STEP's later design phases. After the company said it will not construct its own pilot plants, STEP is the flagship construction programme Tokamak Energy is attached to. The September 9, 2026 Demo4 campaign close is already feeding STEP model-coil testing.
Stand up UK HTS tape manufacturing with Furukawa Electric
Turn the June 2026 Furukawa Electric collaboration into domestic UK production of high-temperature superconducting tape, the single largest supply constraint on HTS magnet programmes - STEP's magnet system alone is expected to need around 30,000 km of it. US-based production at Furukawa subsidiary SuperPower Inc in New York has already delivered first tape batches; UK-based manufacturing capability is still being evaluated.
Advance the UK Infinity Fusion Consortium plant project
Move the consortium formed in May 2026 with Type One Energy and AECOM toward the first private-sector-led commercial fusion power plant project in the UK. After Tokamak Energy said it will not construct future pilot plants itself, its likeliest role is as HTS magnet and systems partner rather than plant owner-builder.
Scale HTS cable and magnet production
Use the July 2026 round to expand high-temperature superconducting cable and magnet manufacturing and the engineering systems behind stellarator production across the Munich, Paul Scherrer Institute and Culham sites, and hire across engineering, manufacturing and operations. Convert the September 9 Lower Saxony HTS-tape MoU into a named site, a technology partner and a funded plant (~€140M first two phases through 2029), and advance the planned Gundremmingen magnet factory that has not broken ground.
Convert the Alpha MoU into binding financing and permits
Turn the February 2026 memorandum of understanding with RWE, the Free State of Bavaria and Max Planck IPP into firm agreements: roughly 20% private financing, roughly 20% Bavarian co-financing against the up-to-€400M High-Tech Agenda, federal funding, and the permitting needed to build at Garching. The July 2026 €411M round met Bavaria's private-match condition, unlocking its €400M pledge; a further ~€1.2B in German federal funding for Alpha remains uncommitted. The Garching plot is still unchosen. STRIDE is a separate, smaller, three-hub research channel (initial €125M for all three hubs equally), not Alpha construction money.
Test a full-scale production pulser module
Complete and test full-scale pulser modules, each roughly 40x the size of Sirius and 3x the June 2026 scaled prototype, to prove the unit that will be manufactured 156 times over for the Demonstration System. Nature reporters visiting the Los Lunas build center in August 2026 found two of the more than 150 full-scale modules already completed. Re-verified as of September 16, 2026: company updates through September 15 still describe demonstrating a full-scale pulser module as the next major technical checkpoint; no public full-scale module test result has been posted.
Build the 156-module Demonstration System
Manufacture and assemble all 156 production-scale pulser modules at the New Mexico campus, supported by component production at the Los Lunas build center, and integrate them with the target chamber and diagnostics. Ground was broken August 25, 2026; Users Program FAQ still dates commissioning to 2028.
Bed in the Spirit AeroSystems integration
With the December 2025 acquisition closed and ~15,000 Spirit employees folded into Boeing, the focus shifts to translating it into measurable quality gains - Boeing reports a 45% drop in 737 fuselage defects since 2024 - and stabilising rail deliveries of 737 fuselages from Wichita, which now builds more than 70% of the 737 MAX airframe, as the line ramps to 47/month and the new Everett North Line opens (July 6, 2026). Boeing backed the integration with a $2.35B Wichita reinvestment (announced May-June 2026) and, with Wichita State University's WSU Tech, a roughly 35,000-square-foot Boeing Workforce Training Center for aerospace technicians expected to open by the end of 2026.
Keep executing the safety and quality plan
Boeing's recovery still depends on sustaining its Safety Management System rollout, factory-level quality controls, and stronger safety governance across Commercial Airplanes as production rates climb. The FAA formally moved from a hard numerical production cap (the 38/month ceiling set after the January 2024 door-plug incident) to a performance-based oversight model in March 2026, delegating more individual airworthiness ('ticketing') authority back to Boeing's own authorized representatives as trust rebuilds. Passing the FAA's capstone review to run the 737 line at rate 47 in May 2026 was an explicit validation of those quality gates; lifting further toward rate 52 (and ultimately 63) remains contingent on holding defect rates down as Spirit AeroSystems is folded back in.
Advance the Boeing-Anduril IFPC Increment 2 interceptor through Phase 2
After the September 9, 2026 Army downselect, Boeing and Anduril move into detailed design, system development, integration and a competitive capability demonstration for the IFPC Increment 2 second interceptor. The design uses an Anduril solid rocket motor. A later production award would put the team into the Army's midrange air-defense interceptor magazine.
Convert Vietnam Airlines MoU and Vietravel LOI into firm orders
On September 10, 2026 in Paris, Vietnam Airlines signed a memorandum of understanding for five A350-900s with deliveries planned for 2033-2034, part of a stated priority to acquire 30 widebodies beyond 2030. Vietravel Airlines signed a letter of intent for 20 A220s and 30 A321-family jets, with deliveries expected from 2029. Neither instrument is a firm order. Watch whether purchase contracts follow, which is what would add the 55 aircraft to Airbus's backlog rather than leaving them as diplomatic paper.
Close the 2026 C919 delivery gap
COMAC delivered 10 C919s through September 18, 2026 - 8 in the first half plus 2 in July (Air China's 12th and China Southern's B-659G on July 21) and none publicly recorded in August or the first half of September - against the 33-aircraft total Chinese airlines have planned for the year, leaving it on track to miss the annual goal by a wide margin. Cumulative deliveries since the 2023 entry into service now stand at about 42. IBA's September 2026 forecast still expects roughly 28 C919s for the full year, rising toward 59 in 2028 and 93 by 2030. The near-term challenge is unblocking LEAP-1C engine supply, easing labor shortages, and lifting monthly throughput after the August-September pause.
Qualify licensed F100-PW-229 production
After the September 15, 2026 Pratt & Whitney Production License Agreement, Hermeus must qualify Pratt-standard F100-PW-229 builds before any demand-driven scale-up. CEO Zach Shore said initial manufacturing will use existing footprints and that Hermeus is becoming a second U.S. source for the engine that powers Quarterhorse and some F-15/F-16 variants, not a clean-sheet engine house. No production rate or start date has been disclosed.
Execute DIU $219M high-Mach payload-release demos
The May 2026 $159M DIU contract modification - taking the program ceiling to $219M - commits Hermeus to a series of intensive 2026-2027 flight tests with the USAF and USN proving uncrewed Quarterhorse can hold high-Mach speeds while carrying and releasing payloads, the technical foundation for a Darkhorse-class operational system. Hermeus cleared the first concrete step in late July 2026, when Quarterhorse Mk 2.1's fifth flight completed an unmanned stores-separation test under DIU's HyCAT program, safely releasing an external payload; the campaign continues toward higher-Mach payload releases through 2027.
Fly Mk 2.2 and bring Mk 2.3 online
The FAA granted Mk 2.2 a special flight authorization published September 11, 2026 for up to six supersonic flights over White Sands through August 20, 2027. Hermeus's aircraft page lists Mk 2.2 for 2026 flight testing with a spike inlet, Gryphon engine and precooler above Mach 2, but a first-flight date has not been announced. Mk 2.3 is in active build for 2027 flight testing with high-temperature materials near Mach 3.
El Segundo HQ ramp and Atlanta production scale-up
The ~62,550 sq ft El Segundo headquarters at 888 North Douglas (a former Northrop Grumman campus) is expected to add more than 200 jobs over the next year as executive, design, and prototyping teams consolidate in California, with full occupancy expected in early 2027; the 175-person Atlanta facility is shifting toward production-rate assembly work for the growing Quarterhorse fleet. Careers still lists El Segundo as headquarters, with a Hawthorne mailing address at 3401 Jack Northrop Ave.
HEAT engine test facility completion
The High Enthalpy Air-breathing Test (HEAT) facility at Cecil Airport in Jacksonville - a ~$135M investment expected to create 100+ jobs over the decade - is being built in phases toward continuous high-Mach vitiated-air ground testing of the F100 and Chimera engines. The first phase came online with sea-level static F100 tests in January 2025; as of September 2026 the site is also running Ramjet-X engine tests. The site is also intended to base high-Mach Quarterhorse flight-test operations into the Eastern Range.
HX-2 hybrid-electric flight demonstration
With the all-electric X1 having completed its piloted maiden flight on August 12, 2026, Heart's next major flight-test milestone is a hybrid-electric demonstration originally described as the X2 prototype. Heart's August 13 first-flight release and current ES-30 page no longer mention a 2026 X2 flight; they say Heart is developing the first pre-production ES-30 at its Los Angeles Pilot Plant, with flight testing scheduled to begin in 2028. Key.aero reported on August 19 that the originally planned X2 is now referred to as that first pre-production ES-30. Wikipedia still lists a 2026 hybrid-electric demonstration target. Heart's newsroom has posted nothing after the August 13 first-flight release as of September 19, 2026, and the company has not announced an X2 flight.
Push BotQ from 1/hour toward 50,000/year
Figure hit a 24x throughput jump in early 2026 to one Figure 03 per hour and crossed 1,000 units produced on July 23, 2026, sustaining that pace over several months. The next phase is holding the rate as volumes climb further, finishing the publicly-stated supply-chain decoupling from China, and scaling toward Adcock's stated 50,000-unit annual target on the way to 100,000 robots across four years. Figure's news page listed no newer factory-count disclosure as of September 19, 2026.
Execute the Catalyst Brands deployment
Stand up the Reno, Nevada Figure 03 fleet on the Joey Pouch sorting/packing system, prove throughput and uptime on a live retail-logistics floor, and use the program as the template for broader Catalyst Brands and other retail-logistics deployments. As of September 19, 2026, neither company has disclosed the robot count or operating scale of the rollout, so it still reads as an early implementation phase rather than standing production.
Scale Helix-02 from demos to standing production
After the 8-hour, extended 40-hour, and 200-hour (~250,000-package, zero-failure) autonomous package-sorting demos in May 2026, the next milestone is keeping Helix 02 in continuous Catalyst-style commercial production rather than one-off marathons, while extending whole-body autonomy across additional logistics, office, and home tasks. Helix 02 remains the stack Figure showed on BMW Spartanburg logistics sequencing; Helix 2.5 (September 17, 2026) is the newer home-task model. No independent reporting as of September 19, 2026 confirms continuous paying production has replaced the demo cadence.
Broaden home-task autonomy on Figure 03
After dishwasher and living-room demos, keep expanding Helix across more rooms, tools, and multistep household routines so Figure 03 can move from staged home tasks to real residential pilots inside the Brookfield-supplied housing footprint. Project Go-Big has already produced zero-shot, language-conditioned home navigation from egocentric human video. Index, launched August 25, 2026, is the main new data pipe, generating about 35 minutes of human-experience data every second. The September 17, 2026 Helix 2.5 release is the biggest step so far: Figure 03 tidied living rooms, folded towels and made beds in 30 unseen Bay Area homes with no data collected or fine-tuning in any of them, succeeding on 56% of whole-task trials. Figure has still not announced a paying residential pilot, so the open question is whether that 56% climbs to service-grade reliability in customer homes.
Push Gemini Robotics foundation models into Atlas
Extend the Boston Dynamics-Google DeepMind partnership from Spot (where Gemini Robotics ER 1.6 now powers AIVI-Learning) into Atlas, integrating Gemini Robotics foundation models for broader humanoid task learning beyond reinforcement learning policies. Google DeepMind shipped Gemini Robotics 2 on July 30, 2026 with true whole-body control (legs, torso, arms, and five-fingered hands under one learned policy) - but the public demonstration platform was Apptronik's Apollo 2, not Atlas. As of September 12, 2026 an Atlas-specific integration milestone still has not shipped. Claims of an already-shipped 'Gemini 3-powered Atlas' continue to circulate in low-quality financial-content aggregators but remain uncorroborated by Boston Dynamics, TechCrunch, The Robot Report, or Google DeepMind's own blog, and are not reflected here.
Conclude the board's CEO search to replace interim CEO Amanda McMaster, who took over from retired founder/CEO Robert Playter on February 27, 2026. In a September 2026 Machina Summit interview she was still titled interim CEO and CFO, said Spot remains the currently adopted revenue product, and said the company would not rush Atlas commercialization. No successor has been named.
Expand Spot with FieldAI in dynamic sites
Build on the March 2026 FieldAI partnership: expand Spot deployments across construction and other changing environments using Field Foundation Models for autonomous inspections, mapping, and monitoring in less structured sites, with Stretch and the electric Atlas named as the next platforms to integrate FieldAI. A February 2026 Boston Dynamics case study reported that construction customers using FieldAI-equipped Spot deployments across Asia, Europe, and North America cut inspection and documentation time by more than 90%. As of September 12, 2026 no new deployment figures or Stretch/Atlas FieldAI integration have been announced.
Scale the DHL and Lidl Stretch programs
Convert DHL's 1,000-unit MOU (plus its separate $15M North American automation investment) and Lidl's Netherlands/Belgium/Austria/Spain rollout into delivered fleets, while widening Stretch's case-picking capabilities beyond container/trailer unloading. Lidl's December 2025 plan called for integrating Stretch into regular import-warehouse operations in the Netherlands and Belgium by May 2026, and into Austria and Spain as a standard solution by mid-2026. As of September 12, 2026 neither Lidl nor Boston Dynamics has published confirmation that those mid-2026 country rollouts are complete, and no new DHL or Lidl unit commitment has been reported.
Broaden Hyundai deployment programs and build the Atlas factory
Turn Hyundai's $26B U.S. investment plan - which funds a robotics factory targeting 30,000 Atlas units/year by 2028 and Atlas deployment at Hyundai's Georgia Metaplant (sequencing first from 2028, assembly by 2030) - into broader Spot, Stretch, and Atlas deployment across manufacturing and logistics programs. At Hyundai Motor's August 26, 2026 CEO Investor Day, the company reaffirmed 2028 U.S. robot production at 30,000 units a year, said an initial committed volume of 25,000 units is already confirmed, and said the factory site will be announced by year-end, with Atlas first deployed at Hyundai Motor Group Metaplant America from 2028. The 25,000+ unit Hyundai/Kia deployment plan drew escalating labor pushback in Korea through most of August 2026: after strikes in July and a fourth round of partial stoppages running August 12-16 pushed cumulative Hyundai Motor production losses to an estimated 42,510+ vehicles and roughly $1.3B in lost revenue, Hyundai's union staged a full-day walkout on August 21 (its first in a decade) before a 17th round of negotiations produced a tentative wage agreement on August 24-25: a 100,000-won (~$72) monthly base increase, a performance bonus of more than 400% of monthly salary, a separate 12.5-million-won lump-sum payment, an extended retirement age, and a commitment to hire 500 new technical workers in 2027-2028. Union members ratified the deal on August 31, 2026 with 61.55% approval (78.63% turnout), ending the 117-day dispute; Hyundai has maintained throughout that the 2028 Atlas plan is not formally part of the wage talks.
Land a first named logistics partner for Spot last-mile delivery
Convert the July 14, 2026 'Bridging the Porch Gap' concept reveal - Spot riding in a delivery van, then walking parcels the final ~50 feet from curb to doorstep using a conveyor payload that carries two packages at a time (Boston Dynamics estimates that covers at least 60% of an average delivery van's load) - into a named pilot. Boston Dynamics said it was in talks with major logistics companies toward a full pilot in which Spot would work alongside a driver to deliver roughly 200 packages a day, five days a week. As of September 12, 2026 the company still has not named a partner, city, or start date.
Resolve the ICE Spot procurement into a signed contract
Follow ICE's Office of Firearms and Tactical Programs' August 27, 2026 sole-source procurement notice for an estimated $1-2M of Spot robots and accessories (place of performance: Fort Benning, Georgia) through to a formal solicitation and, if pursued, a signed contract. The DHS forecast listed an expected solicitation date of September 4, 2026, but as of a September 9 Boston.com report the government had still not issued an RFP. ICE did post a related Sources Sought notice for a quadruped unmanned ground vehicle (Notice ID 192126NFU00000158) around September 2, with responses due that day and the listing marked inactive on September 4; that was a capability survey, not the Spot RFP. Volume, delivery timeline, and whether a noncompetitive award survives scrutiny remain unconfirmed; Boston Dynamics has restated that any attempted weaponization is prohibited under its terms.
Convert Dynamic Creatures into a named guest-facing launch
Dynamic Creatures emerged from stealth on September 8, 2026 as Boston Dynamics' official hospitality and entertainment partner, already working with an unnamed large theme park and a retailer on mobile character robots. Interim CEO Amanda McMaster called it a new frontier for the company's platforms. The next milestone is a named venue, a public guest-facing start date, and whether the characters run on Spot, a custom body, or both.
Land a first named customer using Spot 5.2 agentic dispatch
Spot 5.2, released in September 2026, re-architected Orbit around assets rather than inspections, opened the API to third-party PLC sensors and cameras, and described that connectivity as the foundation for an MCP layer that lets AI agents dispatch Spot when business logic fires, for example sending the robot to a pump after a fixed sensor alerts. Gemini-powered AIVI now covers gauges, sight glasses, pallets, 5S boards and video inspections, plus gas sensing, visual vibration analysis and partial-discharge checks. Boston Dynamics has not named a customer running that agentic loop in production. Watch for a first disclosed site that actually lets an enterprise system send Spot rather than a scheduled mission.
Win Texas governor approval and pick the Crystal Sun site
Lamar CISD voted 7-0 on September 15, 2026 to advance Tesla's $10.1 billion Project Crystal Sun JETI application for a Fort Bend County solar-cell and module factory. Construction is targeted for 2026-2028 with commercial operations in Q1 2029 if Texas wins. Watch whether the governor signs on, whether Tesla picks Fort Bend over the unnamed out-of-state alternative, and whether the 10-year school-district tax limitation actually closes.
Open Tesla Motors Vietnam into commercial auto sales
Tesla Motors Vietnam LLC was registered September 11, 2026 in Ho Chi Minh City with about $3 million of charter capital for auto wholesale, retail, import and distribution. Tesla has not commented and has not set a sales-start date, showroom plan, or model lineup. The watch item is whether the entity becomes a live importer and retailer versus remaining a paper registration.
Grow FSD and software monetization
Active FSD subscriptions reached 1.48M in Q2 2026, up 56% year over year and the largest quarterly add in Tesla's history, generating roughly $791M of annualized software revenue with over 55% of new North American deliveries attaching FSD at purchase. FSD (Supervised) is now cleared in six EU markets after Slovenia approved the system on September 8, 2026 (Netherlands, Lithuania, Estonia, Denmark, Belgium, Slovenia); Tesla Europe said the Slovenian rollout will begin soon. Italy has filed to adopt the Dutch approval, while Germany, France, Spain, and the Czech Republic are holding out for a coordinated EU-wide decision rather than a national clearance Brussels could later overturn. Germany's KBA is not relying solely on the Dutch paperwork and received a dedicated Tesla vehicle in July 2026 for its own independent evaluation, with the Autobahn speed offset a specific sticking point. The next qualified-majority TCMV vote is still preliminarily expected around October 6, 2026, though industry trackers warn continued disagreement over specifics like the speed offset could push a final decision into late 2026 or 2027; separately, reporting around August 6 indicated Dutch approval authority RDW is withholding the underlying FSD safety and evaluation documentation from public view at Tesla's request, adding controversy ahead of the vote. In the United States, Representative Raja Krishnamoorthi wrote Transportation Secretary Sean Duffy on September 9 asking what NHTSA is doing about Autopilot and FSD after 43 publicly posted sleeping-driver videos, citing engineering analysis EA26002 on about 3.2 million vehicles and setting a September 30 reply deadline. The next step is sustaining subscription growth, winning that EU-wide vote, answering the US safety questions, and turning the same AI stack into Robotaxi revenue.
Scale manufacturing capacity toward 20,000 humanoids in 2026
Near-term target is 20,000 humanoid robots shipped in 2026 (vs 5,500 in 2025) - a nearly 4x increase (some September coverage around Unitree's IFA Berlin showcase instead frames it as a 10,000-20,000 range, but the company's own guidance remains 20,000). Medium-term goal over five years is 75,000 humanoids and 115,000 quadrupeds annually (190,000 total). IPO proceeds are earmarked for a smart manufacturing base to replace the current manual/outsourced assembly model and achieve modular, flexible mass production at scale. H1 2026 shipments of roughly 5,900 units imply Unitree needs a sharp second-half acceleration to reach the 20,000-unit full-year target; no updated shipment tally has been published since mid-August.
Broaden the embodied-AI product stack and navigate Pentagon listing, FCC Covered List, and the GUARD Act
Continue expanding from finished robots into embodied models, dexterous hands, robotic arms, sensors, and now the GD01 manned mecha unveiled in May 2026. In late July 2026 Unitree shipped UnifoLM-OmniA-0.3, a unified speech/vision/navigation/manipulation AI model aimed at home-care and wellness tasks. On September 7, 2026 it posted UnifoLM-X2-1.0, claiming the first fully autonomous humanoid combat driven in real time by a world model without teleoperation. The Pentagon's June 8, 2026 Section 1260H listing restricts direct DoD contracting with Unitree from June 30, 2026 and indirect procurement from June 2027. On July 28, 2026 the FCC went further and added foreign humanoid and quadruped robots to its Covered List directly, ahead of the bipartisan GUARD Act (H.R. 9129, introduced June 3, 2026) that had proposed the same step through legislation - new Unitree equipment authorizations in the US now require a Department of War Conditional Approval, while domestic China and non-U.S. markets remain unaffected. A separate bill, the Blocking CCP Spy Tech Act of 2026 (S.4586, introduced by Sens. Rick Scott and Tom Cotton around May 20-21, 2026), also names Unitree among six Chinese tech firms for a national-security investigation that could feed further Covered List additions. As of early September 2026 Unitree had still issued no public legal challenge to any of these measures (re-verified; unlike Alibaba, which sued the Pentagon over its own 1260H listing), even as Beijing imposed broad retaliatory trade measures over the U.S. designations. Adding to the scrutiny, researcher Olivier Laflamme's August 2026 disclosure of two wormable root-RCE flaws in the G1 EDU (CVE-2026-76639, CVE-2026-76640, both since patched) gives critics of Chinese-made robots a concrete security incident to point to alongside the geopolitical designations.
Sell G1+ outside China and clear any new-model US authorization
G1+ launched September 14, 2026 as a new hardware SKU, listed at 95,000 yuan tax-in in China and US $15,000 excluding tax and shipping. Previously authorized G1 units remain sellable from US inventory, but a new model generally needs a Department of War Conditional Approval before FCC equipment authorization. EDU pricing is still unannounced. Watch China and Europe shipments and whether Unitree files for US authorization.
Scale Apollo production through Jabil
Apptronik's manufacturing route runs through Jabil rather than through factories of its own, under the April 2025 collaboration that also puts Apollo to work inside Jabil's own operations. Neither company has published 2026 production volumes. Proving that a contract manufacturer can build humanoids at rate is the supply-side half of the 2027 plan, and it is the part Apptronik controls least directly.
Expand the Robot Park network to customer and partner sites
The Austin Robot Park is the flagship of what Apptronik describes as a growing network of Robot Parks at customer and partner locations, including data-collection workflows at Google DeepMind, Mercedes-Benz and GXO. The June 30, 2026 release said the company plans to open new Robot Park locations in more cities soon. More sites mean more task diversity, which is the input both Apollo 3 and Google DeepMind's Gemini Robotics models depend on.
Build the restructured Georgia plant
Keep the roughly 2,000-acre Stanton Springs North facility on track for production in late 2028 under the restructured plan: an initial-phase capacity of 300,000 vehicles per year supported by a renegotiated DOE loan of up to $4.5B that Rivian expects to begin drawing in early 2027. On August 5, 2026 Rivian released detailed site plans and renderings, unveiled at a press briefing at its new East Coast HQ in Atlanta, showing primary factory structures fully enclosed by 2027 to allow tool installation ahead of production, surrounded by a public-facing 63,550-sq-ft visitor build-out (museum, walking trails, cabins, campsites, and an off-road test track) rolled out in three phases. The plant is set to build R2, the upcoming R3/R3X, robotaxis and delivery vans, and could reach full capacity by the end of the decade. By late August 2026 Rivian said site construction was 'progressing well,' with crews grading roads and installing water and sewer infrastructure across the property.
Roll out the Gen 2 autonomy roadmap toward eyes-off driving
CEO RJ Scaringe said in June 2026 that Rivian will bring supervised, Tesla-FSD-like point-to-point driving to its Gen 2 vehicles within 2026, followed by eyes-off autonomy in 2027. Rivian's own Autonomy+ roadmap page, still current as of September 16, 2026, no longer names Universal Hands-Free 2.0 as a Q3 milestone. It lists Universal Hands-Free on 3.5 million miles of US and Canadian roads plus Lane Change on Command on any marked-line road, then jumps to Point-to-Point assisted driving (limited rollout later in 2026, widespread customer rollout targeted for 2027). A footnote still says Universal Hands-Free will not stop or slow for traffic lights or stop signs. The 2026.31 RivianOS 2 update that shipped September 4 was a UI and AI overhaul, not a surface-street driving upgrade. A late-2026 hardware upgrade begins shipping Rivian's ACM3 (Autonomy Compute Module 3) plus a roof-mounted lidar sensor on R2. The stack underpins the performance milestones tied to Uber's up-to-$1.25B investment and the planned 2028 robotaxi rollout in San Francisco and Miami, expanding to 25 cities by 2031. On its Q2 2026 earnings call Rivian said it also still expects an additional $250M Uber equity tranche - one of several milestone-based investments inside the up-to-$1.25B commitment - later in 2026, subject to completion of certain conditions; it had not yet been received as of mid-September.
Confirm official on-sale pricing and open Da Han deliveries
Following its August 21, 2026 presale launch at the Chengdu Auto Show, the Da Han reached dealer showrooms across China on August 24 and BYD opened firm orders the same day, still at presale pricing (249,900 yuan/~$36,850 for the rear-wheel-drive LiDAR Premium trim, 269,900 yuan for the rear-wheel-drive LiDAR Flagship, and 299,900 yuan - confirmed as the model's price cap - for the all-wheel-drive LiDAR Flagship). As of September 5, 2026 BYD has still not announced a distinct final on-sale price or a delivery start date; based on BYD's typical cadence, formal on-sale confirmation and first deliveries are expected within weeks of the order-opening.
Scale FLASH Charging infrastructure globally
The FLASH Charging network reached 10,000 stations across 325 Chinese cities on August 28, 2026 (up from 7,000 by early July, 6,682 across 321 cities in mid-June and 5,715 on May 5), halfway to BYD's 20,000-station year-end target, with 1.83 million users (nearly one-third non-BYD) and 210 million kWh delivered. Partners include Sinopec, PetroChina, CNOOC, TELD and JD.com. That Sinopec framework became a working template on August 8, 2026, when BYD and Sinopec opened their first joint conversion at a working Sinopec petrol station in Shanghai - ripping out the underground fuel tanks and replacing them with six T-shaped 1,500 kW flash chargers (12 stalls) while keeping the site's 24-hour convenience store, a model BYD and Sinopec can repeat across Sinopec's 30,000+ existing fuel stations instead of building new charging hubs from scratch. Hitting 20,000 by year-end would require about 80 new stations a day. International expansion is scaling in parallel: BYD plans 6,000 overseas FLASH stations within a year - 3,000 across Europe by the end of March 2027, 2,000 in the Americas, and 1,000 in Asia-Pacific - to pair with the raised 1.9-2.0 million 2026 overseas target, and in mid-July 2026 it overhauled its overseas brand structure around channel integration and the charging network. Germany's first 1,500 kW FLASH station opened June 9, 2026.
/ All milestones by year
Every announced milestone, listed by target year.