Flash Charging

China Ultra-Fast Charging Infrastructure: A Four-Point Engineering Audit

Mar 27, 2026 EngiVolt Pro

38 days. That is how long remains before Wang Chuanfu’s first verifiable ultra-fast charging infrastructure commitment lands on a hard public date. 4,597 stations are confirmed operational across 279 cities — but the split between highway and urban stations has never been disclosed, and no operator has published what a non-OEM vehicle actually receives at a cross-brand charger. This audit examines four challenge points derived directly from official announcements as of March 28, 2026.

Every claim in this report carries exactly one confidence tag. {VER} = independently sourced. {CLM} = manufacturer or official claim, not independently confirmed. {INF} = engineering inference. {GAP} = data does not publicly exist.

[evp_confidence_bar verified=”18″ claimed=”24″ inferred=”9″ gap=”6″]

CP-01: OEM Ultra-Fast Network Station Counts — What “Built” Actually Means

BYD: 4,597 Flash Charging stations in 279 cities as of March 13, 2026 {CLM — Xinjingbao, March 13}. Geely: 2,103 stations in 215 cities, 1,216 of which are ultra-fast {CLM — Dongchedi, March 9}. BYD has officially confirmed the network is “fully open for sharing across all brands; all EV models meeting GB standards can connect and charge normally” {CLM — Dongchedi, March 9, citing BYD official}.

The audit question is not whether these numbers are real. It is what they mean for a vehicle that is not the network owner’s own brand. Three structural gaps sit between “built” and “available at peak power for your vehicle.”

Gap Layer 1 — Physical connection: confirmed open for BYD network, GB/T compatible {CLM}. Gap Layer 2 — Power negotiation ceiling: not a single operator has published the power curve a non-OEM vehicle actually receives. Gap Layer 3 — Geely network access policy: no equivalent official open-access statement found as of March 28 {GAP}.

Geely’s published “2,103 ultra-fast stations” headline is a common misreading. The audit-correct figure is 1,216 — 42.2% of the 2,103 total are standard, not ultra-fast charging {INF — calculated from CLM source data}.

// cp01 · Infrastructure · Mar 2026
CP-01: Announced Counts vs. Available Capacity
Data Gap
Manufacturer Claim
BYD: 4,597 Flash Charging stations / 279 cities as of March 13, 2026 {CLM -- Xinjingbao, March 13}. Geely: 2,103 stations / 215 cities / 1,216 ultra-fast as of Feb 28, 2026 {CLM -- Dongchedi, March 9}. BYD: network fully open for sharing across all brands; all EV models meeting GB standards can connect and charge normally {CLM -- Dongchedi, March 9}. BYD Gen 2 on BYD Flash Charger: 8 min 45 sec -- 9 min 24 sec, independently measured [V].
Engineering Reality Official claim only
Built does not equal available at peak power. 4,597 is confirmed. What a non-OEM vehicle actually charges at on those stations: DATA NOT AVAILABLE [?]. Geely ultra-fast correct count is 1,216 -- not 2,103. 887 stations in that figure are standard speed [I]. Charging power curve for non-OEM vehicles on BYD network: DATA NOT AVAILABLE [?]. Geely network open access policy for non-Geely vehicles: not confirmed with BYD-equivalent specificity [?].
? Data Gap Built does not equal available at peak power. Cross-brand charging speed on any OEM network: DATA NOT AVAILABLE. Do not treat station count as operational charging speed guarantee.
DATA GAPS: WITHHELD BY OEM
[?]?

As of March 28, 2026, no proprietary ultra-fast charging operator has published actual charging power curves for non-OEM vehicles on their network. BYD confirmed physical access. What has not been published is the power a non-OEM vehicle actually receives. Fleet operators cannot calculate expected charging time for cross-brand sessions. Procurement decisions that treat 9 minutes as the operational figure for non-OEM vehicles are built on an assumption, not a measured number.

CP-02: Cross-Brand Charging Speed Fragmentation

BYD’s position on open access is clear {CLM}. The core audit question has shifted from whether you can connect to what you actually receive. The audit applies a scenario model until T2 data resolves this: moderate scenario (75% power = approximately 12 minutes for a non-OEM vehicle at a BYD Flash Charger); conservative scenario (60% power = approximately 18 minutes) {INF — engineering scenario from CLM source data}.

Zeekr 001 was independently measured at 1,266.4 kW peak at a Geely ultra-fast charger {CLM — Yiche, March 4, 2026}. BYD Flash Charger single-gun peak: 1,500 kW {CLM — Securities Star, March 4, 2026}. What a non-Geely vehicle receives at a Geely station: DATA NOT AVAILABLE {GAP}. What a non-BYD vehicle receives at a BYD station: DATA NOT AVAILABLE {GAP}.

// cp02 · Infrastructure · Mar 2026
CP-02: Cross-Brand Charging Speed
Data Gap
Manufacturer Claim
BYD: Flash Charger fully open for sharing across all brands; all EV models meeting GB standards can connect and charge normally {CLM -- Dongchedi, March 9, 2026}. BYD Flash Charger single-gun peak: 1,500 kW. Host total: 2,100 kW. Voltage 400V--1,000V {CLM -- Securities Star, March 4, 2026}. Zeekr 001 measured peak at Geely ultra-fast charger: 1,266.4 kW {CLM -- Yiche, March 4, 2026}.
Engineering Reality Official claim only
Cross-brand actual charging power curve: DATA NOT AVAILABLE from all operators [?]. Physical access confirmed. Power allocation policy is separate from access policy -- not addressed in BYD open-access statement. No operator has published what firmware-negotiated power a non-OEM vehicle actually receives. Planning estimate until T2: 12 minutes moderate scenario / 18 minutes conservative [I]. If T2 confirms 75% power scenario, the 9-minute headline is a 12-minute operational figure for non-OEM fleets.
? Data Gap Physical access confirmed. Charging power for non-OEM vehicles: DATA NOT AVAILABLE. Planning estimates only until Q2 2026 independent test data.

CP-03: BYD May Day Highway Station Commitment — Execution Pace

Wang Chuanfu publicly committed to 1,000 highway Flash Charging stations operational before May 1, 2026 {CLM — Jingji Ribao, March 6, 2026}. 38 days remain from data cutoff. The audit calculates the current build trajectory against that commitment.

Stations at launch March 5: 4,239 {CLM}. Stations March 13: 4,597 in 279 cities {CLM — Xinjingbao}. Net additions in 8 days: 358 {INF}. Implied monthly build rate: approximately 1,344 stations/month {INF}. Required rate to reach 20,000 by December 31: approximately 1,621/month {INF}. Current shortfall: 17% below required pace {INF}.

Critical caveat: the station-in-station model applies to the 18,000 urban stations. Highway stations require separate approval from state-owned highway service area operators — a structurally longer chain. The May Day commitment specifically tests the harder segment {INF}.

// cp03 · Infrastructure · Mar 2026
CP-03: BYD May Day Highway Station Commitment
Conditional
Manufacturer Claim
Wang Chuanfu: 1,000 highway Flash Charging stations before May Day 2026 {CLM -- Jingji Ribao, March 6, 2026}. Annual target: 20,000 total stations by Dec 31, 2026 [C]. Station-in-station model: no land acquisition, no new permits, existing wiring {CLM -- Xinjingbao, citing Li Yunfei, March 13}. Partnership agreements with existing charging operators for 18,000 urban slots already signed {CLM -- Guancha.cn, March 7}.
Engineering Reality Official claim only
Current build rate (8-day sample): 1,344 stations/month. Required rate for annual target: 1,621/month. Shortfall: 17% -- calculated from CLM data [I]. City vs. highway station split within 4,597 total: DATA NOT AVAILABLE [?]. May Day milestone cannot be independently tracked before May 1 -- no partial progress number exists [?]. Station-in-station advantage applies to 18,000 urban stations; the May Day test specifically applies to 2,000 highway stations, which require state-owned highway operator coordination -- a longer approval chain [I].
⚡ Conditional Real stations are being built. Current pace is 17% below required rate. May Day 2026 -- 1,000 highway stations committed -- is the first independently verifiable hard checkpoint.
DATA GAPS: WITHHELD BY OEM
[?]?

BYD has not separately disclosed the urban/highway breakdown in any public filing or press release. The May Day 1,000 highway station commitment cannot be independently tracked before May 1. There is no partial verification path. The only data point will be BYD’s own official announcement on or around that date.

May Day Checkpoint
BYD highway station count announcement. Above 1,000: annual target acquires execution credibility. 700--900: execution real but behind pace. Below 700: annual 20,000 target statistically unreachable without step-change in build rate.

CP-04: Fragmentation Cost Impact on Fleet TCO

The fragmentation premium is quantifiable. The audit model: 60-vehicle intercity fleet, charging 40% of sessions off-OEM network, non-OEM power scenario 75% of OEM benchmark (12 minutes vs. 9 minutes). Estimated annual fragmentation cost: RMB 295,200–346,000 {INF — scenario model; actual figure subject to T2 recalibration}.

The highest-weight variable is not the power degradation magnitude — that is outside operator control. It is the frequency of off-OEM charging sessions, which is an operator routing and dispatch decision. Reducing off-OEM charging frequency from 40% to 10% cuts the fragmentation premium by approximately 75% — from approximately RMB 295,000 to RMB 74,000 annually {INF}. This action requires no technology change. It is available today.

// cp04 · Infrastructure · Mar 2026
CP-04: Fragmentation Cost on Fleet TCO
Verified
Manufacturer Claim
BYD Gen 2 on BYD Flash Charger: 8 min 45 sec -- 9 min 24 sec [V]. Non-OEM vehicle on BYD Flash Charger: planning estimate 12--18 min (moderate/conservative scenario) [I]. Ti3 entry fleet unit price: RMB 153,800 {CLM -- Fang Cheng Bao official, March 13}. Post-Year 1 charging rate: industry standard parity, no premium {CLM -- Jingji Ribao, March 6}. Off-OEM charging frequency: operator-variable; baseline assumption 40% [I].
Engineering Reality Independently verified
Moderate scenario (75% power = 12 min): annual fragmentation premium for 60-vehicle intercity fleet at 40% off-OEM frequency: RMB 295,200--346,000 [I]. Highest-weight variable: off-OEM charging frequency -- operator-controlled. Reducing from 40% to 10% cuts premium by 75% to RMB 74,000. This is a route planning decision, not a technology dependency. Actionable today. Cross-brand power curve estimate subject to revision on T2 trigger (Q2 2026) [?].
✓ Verified Fragmentation cost is quantifiable and operator-controllable. Reduce off-OEM charging frequency from 40% to 10% to cut annual premium by 75%. No technology change required.

Evidence Grid: OEM Network Comparison

// OEM Network Comparison — Data cutoff March 28, 2026
Dimension BYD Flash Charging Geely Ultra-Fast Tesla Supercharger (China)
Stations built 4,597 / 279 cities {CLM — March 13} 2,103 / 215 cities (1,216 ultra-fast) {CLM — Feb 28} ~13,000+ as of end-2025 {CLM — EVCIPA}
Single-gun peak power 1,500 kW {CLM} 1,266.4 kW measured at Zeekr {CLM — media test} 250 kW (V3) {CLM}
Open access for non-OEM CONFIRMED {CLM — Dongchedi, March 9} Not confirmed with equivalent specificity {GAP} Limited rollout in China {CLM}
Cross-brand power curve published DATA NOT AVAILABLE {GAP} DATA NOT AVAILABLE {GAP} DATA NOT AVAILABLE {GAP}
City / highway split disclosed No {GAP} No {GAP} Partial — urban/highway segmentation in media

Data Gap Summary

DATA GAPS: WITHHELD BY OEM
[?]?

No equivalent official open-access statement found for Geely network in public reporting as of March 28, 2026. Until confirmed otherwise, Geely ultra-fast network should be treated as OEM-exclusive for non-Geely fleet planning. Monitoring Trigger T8.

// Download Full Report

China Ultra-Fast Charging Infrastructure Fragmentation Audit

Full 30+ page audit report: four challenge points, build rate calculations, fleet TCO fragmentation model, eight monitoring triggers. Data cutoff March 28, 2026.

$49 — Get the Report →