In Europe, the Network Is Not the Bottleneck. The Data Flywheel Is.

TL;DR

Dealer count is the input metric, not the output. By 2027, the first wave of Chinese-brand EVs sold into Europe in 2024-25 will be three years old — the point at which warranty buffers thin and residual value gets re-priced on real ownership data, not projections.

Which Chinese brands enter the 2029 first tier — building service, leasing trust, and localized governance — versus which stay in a long-term second tier dependent on price war and subsidized leasing depends less on how many points of sale they open in 2026 than on whether they build — or rent — the operational stack underneath. Partnership models (Stellantis-Leapmotor, Volvo-Lynk&Co, Ayvens-BYD) are buying time and access; they are not yet a substitute for that build.

Three quick things this isn't claiming. First, it's not saying Chinese OEMs will fail in Europe — they will increasingly succeed. Second, it's not saying European traditional service systems are a permanent moat; they're brittler than insiders like to admit. Third, it's not denying Chinese OEMs' very real weapons — supply-chain cost discipline, software iteration speed, organizational adaptiveness, HQ willingness to underwrite losses for longer than Western boards would tolerate. The argument is narrower than any of those, and harder to dismiss because of it: the next material competitive divergence in Europe will run within the Chinese-brand cohort, not between Chinese and European brands. And the line that splits the tiers is operational.

The next material competitive divergence in Europe runs within the Chinese-brand cohort, not between Chinese and European brands. The line that splits the tiers is operational.

What each dealer is actually for

Most analyst pieces I read about Chinese OEMs in Europe count dealers. BYD is doubling its European points of sale from 1,000 in 2025 to 2,000 by the end of 2026. Leapmotor reached 600 sales-and-service points in 2025, on track for 700+ by year-end, with a target of putting the average drive time to a Leapmotor location below 25 minutes. MG (SAIC) sits on 1,300+ dealer partners in 34 markets and just crossed one million cumulative European deliveries. XPeng now lists 290+ retail points across 28 countries.

These are construction metrics. They tell you how fast a Chinese OEM is putting things up in Europe. They do not tell you what each dealer is actually for.

By end of 2026, the first volume cohort of Chinese-brand EVs sold in 2024-25 will be approaching three years on the road. At that age, warranty buffers thin, first-fix rate matters more than the showroom experience, and the residual-value lens that decides every fleet contract starts pricing real ownership data rather than projections.

A confession of lens: the argument here is Germany-weighted by design. Germany sets the institutional residual-value template that southern, eastern, and UK markets adapt to with different weights — southern Europe more price-sensitive, eastern Europe more so, the UK with its own leasing structure. The argument applies most strongly to Germany; less so elsewhere. But Germany is where 2027’s residual-value standards get set for brands aspiring to mainstream fleet positioning across the continent.

I run IT integration for a Chinese OEM’s European operation. From that seat, you sit at the layer where dealer DMS systems try to talk to factory ERPs, where leasing-partner risk engines try to read VIN-level data, where homologation audits sit on top of OTA pipelines. None of that shows up in registration data — and what you see from here is that the "China playbook" of fast iteration and dense software does not translate the way Chinese boards still seem to expect. This is what I think is being underestimated.

Sales velocity, mistaken for service velocity (and not all brands the same)

Almost every public dealer-count number is bundled. "Sales & service points." You cannot, from any public source I have found, cleanly split BYD’s 2,000 European POS target into a service-bay count, a parts-warehouse count, and an HV-certified-technician count. The same is true for Leapmotor — and notice that two press cycles three months apart show 600 and 800+ POS respectively, which suggests the counting methodology is itself drifting.

One thing worth saying upfront: the argument here does not apply evenly to all Chinese brands. The 2025 USCALE EV Satisfaction Study in DACH put XPeng at the top with an NPS of 81 — the first Chinese brand to outscore Tesla. Polestar and NIO also placed in the top ranks. MG and Smart "disappoint," in the study’s word. This is a wide intra-Chinese spread on customer experience, not a uniform pattern. The downstream argument matters more to brands now putting up the POS without a satisfaction track record yet — BYD, Leapmotor, Voyah, Hozon — and less to brands that have already shown the rest can be built.

The ADAC Pannenstatistik 2026 confirms a second inconvenient fact: there is no entry yet for Chinese brands. The 2026 dataset shows EVs as a group are now meaningfully more reliable than ICE (4-year-old EV breakdown rate of 6.5 per 1,000 vehicles versus 12.5 for ICE). But ADAC requires minimum fleet age and registration base — Chinese brands have either or both missing.

Two readings of that absence are both true. The neutral one: it's normal for new entrants — their cars are too young. The risk-side one: in January 2026, Chinese brands crossed 8% of German EV registrations, with BYD outselling Tesla 2-to-1 and Leapmotor up +344% year-on-year. These vehicles will enter the ADAC dataset on the dataset’s schedule, not the OEM’s — somewhere in 2027-28. There is no public dashboard right now that tells a German fleet manager whether a 2025 BYD Atto 3 in their leasing book is going to need 3.2 unscheduled service visits a year or 0.8.

Sales velocity is being measured precisely. Service velocity is not being measured at all.

So what: evaluating a Chinese OEM’s European trajectory through POS growth means using an input metric. The output metric — service throughput — is currently unmeasurable from outside, and we are 12-18 months away from the first reliable external signal.

The three operational rails you cannot capital-expense your way through

A reasonable counter-argument is that any gap can be closed with capital and time. Three specific rails do not respond to capital the way "open more stores" does. They respond to time, to regulation, and to how fast institutions absorb new entrants.

Rail 1 — Parts logistics latency

KPMG’s March 2025 report on Chinese OEMs in Europe is blunt: the supply chain of Chinese manufacturers in Europe "does not yet fully meet European standards," with cases where vehicles cannot be repaired due to a shortage of parts, creating uncertainty among insurers. Industry shorthand puts the practical gap at weeks of lead time for Chinese-brand parts versus 24-48 hours for established brands. EU-side bonded warehousing, regional depot networks, and just-in-time logistics contracts (Leapmotor signed one with Grimaldi in 2026) take years, not months, to reach 24-hour SLA reliability.

Rail 2 — HV-certified technician supply

EV high-voltage service requires separately certified technicians — a regulated skill, not a learn-on-the-job role. The technician shortage in EU EV service is structural and predates Chinese entrants. Chinese OEMs are the newest brands in a queue where every existing OEM is also hiring; pipeline lead time for HV certifications runs through TÜV, DEKRA, or IHK-equivalent bodies on their schedules. As a service-bay metric: JD Power’s US 2025 EVX Ownership Study found 12% of EV repairs are not completed correctly on the first visit, with parts unavailability cited in 28% of those cases. The pattern — well-trained EV technician shortage plus parts gaps — is the same diagnosis European technicians describe.

Rail 3 — Compliance-bound OTA cycles

UN R155 (cybersecurity management) and UN R156 (software update management) require manufacturers to operate a certified Software Update Management System with three-year recertification and independent audit. In practice this turns OTA pipelines from a "ship weekly" function into a planning-and-evidence function: any safety-relevant update must be planned, validated, and documented against an approval framework before release.

I should be honest about what this argument is risking. Chinese OEMs’ real strengths — supply-chain cost discipline, iteration speed, HQ willingness to underwrite losses, organizational adaptiveness — compress timelines that legacy OEMs assume are immovable. The Chinese-smartphone analogy is real: in 2010, Xiaomi, Huawei, Oppo, and Vivo had no global service network, no software ecosystem, and weak international IP defenses. In under a decade, they closed every one of those gaps. That historical pattern should be in the back of your head as you read this article. The risk in the framing here is mistaking a time differential for a structural moat — and those are not the same thing.

Subsidy and speed do not dissolve the three rails above, though. Capital does not compress TÜV certification cycles for HV technicians, or UN R155/156 audit windows, or the time it takes a bonded-warehouse network to reach 24-hour SLA reliability across 27 jurisdictions. Those are clock-bound, not capital-bound. The Chinese-brand learning curve will close — for some brands. The question is which ones, and how much market share they cede during the closing.

So what: the OEMs that will be in healthy shape in 2028 are the ones who started running these tracks in 2024, not the ones whose first board meeting on parts logistics is in 2027. The brands that wait do not necessarily fail. They get sorted into the second tier.

Why leasing is the actual battlefield (and what's working right now)

Europe is a leasing market. The leasing pricing engine, not the showroom sticker, determines whether a Chinese-brand EV gets monthly-payment competitive in the consideration set most buyers actually use. That engine runs on residual value (RV).

The RV gap between Chinese-brand and established-brand vehicles in Germany has narrowed from roughly 16-17 percentage points in 2023 to about 9 pp in 2025-26; Spain shows a narrower 7-pp gap. That improvement is real — but the gap sits on top of a wider EV-vs-ICE RV problem (one study put EV used-value decline at -46% versus -19% for comparable ICE), and on top of Eurotax’s outlook for 2026 expecting a further 0.7% decline across the EV segment with no relief expected through 2026. The most pointed reading: a three-year-old Škoda Elroq or Kia EV3 is forecast to be worth meaningfully more on the German used-car market than an equally aged Leapmotor B10. For BYD, the same outlook notes that reliable resale experience is "still lacking." That is not a verdict on the product. It is a verdict on the data.

Short-term partnership routes are working right now, though. Ayvens and BYD extended their fleet partnership to 11 European countries by 2025, with a target of supporting 30+ enterprises in the first year through a white-labelled "BYD lease" product for SMEs and retail. Leapmotor’s €49-per-month T03 offer in Germany rides Stellantis-absorbed RV risk. These are not loopholes — they are working strategy for now. The question they don’t answer is whether the Chinese OEM gains the data or the partner retains it. Partner-absorbed-risk lets you sell fleet vehicles today. It does not produce the proprietary RV evidence base you need to graduate from "discount lease" to mainstream fleet pricing in 2028 onward.

The second-hand market is the next domino. The first material wave of Chinese-brand fleet vehicles will hit used auctions in 2026-2027 as 24-to-36-month leases roll off. If those trade below their RV forecast — because parts availability, repair history, or technician capacity made the in-life ownership experience worse than the model assumed — the next RV cycle re-prices lower, and the discount-lease lock-in deepens for whichever brands cannot absorb that feedback fast.

So what: the visible competition for fleet share is happening in 2026 via partnerships. The decisive numbers are what Schwacke, Eurotax, and Glass’s publish in 2027 once the first cohort of used vehicles clears auction. Those numbers are set by what is happening in service bays right now.

The data flywheel that is actually decisive

Traditional OEMs in Europe defend a residual-value moat that is, mechanically, a data moat. A leasing company quoting a 36-month residual on a Volkswagen ID.4 uses 20+ years of Volkswagen telemetry, dealer repair-history files, regional parts-availability indices, and country-by-country resale curves. None of that is available for Chinese brands yet. The 9-pp residual gap is what that absence produces.

The reasonable Chinese-OEM response — which I think is correct in principle but not yet executed — is that car-side data is now better than it has ever been. BMS telemetry, driving-behaviour signals, OTA-driven predictive-maintenance flags. A 2025 BYD or NIO has more recoverable per-vehicle data per month than a 2010 Volkswagen had over its entire warranty life. In principle, that data could power a residual-value model that beats the legacy approach.

In practice, it doesn't yet — and the reason is not technical. It's integration, governance, and trust between four parties who have never built a shared system. Here is what that integration actually looks like, from the inside:

The IT integration backbone of new-vehicle sales runs along a single business line — lead → quote → leasing approval → contract → handover. The data flows across four very different systems: the OEM headquarters ERP (wholesale, production, finance), the local sales subsidiary’s CRM (leads, opportunities, contracts), the dealer’s DMS (inventory, retail, vehicle owner), and the third-party leasing partner’s risk decision system (KYC, credit, loan approval). These four belong to four different legal entities. The interface styles — REST, SOAP, SFTP, batch file — do not agree. The master-data definitions — VIN, customer ID, contract number — do not agree. The SLAs do not agree. The real difficulty is not the protocol layer. It is in three places: field-level semantic alignment (the same word "delivery date" refers to different events in each of the four systems); transactional consistency and idempotency (when any single step fails, who reconciles, who compensates); and cross-entity data compliance (GDPR/DPIA, the data-processing agreement with the credit-scoring partner, the legal basis for cross-border transfer). Local IT on this chain is, in practice, not an engineering function — it is an interface-governance function.

Local IT on this chain is not an engineering function — it is an interface-governance function across four legal entities and several regulators.

That is one chain. The residual-value chain involves the same four-actor topology with different fields and a different cadence. The insurer-claim chain, again. The parts-availability-to-DMS chain, again. Each chain has its own version of "the protocols don’t agree, the master data doesn’t agree, the SLAs don’t agree, and the hard part is the field-semantics and the compliance, not the API."

The reason Chinese OEM’s better car-side data does not yet improve their European RV is that no one has built the governance layer that lets it flow to the lessor’s risk engine in a form the lessor’s regulator will accept. GDPR, the EU Data Act, the data-processing-agreement chain with credit partners, and the cross-border-transfer basis for any Chinese-parent data lake all sit between the OEM and the residual model. The Chinese-side advantage — data abundance — becomes, at the integration boundary, a compliance debt.

There's a counter here I should answer directly. European traditional OEM service systems may not be a moat at all. They are a cost structure — thick dealer layers, high union-driven service costs, slow software cadence, expensive technician certification, legacy customer-experience friction. Chinese OEMs that leapfrog the old system rather than copy it might find a different — and possibly better — equilibrium. There is no rule that the next dominant fleet-RV model has to look like 2025 Volkswagen’s. That cuts both ways for the thesis here: if the next equilibrium isn’t the old equilibrium, the legacy data moat is less defensible than legacy OEMs assume. But the new equilibrium has to be built by someone — by some Chinese OEM, or by a Chinese OEM in partnership with a European one — and that build is the same governance problem described above. The flywheel just runs a different way around. Either way, somebody runs the four-actor governance layer.

A few moves are visibly happening. XPeng’s #1 ranking in the USCALE 2025 DACH study suggests at least one Chinese brand has solved the front-end data-and-service stack at the consumer surface. The Lynk&Co-Volvo MoU of March 31, 2026 hands Lynk&Co’s European retail and 350-service-point footprint to Volvo entirely (the MoU is non-binding and final agreements are still pending). Polestar pivoted in March 2025 from pure agency back to using Volvo’s dealer network, going from 9 to 17 German locations. NIO is trying to build its own and closed Denmark to fund the rest. None of these is yet a settled answer.

So what: the strategic question is not which Chinese OEM has the most dealers in 2026. It is which Chinese OEM first builds — or gets durable institutional access to — a localized OEM-dealer-lessor-insurer data flywheel before 2028’s first material RV reset.

The 2027 inflection and what watchers should track

If the thesis is right, the change in market structure will not be a public crisis. It will be a quiet divergence in five places, all visible from open data.

1. Intra-Chinese brand divergence by name. Watch for at least one Chinese brand to enter ADAC’s reliability table in line with European average — likely XPeng or Polestar first, based on USCALE momentum. That brand effectively breaks the "all Chinese brands are unproven" frame. This is the single most important leading indicator.
2. ADAC Pannenstatistik 2027 or 2028. The first edition that includes Chinese brands by name. The 2-year and 3-year breakdown-rate columns will show who is absorbing parts-and-service execution and who is not.
3. Schwacke / Eurotax 2027 brand-level residual indices. If the Germany gap narrows from 9pp to 5-6pp, the data-trust infrastructure is being accepted. If it widens back toward 12-15pp, the early-cohort used-car experience has fed back into the model.
4. JD Power Europe Customer Service Index, 2027. Where Chinese brands appear, at what gap relative to the EV baseline. The US data already shows BEV CSI trailing mass-market ICE by 51 points; the EU pattern will likely look similar with brand-specific divergence.
5. First roll-off of fleet vehicles to auction. Hammer-price-versus-forecast for the first ~5,000 Chinese-brand fleet returns in 2026-27 is, in residual terms, the closest thing to a verdict day.

Short-term sales market share growth is not a signal to over-weight. Chinese-brand share of the German EV market is on a steep growth curve driven primarily by price and product. That number tells you the front-end works. It tells you nothing about the back-end. It also tells you nothing about which Chinese brands graduate.

Which tier will you be in?

If you’ve read this far, the standard "Chinese OEMs versus European OEMs" frame is the wrong one. Several Chinese brands are already winning in Europe. The sort that matters next is within the Chinese cohort: which Chinese brands move into a 2029 first tier — building service throughput, leasing trust, localized governance, fleet confidence — and which stay in a long-term second tier dependent on price war, subsidized leasing, and partner-absorbed RV risk. Both are viable businesses. They are not the same business. The first-tier margin pool and the second-tier margin pool diverge sharply by 2030.

The standard frame for Chinese OEM expansion into Europe is product-and-tariff. That frame covers what the conventional analyst report needs to say. What it misses is that replicating the network is easy. Replicating operational maturity is not. The China playbook treats hardware as the static layer and software as the dynamic layer. Europe inverts that: hardware — dealer service, parts logistics, technician certification, residual data, regulatory cadence — is the dynamic competitive surface, and software is a constraint-bound asset. Translating between the two is not a deploy command. It is a multi-year governance build across four legal entities and several regulators.

Replicating the network is easy. Replicating operational maturity is not.

For a Head of Europe or Head of Strategy at a Chinese OEM, three implications. POS count as a 2027 success metric will be the wrong number, because the metric that decides 2028 fleet RV is currently invisible to your own dashboard. The team that runs your European interface-governance stack — DMS, ERP, CRM, leasing-partner risk integration, OTA homologation, GDPR compliance — is statistically two people and a contractor; adding eight people there will return more in 2028 RV terms than another fifty POS. And the tier sort is already starting. You do not get to start in 2027 and catch up.

I’ll be writing more on this through 2026. The data above is public; the framing is mine; the disclaimer applies. If any of this is right enough to be worth disagreeing with, the LinkedIn DM is the right channel.

Sources

• BYD aims to double European sales network by end-2026 — Automotive News, November 2025
• Leapmotor European expansion with new market entries and dealer appointments — Stellantis Media
• SAIC's MG brand achieves one million European sales milestone — Just Auto
• USCALE EV Satisfaction Study 2025 — USCALE, July 2025
• ADAC Pannenstatistik 2026 — ADAC, April 2026
• Chinese EVs in the German market, January 2026 — AutoColumn (KBA-referenced)
• Impact of Chinese OEMs in Europe, Part II — KPMG, March 2025
• Leapmotor + Grimaldi Group logistics partnership — Stellantis Media
• JD Power 2025 US EV Experience Ownership Study — JD Power
• Software Update Management Systems per UNECE R156 — UL Solutions
• Can Chinese EV brands flourish in Europe? — Autovista24
• The Second-Hand Market in the Electric Vehicle Transition — MDPI, 2025
• Ayvens extends its partnership with BYD to seven new European countries — Ayvens, 2025
• Leapmotor expands in Germany with EV costing 49 euros a month — NAI500, May 2026
• Lynk & Co in Europe to be operated by Volvo Cars — Lynk & Co newsroom, March 2026
• Polestar to use Volvo dealer network — electrive, February 2025