CATL's FY2025 annual report disclosed an overseas gross margin of 31.44% against 24.00% domestic. That 7.44 percentage point premium is the clearest fingerprint of the VAT export rebate's role in Chinese battery export economics. Mysteel's January 2026 analysis, citing H1 2025 semi-annual filings across the sector, stated it without hedging: "Due to the support from export tax rebates, the gross profit margins of listed companies' overseas operations are generally higher than those of their domestic operations."
The rebate dropped from 9% to 6% on April 1, 2026. It goes to 0% on January 1, 2027. The schedule is published, the mechanism is central-government fiscal policy, and the direction is one-way. The most precise consolidation map available for Chinese battery exports right now is an assessment of which producers can absorb the full elimination and which cannot.
Disclosed Margin Spreads as Anchor Data
Three publicly listed Chinese battery producers disclose overseas-versus-domestic margin splits with enough granularity to work with (Tier 1 sourcing, public filings):
| Producer | Overseas GM | Domestic GM | Premium | Source |
|---|---|---|---|---|
| CATL | 31.44% | 24.00% | 7.44 pts | FY2025 annual report |
| EVE Energy | 21.71% | 15.88% | 5.83 pts | H1 2025 semi-annual |
| Gotion High-Tech | ~19% | ~15% | ~4 pts | H1 2025 semi-annual |
CATL's H1 2025 split was 29% versus 23% per Mysteel; the widening in H2 likely reflects product mix and pricing power rather than any change in the rebate mechanism.
The pattern holds across all three: 4 to 7.5 points of overseas margin premium during the period when the rebate stood at 9%. But the observed premiums are smaller than the rebate's total contribution to export economics, because producers were not keeping the full benefit. Industry bodies have acknowledged that rebates "were passed directly to overseas buyers as price discounts," contributing to sharp declines in export pricing. The overseas margin premium reflects only the portion producers retained; the rest was shared with customers as competitive pricing. When the rebate disappears, producers lose both portions simultaneously: the margin they kept and the pricing advantage they were funding. Raising export prices to recapture the passed-through portion restores margin but surrenders the competitive position the rebate enabled.
Mysteel estimated the cell-level cost impact of full elimination at RMB 0.03–0.04/Wh, based on a cell price range of RMB 0.35–0.40/Wh. That translates to 8–11% of the cell's export price absorbed as non-recoverable VAT.
The full rebate loss, not just the premium portion visible in disclosed margins. The subtraction exercise below uses the Mysteel cost-impact range rather than the observed premiums, because it captures the total margin hit including the portion previously passed through.
Subtracting the Rebate
CATL at 31.44% overseas gross margin, losing approximately 8–11 points of rebate-supported margin, lands in the range of 22–24%. That is at or near its domestic gross margin. The overseas business becomes less profitable but remains structurally viable. CATL's scale, technology licensing revenue, and product mix advantages persist independent of the rebate. Absorption capacity: unambiguous.
EVE Energy at 21.71% overseas gross margin lands approximately 13–15% post-elimination. Tighter. The segment-level data matters here: EVE's energy storage segment gross margin was only 12.28% for FY2025, already compressed 2.44 points year-on-year, well below the company's blended overseas figure. For a producer in the 20–25% overseas margin band, the outcome depends on which product lines carry the export volume. EVE's power battery exports, likely closer to the blended overseas margin, can absorb the hit. Its ESS exports, starting from a lower base, face a post-rebate residual that could be single digits (Tier 3, editorial inference based on the gap between segment and blended margins).
Gotion at approximately 19% overseas gross margin lands around 10–12%. Its FY2025 overall gross margin was 16.17%. An overseas margin of 10–12% would sit below its blended domestic margin, meaning exporting becomes actively less profitable than selling at home. Gotion has a Volkswagen strategic relationship and global capacity commitments that create non-economic reasons to continue exporting. But the financial incentive to push incremental volume through export channels weakens materially. Gotion's financial expenses surged 71% in FY2025 on FX losses alone, compounding the margin pressure from a direction the rebate phase-out doesn't touch.
REPT Battero is where the math turns existential. FY2025 blended gross margin: 11.2%, with power batteries at 11.9% and energy storage at 10.8%. REPT does not separately disclose overseas-versus-domestic margins. Applying the observed 4–6 point overseas premium pattern from Gotion and EVE (Tier 3, editorial inference based on structural similarity) suggests an overseas gross margin of perhaps 15–17%. Subtract the full rebate elimination and the residual is 6–8%. After operating expenses, SG&A, and financial costs, the export business is at or below breakeven. REPT posted its first annual profit in 2025, RMB 681 million, after a RMB 1.35 billion loss in 2024. The margin of safety is negligible.
CALB does not disclose geographic margin splits, but its FY2025 net margin was 4.7%, implying gross margins well below the Tier-1 range. Post-rebate export economics are likely in the same territory as REPT's.
| Producer | Pre-elimination overseas GM | Post-elimination est. range | Sourcing | Assessment |
|---|---|---|---|---|
| CATL | 31.44% | 22–24% | Tier 1 | Viable; at domestic parity |
| EVE Energy | 21.71% (blended) | 13–15% (blended); ESS possibly single digits | Tier 1 / Tier 3 for ESS | Survivable; ESS line stressed |
| Gotion | ~19% | 10–12% | Tier 1 / Tier 2 | Below domestic margin; export incentive weakens |
| REPT Battero | ~15–17% (est.) | 6–8% | Tier 3 | At or below operating breakeven |
| CALB | Not disclosed | Likely similar to REPT | Tier 3 | Structurally challenged |
The Uneconomic Band
Based on the disclosed data and the triangulation above, I assess the critical margin band as follows (Tier 3, editorial construction):
Below approximately 18% overseas gross margin pre-elimination: The post-rebate residual falls to 9–11%. After operating expenses, the export channel generates negligible or negative operating profit. REPT, CALB, and the long tail of sub-scale producers sit here. TrendForce's assessment, relayed by InfoLink, was direct:
"The cancellation will directly breach the cost threshold for low-margin companies."
This is also the band where the rebate phase-out does consolidation work that regulation cannot reach. GB 38031-2025 applies to traction batteries, not ESS cells, and its enforcement date for existing models is July 2027. The rebate elimination hits ESS exporters on the same schedule as power battery exporters, making it the operative consolidation mechanism for the storage segment specifically. REPT's ESS-specific gross margin of 10.8% and EVE's 12.28% are the numbers that make this concrete.
Between 20% and 25%: Survivable but compressed, and the outcome diverges by product line within the same company. EVE and Gotion sit here at the blended level. A producer whose export mix skews toward ESS cells, where segment margins already run 5–9 points below blended overseas margins, faces a materially worse post-rebate position than one exporting primarily power batteries with higher ASPs and better margin structures. The ability to pass through cost depends on customer concentration and contract structure: producers locked into fixed-price annual agreements with large integrators have less room than those selling spot into fragmented markets.
Above approximately 25%: Post-elimination residual above 16%, which is workable. CATL occupies this position. BYD likely does as well, though BYD's geographic margin split is not disclosed with comparable granularity.
For procurement teams sourcing Chinese cells, the implication is specific: post-January 2027, Tier-2 exporters either raise prices by approximately RMB 0.03–0.04/Wh to maintain margins, or continue exporting at current prices and accept operating losses on every shipment. The first option narrows their price advantage over Tier-1 producers and non-Chinese suppliers. The second is not sustainable beyond a few quarters.
Why Provincial Governments Cannot Offset This
Chinese battery overcapacity has been a consensus concern for three years without producing the consolidation that analysts keep predicting. The reason is institutional. Provincial governments subsidize local producers for employment and GDP contribution. A factory running at 40% utilization but employing 2,000 people in a Tier-3 city has a political constituency that keeps it alive regardless of its return on capital. Overcapacity can persist indefinitely under this logic.
The VAT export rebate is a central government fiscal mechanism. Provincial governments cannot replicate it. They cannot issue local VAT rebates or create alternative export tax incentives. They can keep a factory running for domestic sales. They cannot make exporting profitable when Beijing has decided to withdraw the margin support.
The ratchet property: the phase-out does not reverse with commodity cycles. Lithium carbonate can spike above RMB 200,000/MT (as it did in May 2026) or collapse back toward RMB 73,000 (as in October 2025), and the rebate schedule does not change. A producer can cut costs, improve yields, negotiate better cathode pricing. The schedule still does not change. The only variable is the published step-down: 6% since April 1, 0% in January.
The Pull-Forward as Revealed Preference
Q1 2026 battery exports reached 84.1 GWh, up 36.7% year-on-year. March alone hit 36.1 GWh, a 57.1% year-on-year surge. Storage battery exports in March rose 96.9% month-on-month. In dollar terms, Benchmark Minerals reported HS 8507.60 exports up 55% in Q1 to $23.95 billion.
The customs declaration deadline was March 31. Goods declared by that date received 9%. Goods declared April 1 or later receive 6%. The pull-forward was mechanically predictable, and the data confirms it arrived on schedule.
The composition matters more than the aggregate. REPT doubled its exports year-on-year in Q1. China Automotive New Energy posted a 723% increase. The surges track the customs declaration deadline, with producers and customers accelerating shipments to capture margin being withdrawn on a published schedule. Benchmark flagged the same logic predicts a second pull-forward in Q4 2026, ahead of the January 1 elimination of the remaining 6%.
The Q4 surge will be the last. After January 1, 2027, there is nothing left to pull forward. Some Tier-2 producers will attempt to localize production in Southeast Asia, where the Chinese VAT mechanism does not apply. REPT's Indonesia manufacturing base is planned at 8 GWh but is under construction, not producing. I do not have sourced data on REPT's expected qualification timeline for that facility; typical cell manufacturing greenfield-to-qualified-production timelines in ASEAN run long enough that this is a 2028–2029 solution at the earliest to a January 2027 problem.
Provincial governments kept overcapacity alive for three years. The central government is now delivering consolidation through fiscal withdrawal. Quiet, scheduled, irreversible. The post-rebate export landscape is one where CATL's cost absorption capacity widens into an even deeper moat. The rebate compressed the visible gap between Tier-1 and Tier-2 export economics for years. January removes the compression.
Things to follow up on...
- Q4 2026 pull-forward surge: Benchmark Minerals flagged the likelihood of a second export acceleration in Q4 2026 as buyers and producers rush to capture the remaining 6% rebate before the January 1, 2027 elimination.
- Post-April export volume data: The first full month of customs data under the reduced 6% rebate regime should surface in May–June reporting; Benchmark's Q1 tracking provides the pre-cut baseline against which to measure any volume drop-off.
- Gotion's overseas revenue reversal: Gotion's overseas revenue declined 7.47% in FY2025 even before the rebate step-down, a notable shift from 31% of revenue in 2024 to 22.6%, per BESSM's annual results analysis.
- REPT Indonesia qualification timeline: REPT's planned 8 GWh Indonesian manufacturing base is under construction but not producing; the company's FY2025 results disclosed the facility without a commissioning date, making it a critical variable for whether REPT can route exports outside the Chinese VAT regime before margins compress further.