CATL's net finished goods inventory rose from RMB 22.6B to RMB 47.6B in the first half of 2026, a 111% increase. Goods dispatched but not yet recognized as revenue grew 4.7%. Contract liabilities, meaning customer prepayments against product not yet delivered, fell 25.9%, from RMB 49.2B to RMB 36.5B. Reported capacity utilization held at 94.86%.
Read together, those four figures describe a producer running near full capacity, accumulating the resulting surplus as finished product, and carrying it without any visible financial pressure to move it. Neither of the reflexive interpretations survives contact with the numbers. Demand strength does not produce a 4.7% dispatch increase alongside a 26% drop in prepayments. Distress does not produce a RMB 372B cash position and 21% domestic gross margin.
All figures below come from CATL's H1 2026 interim report, filed through HKEX on July 26, roughly four weeks after the June 30 reporting date. This analysis follows about four weeks after that filing, which puts the balance sheet position described here at nearly eight weeks old. CnEVPost and Reuters, both on July 24, covered headline results, revenue up 54.8% and net profit up 42%, but broke out neither inventory composition nor contract liabilities. A Wallstreetcn relay on July 26 came closest, reporting net finished goods, goods shipped, and the contract-liability decline. For all three signals with matched gross, provision, and net detail, the filing remains the only complete English-language source.
One sourcing discrepancy to clear first. Some coverage quotes a 97% increase in finished goods. That references gross finished goods, RMB 26.1B to RMB 51.4B, +96.9%. The net figure, after provisions for inventory write-downs, gives the 111% increase I use throughout. The provision itself barely moved, RMB 3.5B to RMB 3.7B. CATL is not writing down the inventory it is accumulating, which becomes relevant below.
The three signals
| Signal | Dec 2025 | Jun 2026 | Change |
|---|---|---|---|
| Net finished goods | RMB 22.6B | RMB 47.6B | +111% |
| Goods dispatched | — | — | +4.7% |
| Contract liabilities | RMB 49.2B | RMB 36.5B | −25.9% |
| Utilization | — | 94.86% | — |
Finished goods: +111%. Production outran deliveries by a wide margin. The trajectory matters more than the level. Net finished goods also roughly doubled across H1 2025, from RMB 10.8B at year-end 2024 to RMB 22.0B by mid-year, a 103% increase, then went flat through H2 2025, rising 2.6% to RMB 22.6B by December. H1 2026 repeats the H1 2025 pattern from a higher base.
Two observations of the same half-year shape do not distinguish a seasonal production pattern, where output runs ahead of second-half delivery schedules, from a structural one. A third (H1 2027) would. The filing offers nothing further to work with: CATL does not segment finished goods by product type, chemistry, geography, or order status. The entire RMB 47.6B sits in one undifferentiated line.
(Comparability note: as covered in Issue #10, CATL's restated H1 2025 comparatives do not always reconcile with originally reported figures because of reclassifications. The trajectory comparisons above use same-vintage figures where available.)
Goods dispatched: +4.7%. This is what sharpens the finished-goods number. "Goods dispatched" covers product that has physically left CATL's facilities but has not yet been recognized as revenue, typically in transit or awaiting customer acceptance testing. If the finished-goods buildup were pre-positioning for imminent delivery, dispatched goods should be rising in something like proportion. A 4.7% increase against 111% means the distance between what CATL produced and what it shipped widened considerably.
The half-yearly sequence points the same way. Dispatched goods jumped 51.5% during H2 2025 while finished goods sat flat; in H1 2026 the two reversed.
Contract liabilities: −25.9%. Contract liabilities are consideration CATL has received, or that has become due, for goods it has not yet delivered. Customer prepayments, in practice, against future shipments. The 26% fall from the December 2025 peak of RMB 49.2B looks severe until you check the June 2026 level against a year earlier: RMB 36.5B against RMB 36.6B, essentially unchanged. Contract liabilities built through calendar 2025, up 31.6% in H1 and another 34.4% in H2, then gave it all back.
So was the H2 2025 peak the anomaly, driven by a wave of energy-storage pre-orders or customers locking terms ahead of policy changes? Or is the H1 2026 drawdown genuine softening in forward commitments? Both readings are plausible. The filing discloses contract liabilities only as a consolidated "sale of goods" figure, with no product, segment, or geographic split, so the disclosure doesn't distinguish them.
Reading the three together
Demand pulling product through the system at the rate 95% utilization implies would show up as growing dispatches and stable or rising contract liabilities, as customers competed for allocation. Neither appears.
The financial position, though, rules out distress cleanly. CATL reported RMB 372B in monetary funds at mid-year and RMB 60.2B in operating cash flow. Gross margins remain positive: 21.16% domestic, 29.97% overseas (CATL H1 2026 filing, geographic revenue note). And the inventory provision moved barely at all against a doubling in finished goods, which means management's assessment, audited, is that this product is saleable at or above its carrying value. Expecting to sell below carrying value would require a larger write-down under the applicable standards.
The carrying cost is also close to nil. As covered in Issue #10's supplier-credit analysis, CATL runs a negative cash conversion cycle: it collects from customers before it pays suppliers, financing operations through extended payment terms upstream (combined payables of RMB 355.5B in H1 2026, 96.6% of it in bank acceptance bills). The working capital embedded in that RMB 47.6B of finished goods was substantially funded by CATL's own supply chain.
My reading, and this is inference rather than disclosure: CATL is running at high utilization for reasons other than current-period delivery requirements. Unit cost absorption at scale, share positioning, capacity qualification against future contracts, some mix of the three. The surplus that produces is being held rather than pushed into the market at a lower price.
The procurement read
Surplus accumulating at a financially constrained producer eventually reaches the market as distressed tonnage, because the producer needs cash. Surplus accumulating at a producer with RMB 372B in monetary funds, positive margins, and a demonstrated tolerance for running full regardless of dispatch rates does not. It stays where it is until the holder chooses otherwise.
That has a direct effect on what buyers see. RMB 47.6B in finished goods held off the market is supply that does not appear in spot or near-term contract negotiations, and CATL can keep it there indefinitely on current cash flows. The absence of write-downs reinforces the point: CATL's auditors are certifying that the product moves at or above carrying value. A producer that accumulates rather than discounts is showing you where its price floor sits.
The contract-liability decline is the more useful signal for buyers. A 26% drop in prepayments, even allowing for the elevated comparison base, indicates a forward order book that is not tightening. Customers are not bidding harder for allocation. That may not show up as lower cell prices, because CATL has no need to concede on price, but it plausibly shows up elsewhere: delivery flexibility, warranty terms, payment schedules, minimum-commitment thresholds. Teams negotiating with CATL in H2 2026 should be pressing those dimensions rather than waiting for movement in headline $/kWh.
Four variables would change this reading:
- A second consecutive half of inventory doubling. RMB 95B in finished goods by December would test even this balance sheet.
- A meaningful increase in the inventory provision, signalling that CATL's accountants no longer believe the product moves at cost.
- A reversal in operating cash flow, which would make the carrying cost real rather than supply-chain-subsidised.
- A decision to price for share rather than hold. The hardest to observe from filings, and the most consequential. Procurement teams would see it in quote behaviour months before it surfaced in a balance sheet.
Three gaps worth naming
Product mix. Whether this inventory is LFP cells, NMC packs, storage containers, or a blend changes the interpretation materially. Storage inventory could reflect project-delivery timing and milestone-based acceptance. EV cell inventory could reflect OEM schedule slippage. The filing gives no basis for distinguishing, and I have not found an English-language source that does.
Geographic distribution. Overseas revenue grew 42% to RMB 87.1B in H1 2026, 31.5% of the total (CATL H1 2026 filing). Longer logistics chains on export orders could contribute to finished-goods accumulation, product built for shipment but not yet dispatched. No geographic inventory data is reported, and as covered in Issue #8, the overseas revenue line itself blends China-origin exports, German production, and Hungarian assembly without identifying the contribution of each.
Utilization denominator. CATL is the only major Chinese cell producer that reports utilization at all, which is what makes 94.86% worth having. The filing gives 525 GWh of battery-system capacity as the denominator but does not say whether that base includes facilities still ramping or commissioning, or how capacity is defined for lines running multiple formats. Selective denominator definition would flatter the ratio. I have no evidence CATL does this. The methodology simply isn't disclosed at a level that lets me rule it out.
CATL can hold RMB 47.6B in finished goods indefinitely on its current cash position, so the inventory buildup should not be read as a signal of coming price concessions. The contract-liability decline is the more actionable number: a forward order book that has stopped tightening gives buyers leverage on terms, flexibility, and commitment structure rather than on headline price.
- Industry-wide production-sales gap: January–July 2026 output exceeded reported sales by 122.3 GWh — already more than double the 55.1 GWh gap for all of 2025 — though the CABIA series measures reported production minus reported sales, not physical inventory.
- BYD's comparable signals: BYD's H1 2026 results are scheduled for August 28, and its inventory composition and contract liabilities will provide the first direct comparison to CATL's accumulation pattern at the second-largest producer.
- September tax basis break: The 2% consumption tax taking effect September 1 will change the invoice basis for cell and pack transactions, making pre- and post-September finished-goods and dispatch figures non-comparable without explicit tax adjustment.
- Tier-2 utilization opacity: CATL remains the only major Chinese producer disclosing utilization, so the single largest driver of unit cost across the rest of the industry — including REPT, Gotion, and EVE — remains invisible in public filings.