API Down 22.5%, CDMO Up 41%: Apeloa Trades an Ugly Quarter for a Structural Rebuild

Apeloa's Q3 2026 report (22 October): nine-month revenue of RMB 6.835B (-11.97%), net profit RMB 614M (-12.37%) and non-recurring net profit RMB 568M (-12.22%). Yet CDMO revenue reached RMB 2.05B (+21.38%), with Q3 alone at RMB 642M (+40.96%), while API was deliberately shrunk to RMB 4.028B (-22.53%). After the results, 185 institutions met management, which disclosed more than RMB 7.8B of commercial backlog to be delivered within three years and 1,715 total projects (+38%).

Apeloa published its third-quarter 2026 report on 22 October and within a week had hosted 185 institutions — the busiest healthcare name of the reporting window and second across all sectors. The headline numbers are not flattering: nine-month revenue of RMB 6.835B, down 11.97%; net profit attributable to shareholders of RMB 614M, down 12.37%; non-recurring net profit of RMB 568M, down 12.22%. In the third quarter alone, revenue was RMB 2.011B (-13.28%) and net profit RMB 91.3M (-33.47%).

Read only the headline numbers, though, and you miss what the report is actually about.

A quarter where the total goes down and the structure goes up

The divergence across three businesses is unusual. CDMO revenue for the nine months reached RMB 2.05B, up 21.38%, at a 40.39% gross margin; in Q3 alone it was RMB 642M, up 40.96%. API revenue fell 22.53% to RMB 4.028B — management's framing is a deliberate exit from low-margin business, layered on top of florfenicol prices hitting historic lows. Finished-drug revenue fell 14.48% to RMB 716M at a 56.16% gross margin.

Two non-operating drags were named explicitly in the investor call: foreign exchange, worth about RMB 63M, and staff compensation, about RMB 50M. The first is a revaluation of dollar assets; the second is the price of hiring for CDMO. Neither means the business got worse, but both landed on the income statement.

That produces a tension worth flagging: non-recurring profit fell 12.22% over nine months, yet management maintained a full-year target of 10% growth in non-recurring net profit. Meeting it requires the fourth quarter to carry a substantial share of the annual figure while three conditions hold simultaneously — no further FX deterioration, commercial CDMO orders delivered on schedule, and no further slide in API prices. It is the single line in this report most in need of later verification, and the hardest to separate from a claim of deliberate contraction.

Four numbers in the CDMO story: 7.8B, 1,715, +38%, 35.52%

The first is RMB 7.8B: commercial backlog to be delivered over the next three years, above RMB 7.8 billion. The second is 1,715: total CDMO projects, up 38% year on year, with management describing a project funnel that has taken shape and continues to widen. The third is the same +38% read from the other direction — the mouth of the funnel is expanding far faster than revenue, the classic signature of reserves moving ahead of conversion.

The fourth, 35.52%, is the most easily misread. Q3 CDMO gross margin was 35.52%, below the 40.39% nine-month figure. The cause is not price: management said Q3 carried a higher share of clinical-stage deliveries and a lower share of commercial ones, and expects commercial delivery to exceed clinical from Q4 into next year. Clinical batches are small and process-scale amortisation is incomplete, so margins are structurally lower than on commercial campaigns.

On price, the wording matters: order pricing has stayed broadly stable, negotiated on a basis of reasonable cost plus reasonable margin. In other words, there is no price war at this tier. Competition happens on whether you can take the order and whether you can deliver it on time — consistent with what first-tier CDMOs are saying, and a reminder that the industry's dividing line in 2026 is capacity and delivery record, not price.

The capability list is worth recording: ADCs, peptides, cyclic peptides, molecular glues and oligonucleotides. Three of the five are chemical macromolecules, two are conjugation and hard-synthesis directions — a portfolio that extends Apeloa's existing API process chemistry rather than jumping across modalities.

A capex decision worth close to RMB 2B

Early in 2026 the company began a new capacity programme: five new workshops, two retrofits, roughly RMB 2B of investment. R&D headcount is above 1,500 and is expected to reach 2,000 in the CDMO segment within one to two years.

Set that against peers. Jiuzhou Pharmaceutical has capped 2026 fixed-asset investment at RMB 1B, expects a similar level in 2027, and is prioritising small-molecule and peptide commercial lines in Suzhou, an ADC line in Taizhou and oligonucleotide capability. WuXi AppTec's Changzhou expansion is a RMB 10.1B, 800-tonne API programme. Apeloa's RMB 2B sits in the middle of that axis — a level of capital intensity no defensive API company would carry.

More interesting is where the money comes from. Apeloa's cash flow historically came from commodity APIs; today API revenue is being cut by 22.53% on purpose so that workshops, people and capex can move to CDMO. It is a strategy of paying for the new ticket with the old business's cash flow, and the cost is a report that looks bad for a few quarters.

Why 'API steps aside' is an industry signal

Line up three companies and the same curve appears. Apeloa deliberately shrinks low-margin API and converts workshops to CDMO. Jiuzhou disclosed that improving CDMO business has created short-term bottlenecks at certain plants, so a workshop previously making anti-infective API is being converted to commercial innovative-drug production — which directly reduced specialty API sales. Asymchem keeps small molecules as the base while new modalities open the increment.

Different routes, same substance: China's second-tier CDMOs are trading the plants, process know-how and EHS capability accumulated in the API era for order eligibility in the new-modality era. The advantage is a low starting cost and smooth customer migration. The cost is a structurally lower gross margin for a while, because what gets cut is the steady margin on mature products and what arrives is low-margin clinical batches plus the fixed costs of ramping capacity.

There is a counter-intuitive conclusion hiding here about Section 232. Under the US tariff schedule, qualifying generics and commodity APIs carry 0% while patented medicines and their APIs carry 100%. Commodity API exports therefore have a safer tariff standing than patented originator APIs. The real pressure on Chinese API makers in 2026 is not the tariff but the price cycle — overcapacity, channel destocking and historic lows in products such as florfenicol. Jiuzhou's language was that the antibiotics cycle is under pressure, many facilities are near break-even and the industry is clearing slowly.

Apeloa's choice is therefore not being pushed by tariffs. It is being pushed by the relative slope of two curves: the bottom of the commodity API cycle, and the order visibility of CDMO.

Editorial note: the questions institutions ask have changed

What 185 institutions pressed on after the report was not CDMO growth rates but four harder items: the revenue share of clinical-stage projects in Q3, the delivery outlook for Q4 and next year, order pricing trends and gross-margin trajectory. All four point at one thing — the efficiency with which backlog converts into revenue and profit.

This aligns with Goldman Sachs' 25 October framework: CDMO visibility increasingly depends on late-stage and commercial exposure rather than early-stage financing. For a company like Apeloa, still weighted toward clinical batches and migrating toward commercial, the valuation anchor therefore shifts from project-count growth to the commercial share of deliveries.

Three checkpoints are verifiable. First, whether the commercial share of Q4 deliveries rises as promised — that decides whether 35.52% was a quarterly wobble or a structural step down. Second, the conversion pace of the RMB 7.8B commercial backlog — spread evenly over three years that is roughly RMB 2.6B a year, against RMB 2.05B of CDMO revenue in the first nine months of 2026, so the slope has to show up in 2027. Third, whether utilisation can climb past break-even within 12 to 18 months of the RMB 2B of capacity coming online — the same question every second-tier CDMO will face in 2027.