On 27 August, Drewry’s World Container Index closed at USD 4,473 per 40ft box, down 1% on the week but 111% above its level a year earlier. The same week, COSCO SHIPPING Holdings published its interim results and disclosed average revenue per TEU on international services of USD 1,180.68, against USD 1,205.95 a year earlier.
The two readings sit at opposite ends of the same pricing mechanism. Only one of them records what COSCO actually collected.
Table 1 | The Pricing Gap
Year-on-year change across three different measures of container pricing
| Measurement | Source | Year on year |
|---|---|---|
| WCI spot composite, 27 August | Drewry | +111% |
| CCFI half-year average, 1,249 points | COSCO interim report | −0.26% |
| COSCO realised revenue per TEU, international | COSCO interim report | −2.10% |
The middle line matters most, because it covers the same six months as the bottom one. COSCO’s report records the China Containerized Freight Index averaging 1,249 points across the half, marginally below last year, while the second quarter ran 19.5% above the first. These three are not interchangeable measures of one price. The WCI assesses spot rates on selected East–West corridors in dollars per 40ft box on 27 August; the CCFI carries different route weightings and incorporates contractual as well as spot pricing; and COSCO’s figure is carrier-specific realised revenue shaped by trade mix, customer mix, contract cover and surcharges. Read as triangulation rather than comparison, they show something precise: market rates strengthened through the half and now stand far above year-ago spot levels, yet COSCO’s half-year realised revenue per box still came in below last year’s. Part I of this series argued that rates had returned without pricing power but could only show the spot half. COSCO supplies the half that matters.
The revenue line that did not move
Container shipping revenue reached RMB 107.30 billion, up 2.38%. Within that figure, revenue from supply chain services other than ocean transport reached RMB 24.09 billion, up 11.61%. Subtract the second from the first and the ocean freight business produced RMB 83.21 billion, against RMB 83.22 billion a year earlier.
Flat, therefore, to within a rounding error, on volumes that grew 7.52% to 14.28 million TEU. On the reported segment numbers, essentially all of the division’s revenue growth came from supply chain services outside ocean transport. Almost none of it came from carrying boxes across oceans.
What the exchange rate hides, and what it does not
One correction belongs here, and it cuts against the direction most readers would assume. The average renminbi-dollar rate moved from 7.1793 to 6.9043, so the dollar presentation flatters the renminbi accounts, and the liner division grows 6.5% in dollars against 2.38% in renminbi. The per-TEU line is not exempt from that effect, because COSCO’s dollar route figures are translations of renminbi amounts at those same average rates rather than natively dollar measures.
Reconstruct the same line in the reporting currency, from the disclosed renminbi route revenue and volumes, and it falls about 5.8% against the 2.1% shown in dollars. That reconstruction is ours rather than the company’s, and it is not a second reading of the freight rate. It measures accounting yield per TEU in renminbi. International container freight is customarily priced in dollars, so USD 1,180.68 against USD 1,205.95 stays the better measure of what shippers actually paid, while the renminbi figure shows what the same revenue became in COSCO’s reporting currency. Both fell. The currency move decided how far apart the two declines look.
For the COSCO SHIPPING Lines brand alone the reported dollar decline runs steeper still, USD 1,234.85 against USD 1,279.33. For context, the company’s realised international rate across the whole of 2025 was USD 1,187.46, so the first half of 2026 came in below even that full-year average.
Where the money went missing, trade by trade
The trade-level table is where the container shipping pricing power argument stops being rhetorical.
Table 2 | Where Volume and Revenue Diverged
COSCO container volume and dollar revenue by trade, H1 2026 against H1 2025
| Trade | Volume | Revenue (USD) | Implied change in revenue per TEU |
|---|---|---|---|
| Transpacific | +9.72% | +1.45% | −7.5% |
| Asia–Europe incl. Mediterranean | +12.44% | +6.56% | −5.2% |
| Intra-Asia incl. Australia | +5.34% | +11.18% | +5.5% |
| Other international | −0.13% | −3.22% | −3.1% |
| Mainland China | +9.97% | +9.33% | −0.6% |
| Total | +7.52% | +4.92% | −2.4% |
The Transpacific offers the sharpest case, since COSCO moved nearly a tenth more cargo and collected only 1.45% more dollars for the privilege. Part II described ZIM concentrating its growth in the same ocean: Pacific volumes rose 20.3% year on year, while its company-wide average realised rate climbed 21.4% sequentially from the first quarter. Both accounts hold, and both can be compared on the same sequential basis. COSCO’s first-quarter report discloses international revenue per TEU of USD 1,110.36 against USD 1,311.09 a year earlier, which lets the second quarter be derived the same way as the profit line: roughly USD 1,247, some 12.3% above the first quarter. ZIM’s company-wide rate rose 21.4% across the same two quarters. The more telling comparison sits inside COSCO’s own accounts, since the CCFI’s second quarter ran 19.5% above its first while the realised rate moved about two thirds as far.
Within COSCO’s reported trade mix, intra-Asia was the only major segment showing positive implied unit-revenue growth, on dollar revenue up 11.18% against 5.34% more cargo. Drewry’s subsequent August readings point in the same direction rather than against it, since its Intra-Asia Container Index rose 6% to USD 1,028 on 13 August, 6% again to USD 1,091 on 20 August, and another 10% to USD 1,199 on 27 August. Where COSCO’s first-half unit revenue actually improved, it improved inside Asia.
The quarter hiding inside the half
COSCO does not publish standalone second-quarter financial statements, so subtracting its first-quarter figures from the half-year figures on the same accounting basis produces one.
Table 3 | The Quarter Inside the Half
Derived by subtracting first-quarter from half-year disclosures on the same accounting basis; COSCO publishes no standalone Q2 financial statements
| Q2 2026 (derived) | Q2 2025 | Change | |
|---|---|---|---|
| Revenue | ~RMB 60.12bn | RMB 51.14bn | +17.6% |
| Attributable net profit | ~RMB 7.54bn | RMB 5.84bn | +29.1% |
That sits against a half in which attributable profit fell 23.5% to RMB 13.42 billion and earnings per share dropped from RMB 1.12 to RMB 0.88. The comparison base matters here, because the first quarter of 2025 was exceptional. A weak half containing a strong quarter is precisely the shape Part I identified at OOCL, and it now repeats at the parent.
The broader industry shows the same second-quarter recovery. Sea-Intelligence reported combined second-quarter revenue of USD 43.4 billion, up 15.9%, with eight carriers posting combined EBIT of USD 2.69 billion against USD 1.70 billion, a rise of 58.2%. Deriving COSCO’s quarterly volume the same way gives roughly 7.36 million TEU against 6.80 million, growth near 8.3%. Sea-Intelligence had no COSCO figure when it published, so inserting that derived number into its comparison would put the group behind OOCL at 8.8% and ahead of CMA CGM at 6.0% and ONE at 2.9%, against a then-reported carrier average of 4.7%. OOCL is already consolidated within the COSCO group, so those two are not independent observations.
What the quarterly split actually shows
The half-year figures conceal two very different quarters. In the first, volume rose 6.70% while international revenue per TEU fell 15.3% to USD 1,110.36. In the second, volume rose about 8.3% and the derived rate recovered about 13.1% to roughly USD 1,247. The half nets out at 7.52% more cargo for 2.1% less per box.
The second quarter was good by every measure the company reports. Deriving the liner segment the same way gives revenue up 17.9% year on year, EBIT up 21.8% and the segment margin slightly wider at 14.25% against 13.78%. Any account of this half that leaves that out is not describing the business.
What survives the quarterly split is therefore narrower and harder to dismiss. Even after the recovery, COSCO’s realised international rate sits below the USD 1,374.86 it averaged across 2024, and between the two quarters it moved about two thirds as far as the CCFI did. Whether that gap reflects contract lag, trade mix or something else is not something the disclosures settle.
Costs ran 2.4 times faster than revenue
Group operating costs rose 6.24% against revenue growth of 2.59%, a multiple of 2.4, and the consolidated gross margin fell 2.80 points to 18.34%. Within liner the multiple reaches 2.8, since costs rose 6.62% against revenue growth of 2.38%, with voyage costs up 10.46% and vessel costs up 9.12%. Container shipping EBIT dropped to RMB 15.19 billion from RMB 21.51 billion, while the segment margin collapsed from 20.52% to 14.15%.
Below the operating line the movement runs the same direction. Net finance income of RMB 2.10 billion a year earlier became a finance cost of RMB 356.7 million, a swing of RMB 2.45 billion. The two main drivers were RMB 1.11 billion less interest income and a RMB 1.46 billion deterioration in foreign exchange, from a RMB 653 million net gain to a RMB 810 million net loss. Total investment income held near RMB 3.00 billion, including RMB 2.82 billion from associates and joint ventures, and amounted to roughly a sixth of group pre-tax profit, a material contribution for an asset-heavy liner group.
Meanwhile operating cash flow fell 9.49% to RMB 23.33 billion, net cash used in investing activities increased to RMB 11.91 billion from RMB 10.50 billion, and construction in progress rose 30.99% to RMB 21.83 billion, almost entirely accounted for by vessels under construction. The interim dividend fell to RMB 0.43 per share from RMB 0.56, while the payout ratio stayed near half of attributable profit. Cash generation down, vessel investment accelerating, payout ratio still close to 50%. The configuration is tighter than a year ago, but first-half cash flow shows no funding stress.
The demand nobody admits borrowing
Part I set out a dispute. Four carriers described part of the second-quarter surge as cargo pulled forward, while Maersk’s Vincent Clerc rejected the framing and told analysts he saw no sign of pre-buying anywhere in the numbers.
Sea-Intelligence, working from port-level data, supplies a date. Its assessment attributes the North America West Coast surge to shippers front-loading ahead of a hard 24 July 2026 deadline for new permanent trade tariffs following earlier judicial reversals. Major West Coast ports handled 3.7 million TEU of laden imports, up 7.1%, and the intra-quarter profile is sharper still, because after a soft April, May total volumes rose 12.7% and laden imports 19.8%.
Two qualifications are owed. Sea-Intelligence notes that the May figure flatters against a depressed 2025 base, itself distorted by that year’s tariff deadlines. Furthermore the movement was regional as well as temporal, since Los Angeles rose 13.8% and Long Beach 12.0% while the Pacific Northwest fell across the board, Vancouver by 4.8%, the Northwest Seaport Alliance by 9.2% and Prince Rupert by 12.4%. Neither qualification moves the date. The deadline passed on 24 July, so the third quarter becomes the first full quarter on the far side of it. COSCO’s Transpacific growth of 9.72% and ZIM’s Pacific growth of 20.3% were recorded against the same broader front-loading environment, although neither carrier discloses enough cargo-timing data to attribute those increases directly to the tariff deadline.
What actually holds the market up
Part I argued that effective capacity rather than vessel count decided the quarter, while conceding the mechanism resisted measurement. It now has a number. Sea-Intelligence calculates that 5.0% of global deep-sea capacity currently sits absorbed by vessel delays against a 2011 to 2019 average of 2.2%, which means the market is missing 1.7 million TEU, a fleet that would rank eighth in the world. The excess over the pre-pandemic baseline runs to 1.0 million TEU, roughly the entire capacity of HMM.
The operational conditions behind that absorption worsened in July. Schedule reliability fell 6.1 points to 56.4%, the lowest of 2026, down 8.8 points year on year, while the average delay for late arrivals rose to 6.06 days, the highest since January 2024. Maersk led at 73.7% and Hapag-Lloyd followed at 69.3%, the only carriers above 60%, so the Gemini reliability claim survived the industry’s worst reliability month of 2026. Wan Hai finished last at 29.8%, down 24.4 points, which deserves reading alongside its August commitment to 42 vessels.
Drewry names the same mechanism from the other side, describing congestion at Asian gateways, weather disruption, Panama Canal constraints and the gradual return of Suez services as producing a market where additional capacity fails to deliver smoother operations. That represents the strongest case against this analysis, and it deserves stating plainly. It also confirms the narrower point flagged in Part I: the gradual return of Suez services has not yet translated into smoother operations or an obvious release of effective capacity. Congestion and other disruptions continue to absorb part of what additional nominal capacity should otherwise deliver.
Rates have already turned
The freight market did not wait for the argument to conclude. Drewry’s World Container Index climbed through the first three weeks of August and then slipped.
Table 4 | The August Rate Turn
Drewry World Container Index, USD per 40ft container
| Week ending | Composite | Shanghai–LA | Shanghai–NY | Shanghai–Rotterdam | Shanghai–Genoa |
|---|---|---|---|---|---|
| 6 August | $4,297 | $4,653 | $5,506 | ||
| 13 August | $4,339 | $6,244 | $8,706 | $4,425 | $5,080 |
| 20 August | $4,526 | $6,802 | $9,507 | $4,401 | $4,955 |
| 27 August | $4,473 | $6,818 | $9,333 | $4,287 | $4,866 |
Shanghai–Rotterdam has slipped for three weeks running in the table above, to USD 4,287. Across the major East–West trades, 45 blank sailings are expected from week 36 to week 40, a 6% cancellation rate with 94% of scheduled sailings operating, against 49 blanks and 7% in Drewry’s forecast a fortnight earlier. The two head-haul trades point in opposite directions: Drewry records Transpacific blank sailings for the following week falling to four from seven, indicating increased capacity, while Asia–Europe blankings rose to four from two, reflecting constrained capacity. Congestion at Shanghai lengthened the average vessel wait to 96 hours from 35 a week earlier.
None of this constitutes collapse, because the index remains at more than double last year’s level and Drewry describes demand as resilient. The point is narrower. Several of the mechanisms supporting rates do not originate in stronger carrier pricing: delay absorption and congestion constrain effective capacity, blank sailings manage nominal capacity, the tariff deadline has expired and the Suez return remains incomplete. Their persistence differs, but none by itself demonstrates commercial pricing strength. That is exactly why container shipping pricing power remains the variable to watch.
The Gulf, with dates attached
The war began on 28 February 2026. On 17 June the United States and Iran signed a fourteen-point memorandum of understanding brokered by Pakistan, which formalised the ceasefire, ended the U.S. naval blockade of Iranian ports and reopened the Strait of Hormuz toll-free for a sixty-day negotiating window. CMB.TECH’s regulated disclosure records what that meant at sea, since Hormuz transits had collapsed from roughly 120 daily crossings to about 10 between March and mid-June, then recovered rapidly to more than 40.
It did not hold. Iran struck commercial shipping in the strait, the United States retaliated, and Washington declared the memorandum over on 8 July. CMB.TECH records the same re-escalation at sea, with crossings subsequently falling back towards 20 a day. Tehran suspended its commitments on 18 July, and the sixty-day negotiating period expired on 17 August without a final agreement. On 26 August the Revolutionary Guard declared that the strait will not reopen until the United States returns to the memorandum, while Iran and Oman negotiate a governance arrangement for the waterway. CMB.TECH separately reports Bab el-Mandeb transits at multi-month lows on renewed Houthi activity.
So the scarcity was not merely borrowed. It was borrowed against a ceasefire that Washington declared dead within three weeks, on a strait that Iran still declares closed as this is published. Meanwhile the demand signal weakened as well. China’s official manufacturing PMI fell to 49.2 in July from 50.3 in June, dropping below the 50 threshold for the first time since February. More directly relevant to container trade, the new export order index slipped from 50.1 to 49.6 while the import index fell from 49.6 to 47.5. Those national indicators point to weaker trade momentum, although they cannot establish how much of it will fall specifically on the Asia–Europe and Transpacific lanes where COSCO grew volume hardest.
The ports told a quieter and more interesting story
COSCO offers the cleanest available test of diversification, because its ports arm reports separately. COSCO SHIPPING Ports handled 80,157,047 TEU, up 7.9%, on revenue of USD 905.3 million, up 12.3%, with profit attributable to equity holders up 28.5% to USD 233.7 million. Profit up 28.5% at the subsidiary, therefore, against a 23.5% fall at the parent.
Yet the primary filing will not support the simple version of that contrast. Gross profit rose 9.3% against revenue growth of 12.3%, compressing the gross margin by 0.7 points to 26.5%, while total terminal profit actually fell 3.1% to USD 234.2 million. The geographic split was severe: Chinese terminals rose 14.8% to USD 212.1 million while overseas terminals fell 61.4% to USD 22.1 million, with the Mediterranean and Middle East down 48.4% and newly commissioned terminals still absorbing ramp-up losses. The bridge to the headline 28.5% increase in attributable profit sits elsewhere in the accounts: net other operating income rose from USD 4.3 million to USD 63.4 million, including a USD 53.7 million reversal of a provision recognised in prior years for a contractual obligation, released after the group reached agreement with the counterparty.
The throughput split shows where the volume growth sits. Terminals under CSP control grew 2.5%, while non-controlling terminals grew 9.4% and now carry 78.9% of total throughput. The pattern holds on an ownership-adjusted basis too: equity throughput at controlled terminals grew 2.6% against 10.3% at non-controlling terminals, which supplied 59.4% of equity throughput. The company confirms the logic, stating it will pursue controlling stakes in strategic hubs while taking minority stakes in key gateway ports. The diversification thesis survives, then, but in a more qualified form than Part I could assert: ports grew faster than liner earnings, yet the volume acceleration sits largely outside controlled terminals, China carried the terminal-profit base, and overseas terminal profit fell by nearly two thirds.
Abu Dhabi lost 44% to the war that raised the rates
The geopolitics that lifted freight rates simultaneously cut terminal throughput, and both appear in one group’s accounts with opposite signs. CSP Abu Dhabi fell 44.3% to 442,977 TEU from 795,758 TEU, which the company attributes directly to Middle East tensions. That amounts to 352,781 boxes lost to the same conflict that was raising rates on COSCO’s ships.
Piraeus fell 2.9% to 1,995,150 TEU, attributed to softening Mediterranean demand and adverse weather rather than to the Gulf. Chancay rose 68.2% to 201,773 TEU on a network that reached three main lines and five feeder services. Overseas terminals as a whole grew 18.0%.
The promotional account describes Abu Dhabi as having strengthened its role as a regional hub, and the primary document both supports and undercuts that reading within a few pages. CSP names Abu Dhabi among the key hubs whose service capacity it intends to raise, in the very release reporting that the terminal lost 44% of its volume, then commits to refining contingency plans in response to the Middle East situation. Strategic weight rose while throughput halved. Both statements are true, although only one of them is a number.
The order
On the same day the board approved these results, COSCO Asset Management signed for eighteen container ships across two contracts. China CSSC Holdings disclosed that its wholly owned Shanghai Waigaoqiao Shipbuilding, with China Shipbuilding Trading, contracted twelve 21,700 TEU LNG dual-fuel vessels for USD 2.688 billion, payable in dollars, delivering between 2028 and 2030, under Hong Kong law with arbitration in Hong Kong. A second contract covers six 3,200 TEU wide-beam feeders at the fellow CSSC yard Huangpu Wenchong for RMB 2.039 billion, about USD 300 million, taking the August package to eighteen ships and 279,600 TEU of nominal capacity.
Two details deserve attention. The larger vessels are 21,700 TEU rather than the 22,000 TEU class name in circulation, so the twelve of them account for 260,400 of the package’s 279,600 TEU. And at roughly USD 224 million each they cost more per ship than January’s 18,000 TEU order yet about 8% less per slot when both are converted at the half-year average rate, consistent with one of the economic arguments for building larger.
Where the ships are going
Note the intended deployment. COSCO says the larger vessels are meant for routes including Far East–Northwest Europe. On announcement day, Shanghai–Rotterdam spot had been sliding for three weeks in the table above, while across the half just reported COSCO carried 12.44% more cargo on Asia–Europe for 6.56% more dollars.
Table 5 | The Ordering Cycle Continues
COSCO container newbuilding orders placed in 2026, confirmed by disclosures
| Date | Vessels | Yard | Value | Capacity |
|---|---|---|---|---|
| 13 January | 12 × 18,000 TEU LNG plus 6 × 3,000 TEU | Jiangnan, COSCO Zhoushan | ~USD 2.7bn | 234,000 TEU |
| 29 April | 12 × 13,600 TEU LNG (OOCL) | Hudong-Zhonghua | USD 2.22bn | 163,200 TEU |
| 28 August | 12 × 21,700 LNG plus 6 × 3,200 | Waigaoqiao, Huangpu Wenchong | ~USD 2.99bn | 279,600 TEU |
| Confirmed 2026 total | 48 vessels | ~USD 7.9bn | 676,800 TEU |
Set that against the scarcity holding the market up. The excess capacity currently absorbed by delays above the pre-pandemic norm is 1.0 million TEU, so one carrier has contracted future nominal capacity equivalent to roughly two thirds of that in eight months, against an operating fleet of 3.66 million TEU. Citing MDS Transmodal, Hapag-Lloyd puts the global orderbook at 12.2 million TEU by end-June, up from 10.3 million at the close of 2025, lifting the orderbook-to-fleet ratio from 31.8% to 38.5% in six months.
The 2027 arithmetic that nobody actually disputes
What makes the order remarkable is not that it happened, but that COSCO’s own report contains the case against it. The company records global fleet capacity growing 2.1% during the first half, then cites full-year forecasts of around 4.2% supply growth against demand growth of 2.5% to 3.0%. The forecasts in its own report therefore set supply growth above demand growth, before the company commits nearly USD 3 billion to eighteen ships on the same day.
The external numbers run harder still. Industry forecasts in Hapag-Lloyd’s investor report show net capacity growth of 8.1% against container volume growth of 3.2% for 2027, and a gap of almost five percentage points makes commercial execution alone a difficult balancing mechanism. CMB.TECH goes further and forecasts container demand falling 5.8% in billion TEU-miles during 2027 should Red Sea routing normalise. Meanwhile the macro backdrop offers little offset, since the World Bank projects global growth moderating from 2.9% in 2025 to 2.5% in 2026, while the IMF has cut its 2026 forecast to 3% citing Middle East tensions.
China provides the one genuinely supportive figure. Goods trade reached RMB 25.47 trillion in the first half, up 16.9%, with exports up 13.4% and imports up 22.1%. That is a favourable backdrop for COSCO’s 7.52% volume growth. It does not explain why revenue grew materially more slowly than volume, leaving realised revenue per box lower across the half, which returns us to the container shipping pricing power problem by a different route.
The owner who did the arithmetic and declined
Here the strategic structure becomes visible, and it is not flattering. Linerlytica’s Tan Hua Joo put the mechanism plainly earlier this year: none of the carriers will surrender market share, so the ordering continues. The shape resembles a prisoner’s dilemma, in that restraint would improve the collective outcome while unilateral restraint risks surrendering network scale and share to rivals who do not exercise it. Ordering therefore stays individually rational even where collective ordering worsens the industry’s future supply balance.
CMB.TECH demonstrates what happens when a player sits outside that matrix. During its Q2 call, analyst Climent Molins observed that containership owners had been ordering heavily in recent months and asked why CMB.TECH had not. Alexander Saverys replied that the company had not seen an opportunity interesting enough to move on, while continuing to monitor the market.
His results release explains the reasoning without ambiguity. CMB.TECH expects container trade to grow 3.0% in 2026 measured in billion TEU-miles, yet forecasts fleet growth exceeding demand on the strength of an orderbook representing approximately 38% of the existing fleet, which independently corroborates the 38.5% from MDS Transmodal. Moreover the company states that any eventual normalisation of Red Sea routing would reduce tonne-mile demand and increase effective vessel supply, producing a forecast 5.8% decline in container demand for 2027. Its own four containerships plus one newbuilding sit on charters of ten to fifteen years, consistent with an owner prioritising contracted cash flow over spot-market exposure.
The same logic, applied to his own market
Saverys applies identical logic to his core market. He describes the crude tanker orderbook, now 620 VLCCs and Suezmaxes and the heaviest newbuilding investment in fifty years, as the central medium-term risk, then sells into strength. CMB.TECH had sold nine VLCCs and four Suezmaxes across 2026, and the two VLCCs and one Suezmax delivered in the second quarter alone generated USD 127.4 million of gains, while its VLCCs earned USD 126,790 a day.
Why the two decisions differ
The comparison needs one qualification, and the qualification is the entire point. CMB.TECH is a tonnage provider with no liner network to defend, so it can decline to order because it has no share to lose. COSCO faces a far higher strategic cost from declining, which is precisely the mechanism Tan Hua Joo described. The difference between the two decisions is not analysis. It is structure, and structure does not respond to argument.
The market moved before the results, not after
Part I found that the most useful number of the quarter was one no carrier published, namely the share price reaction to a specific event. Here the reaction had largely happened before the event. COSCO’s Hong Kong line rose from HK$15.49 on 14 August to HK$16.44 on 17 August, a 6.1% jump in one session, and reached HK$17.95 by 21 August, up roughly 16% in a week, all before the board met on 28 August. On Monday 31 August, the first full session after the results and the order, the shares closed at HK$17.22, up 0.12% from Friday, on volume of 39.6 million shares, heavier than Friday’s. Most of the August repricing therefore occurred ahead of the results, not after them; on the day the market had them in full, the stock barely moved.
That timing is the tell, though not its cause. By the time the half-year report and the eighteen-ship order landed, the equity had already made most of its August move. The rate recovery documented throughout this analysis is the market backdrop, but the price action alone cannot establish what investors had already priced in. What Monday establishes is narrower and firmer: with the results and the order fully public, and on heavier volume than Friday, the stock added almost nothing.
Sell-side expectations remain more cautious than the August run. The consensus on the Hong Kong line is neutral, with the average twelve-month target around HK$16, below Monday’s HK$17.22 close. Morgan Stanley sits at sell with a HK$10.90 target and Goldman Sachs at sell with HK$10.40, against JPMorgan at buy with HK$21.00, while Citi cut the stock to hold on 26 August with a HK$17.40 target. The dispersion is the point: the desks disagree sharply on fair value, yet the aggregate target still sits below the price the market reached before the results landed.
The other structural fact is the state. On 3 March 2026 COSCO SHIPPING Holdings confirmed completion of the gratuitous transfer of 2,610,063,089 A shares to China COSCO SHIPPING Corporation, lifting the group and its concert parties’ combined holding to about 45.78%. No consideration, no takeover offer, no market purchase. The direct controlling shareholder changed, but the ultimate controller did not: it remained the State-owned Assets Supervision and Administration Commission of the State Council. The interim report’s risk section lists geopolitical risk first, immediately followed by port investment risk, placing side by side the two exposures that intersect most visibly in Abu Dhabi.
What the three parts add up to
The first quarter of 2026 established the pattern: volumes returned and pricing did not. Part I found the pattern surviving a rate recovery. Part II found it surviving a company that beat its quarter. COSCO complicates the pattern before confirming it. Its second quarter was genuinely strong, with volume up 8.3%, the realised international rate up 13.1% and liner EBIT up 21.8%. Yet the rate moved only two thirds as far as the index, it remains below where it stood in 2024, and several of the mechanisms supporting it are contingent rather than commercial: a strait Iran still declares closed, a tariff deadline that expired on 24 July, congestion absorbing effective capacity, and a Suez return that remains incomplete. Container shipping pricing power did come back in the second quarter of 2026. It came back on loan, and COSCO answered it with nearly USD 3 billion of ships arriving from 2028.