Hong Kong Grade A Office Rents Rise for a Fourth Quarter, but Non-Core Districts Keep Falling

Hong Kong's Grade A office market tightened again in the third quarter, with vacancy down and overall rents up for a fourth consecutive quarter. The same report shows non-core districts still sliding. For landlords and leasing teams, the gap is no longer about location alone — it is about execution.

What actually happened this quarter

Start with the numbers.

Third-quarter leasing volume for Hong Kong Grade A offices came to 1.2m sq ft, down 6% quarter on quarter, but year-to-date take-up reached 3.4m sq ft, up 6% year on year. Strong net absorption pushed the vacancy rate down to 15.5%, a fall of 0.8 percentage points quarter on quarter and the largest single-quarter drop since the second quarter of 2015.

Rents followed. Overall office rents rose 3.2% quarter on quarter, a fourth consecutive quarterly increase, and are up 6.7% year to date. Central rose 15.8% year to date, with Grade A1 offices there up 24.3% — the sharpest rise since the fourth quarter of 2010.

Then there are the non-core districts, where rents kept falling.

Retail tells a similar story. Vacancy on major streets in core districts fell 0.5 percentage points quarter on quarter to 6%, the second-lowest level since the fourth quarter of 2019, while shop rents rose for a 17th consecutive quarter, up 2.7% year to date.

In short: in the same city, in the same year, in the same report, core and peripheral assets are moving in opposite directions.

Where the gap shows up in execution

Operators in non-core districts cannot control the location. They can control two things: how fast they respond, and how accurate their books are.

The common pattern is that leasing leads arrive from platforms, agents or referrals from existing tenants, land in someone's WeChat, and sit there. Nobody picks them up that day; two days later the prospect is touring the building next door. Someone asks about floor area, rent-free periods or whether the layout can be changed, and a slow reply loses them.

Then there is the billing. A single building or a specialised market can hold dozens or hundreds of billable units, with monthly rent escalations, expiring rent-free periods, utility meter readings and management fees all calculated layer by layer in spreadsheets. Collections run on phone calls. By the time someone notices a tenant is three months in arrears, the quarter is over.

Neither problem is strategic. Both are operational. Yet together they determine year-end occupancy and cash collection.

Where peers are putting AI

Two approaches are visible.

One puts AI at the front end: mass-producing leasing content, buying traffic, building a founder's personal brand, widening the top of the funnel. The logic is sound, but if the back end cannot keep up, the traffic is wasted.

The other embeds AI directly into existing ledgers and customer management, letting it assign the work itself: flagging leads for follow-up on schedule, generating bills on schedule, and escalating arrears collection by number of days overdue.

A useful reference point comes from other industries. According to Southern Finance Omnimedia, third-party figures show 406 large-model tenders in the financial sector in the first half of 2026, up 110% year on year, with banks accounting for close to half of both the number and the value. Those projects sit in customer acquisition and marketing, credit due diligence, loan approval and disbursement, and post-loan risk control — all labour-intensive, repetitive work. The logic holds: the more standardised, data-dense and people-heavy the task, the better suited it is to being handed over. Rent bills and utility meter readings fit that description exactly.

How to put it to work

There is no need to replace an entire system at once. Standardise two things first:

• Every leasing lead must have an owner and a specific time for the next follow-up.
• Every billable unit must have a bill that can be calculated automatically under the contract rules.

Work of this kind can sit on SHEYU AIMARKET. It puts asset and lease control, leasing CRM and a contract billing engine into a single ledger, with an AI steward handling leasing follow-up, bill collection and meter-based billing. When a follow-up date passes without action, it prompts. Rent escalations, rent-free periods and utility readings in the contract generate bills according to the rules, and collection is escalated in tiers according to how far a tenant is behind. For a building in a non-core district, getting these two things running smoothly will not improve occupancy out of thin air — but it will at least stop revenue leaking away unnoticed.

Whether rents rise is up to the market. Whether the books are clear and tenants are answered is up to you.

When the market splits, the cash flow that can be defended is usually the cash flow clawed back from these two stages.