On 21 October 2026, Google, Bain and Temasek gave private banks and family offices three figures that do not belong in one line. Southeast Asian digital sales are expected to rise 17 per cent to US$368 billion this year. The same assessment put first-half private funding at US$21 billion, against US$11 billion for the whole of 2025, and attributed that difference to two data-centre deals. It also cited IPO proceeds of US$8.5 billion. Sales describe what digital businesses across the region are expected to take in. The funding comparison describes where private capital went, and the attribution says the jump is concentrated in infrastructure transactions. The proceeds figure describes cash raised from public investors. Private wealth investors across Asia-Pacific, whether the account is booked in Singapore, Hong Kong or another regional centre, should keep those three lines apart when they set 2027 growth exposure.
The sales forecast is the operating case, and it should be used only as that. A rise of 17 per cent to US$368 billion, if companies actually deliver it, is a reason to look at listed platforms and at private businesses whose revenues move with digital commerce, payments and related services. It is not a reason to increase every growth line in the mandate by the same proportion. Discretionary and advisory accounts at UBS Global Wealth Management, HSBC Global Private Banking, DBS Private Bank, Bank of Singapore, Julius Baer, LGT, Pictet, Lombard Odier, Citi Private Bank, J.P. Morgan Private Bank and Standard Chartered Private Bank can all hold this exposure. So can multi-family offices and single family offices that buy the same companies directly or through funds. The decision that belongs to the investor is the weight, the benchmark and the cash set aside to meet obligations elsewhere in the book. A regional sales total does not choose those terms.
The funding line needs a harder reading. US$21 billion in the first half, set against US$11 billion for all of 2025, will appear in manager presentations as a rebound. The assessment attributes the difference to two data-centre deals. Two infrastructure transactions can lift a regional funding total without reopening the venture and growth market that private wealth uses for earlier-stage companies. A data-centre asset is land, power and long-dated contracted cash flow. A venture commitment is a drawdown into corporate equity, often blind or only partly identified at the time of signing, with cash returned only when a later round, a sale or a listing occurs. The fee, the call schedule and the ability to stop do not move from one to the other because both can be labelled private capital. Investors who set a 2027 venture budget from the US$21 billion figure are sizing an equity commitment off an infrastructure print. That is a category error.
Two infrastructure transactions can lift a regional funding total without reopening the venture market that private wealth uses for earlier-stage companies.
IPO proceeds of US$8.5 billion are a liquidity fact of a different kind. The assessment cites that sum. It shows that public investors have provided capital to issuers in the market being described. It does not show that a co-investment on a family office balance sheet, or a line inside a fund, has a buyer at a price the holder will accept. A proceeds total mixes issuers, venues and selling shareholders. For a 2027 policy, the useful cut is between exposure that is already listed and can be reduced on an exchange calendar, and exposure that is committed and can be reduced only on the fund's terms. Those terms include the investment period, any extension, any recycling of distributions and the notice required to decline a call where the documents allow it. None of that is supplied by a regional proceeds figure. Borrowing against a listed holding to meet a venture call is a separate credit decision, and it should be capped on its own, not treated as a free bridge between the sleeves.
Fees have to be read the same way, vehicle by vehicle. Listed exposure inside a discretionary mandate carries the mandate fee agreed for that account. Advisory exposure carries the advice fee and dealing costs. A venture fund carries the management fee and performance fee stated in its partnership agreement, whether on committed or invested capital, and any hurdle that has been negotiated. A data-centre fund, or a co-investment beside an operator, carries a longer asset life and distributions tied to the asset's leases and debt, not to the path of regional digital sales. The US$368 billion forecast does not set any of those terms. For 2027, the question to each manager is which of the three lines the vehicle serves, what fee is charged on it, and how quickly the position can be reduced. An answer that recites the sales forecast, or the US$21 billion funding print, has not addressed the fee or the exit.
The governance consequence is a 2027 investment policy with three maxima, not one digital bucket. Operating exposure can lean on the sales forecast, and should be reviewed if the companies held do not share in that 17 per cent. Data-centre exposure should be underwritten as infrastructure. Venture exposure should be sized from the manager's pipeline and unused commitment capacity, not from a half-year total already attributed to two deals. The mandate has to record that split before commitments for the coming year are signed.