Who buys next? The question on investors' minds at SuperReturn Asia 2026

Across four days at SuperReturn Asia 2026, one question connected discussions spanning AI infrastructure, private equity, private credit and Asian markets: Who will buy next, and what price will they pay?

Drawing on insights shared on stage, Juliana Bambaci, Founder and CEO of JB Content, examines what the answers could mean for LPs and private markets managers.

Four days of panels in Singapore circled the same concern: who takes the loss if data centres, fund stakes, loans and Asian companies sell below their stated value?

A US public pension’s head of private markets closed a fundraising panel in Singapore with a warning: “If you’re six or seven years into a fund and you’re just now thinking about how I’m going to get liquidity, you’ve already lost.”

An Asian technology investor described the opposite habit: “When we underwrite a deal, we actually start with underwriting what the exit looks like … from day one.”

That habit set the tone at SuperReturn Asia 2026. Investors still wanted exposure to AI and Asia’s growth. But on panel after panel, the conversation started with the exit, and with what the next buyer would pay.

Those prices tell the story so far. One buyer of Asian fund stakes said those stakes sell at more than 50% below their stated value. Listed US funds that hold private loans traded at about 74 cents on the dollar this year. Each discount shows how far the next buyer trusts the seller’s numbers.

If you’re six or seven years into a fund and you’re just now thinking about how I’m going to get liquidity, you’ve already lost.
A US public pension’s head of private markets

This article draws on 24 of the conference’s more than 90 sessions. The panels chosen speak to four decisions, one per section: whether to own or lend to AI and energy assets, how to get cash out of private equity funds, how to hold private credit and which Asian markets have local buyers.

Other sessions at the conference covered real estate, sport as an asset class, private wealth and specific energy technologies including nuclear, batteries and solar.

Each section ends with one takeaway for limited partners (LPs), the institutions and wealthy families that invest in private funds, and one for the managers who run them.

AI and energy: Many infrastructure investors prefer to lend to data centres

The AI build-out needs more money than its owners have.

A multi-asset investment firm estimated that the eight biggest AI spenders—a group that includes cloud companies, model developers and chipmakers—need about $3 trillion through 2028. Their own cash flow covers about half. That leaves at least $1.5 trillion to come from someone else.

Several infrastructure investors said they prefer lending that money to owning the data centres. The reason is a mismatch inside each one. The building and its power connection last for decades. The chips inside lose most of their value within a few years, and no one can say who will buy them—or the building—10 years from now.

An emerging-markets infrastructure firm said it could not predict how chips and cooling needs would change “over the course of the five years, 10 years down the road”, so it would “like to be a credit investor”.

A global bank’s alternatives head said its firm preferred “providing power to data centers” to owning the computing equipment inside them. A Gulf sovereign wealth fund added that nobody would “take a 30-year mortgage out” to buy a MacBook.

Power is the safer bet because demand for it does not depend solely on AI. “Power is a binary constraint on data centers,” an Asia-Pacific infrastructure lender said: without a grid connection, a finished building earns nothing.

In emerging markets, data centres drive only about 5% of the growth in electricity demand, compared with approximately 20% in the US, an infrastructure fund head said. Ordinary economic growth drives most of the remainder.

Price adds to the caution. One global infrastructure investor said it had paid an average of 12 to 14 times earnings for data-centre businesses. Today, they sell for more than 20 times, so “a lot more has to go right” just to stay on plan.

There are two catches. Lenders still depend on the same few borrowers: bonds and loans for AI “all triangulate” to a few AI laboratories and about five hyperscalers—the largest cloud companies—the bank’s alternatives head said.

The biggest gains may also go to owners of AI applications. One venture investor said 98% of the value placed on AI companies today sits in infrastructure, but in the web and mobile eras, most of the value ended up in applications.

• For LPs: Add up your exposure to the largest AI companies across venture, infrastructure and credit funds. The total can far exceed your stated AI allocation.

• For managers: Lead with power and contracted assets. For an equity pitch, name who will buy the finished asset.

Private equity: Funds return less cash than they used to

Private funds are sending back less cash.

In good years, they pay out close to 20% of the value of their holdings each year, one large institutional investor said. For the past three to five years, they have paid out about half that amount.

The hold-up is price. Managers value their own holdings, and the industry calls those valuations marks. Buyers will not pay them.

“If you have a valuation that nobody wants to pay, then it’s probably not the right mark,” said an investor who buys fund stakes at deep discounts.

The gap is widest in Asia. Investors who want to leave a fund early sell their stakes to specialist buyers through what the industry calls the secondaries market. One such buyer said Asian stakes sell at more than 50% below their stated value. In the US, the discount is about 10%.

If you have a valuation that nobody wants to pay, then it’s probably not the right mark.
An investor who buys fund stakes in the secondaries market

Managers have responded by constructing their own exits. In a continuation fund, a manager moves a company from an old fund into a new one that it also runs. New investors provide cash, which pays out the existing investors who want to leave.

Secondary transactions of all kinds reached approximately $250 billion last year. A US secondaries specialist called continuation funds private equity’s “fourth exit path”, after selling to a company, selling to another fund or listing on a stock exchange.

The push for cash has critics. The industry measures that cash with DPI, short for distributions to paid-in capital: the cash a fund returns for each dollar investors put in. A US growth investor said managers “sell their best assets too early because they need the DPI”.

An Asian buyout firm takes the other side. “DPI is the new IRR,” its co-head said, meaning that cash returned now counts for more than the internal rate of return—the annual return figure funds previously led with.

The firm has a separate committee to decide exits because investors can “fall in love with the companies”.

• For LPs: Treat the secondaries discount as the market’s price for a fund’s holdings. Ask managers who decides when to sell.

• For managers: Plan the buyer when you invest. Sell part of a winner when its price rises and retain the rest.

Private credit: Funds promised quick exits on loans that take years to repay

Private credit funds lend directly to companies, and the exit comes built in.

“In credit, the exit is repayment or refinancing,” a panel moderator said. A venture lender said its liquidity works differently from that of an equity fund: “It’s contractual, it’s not hoped for.”

That certainty is what the infrastructure investors above are buying when they lend.

The trouble started when managers sold that cash flow to individual investors with a faster exit attached.

From 2022 to 2024, individuals poured money into two kinds of funds. Business development companies, or BDCs, are US funds that hold private loans, and many trade on the stock market. Semi-liquid funds allow investors to withdraw money each quarter, up to a set limit.

This year, more investors wanted to withdraw money than those limits allowed. Listed BDCs traded at about 74 cents per dollar of stated value.

A digital-assets manager explained the mechanics: a fund that allows 5% of its money to leave each quarter cannot pay 20% of its investors if they want to withdraw at once. “Liquidity does not just magically get created,” the manager said.

The loans themselves have held up so far. Rated BDCs still value their loans at close to 99% of cost, a rating analyst said. Defaults have risen, but “not enough to justify the overall outflows”, another analyst said.

Institutions kept buying. A global private credit founder said they invested more money in the first half of the year than during any half-year period since at least 2017.

Liquidity does not just magically get created.
A digital-assets manager

The bear case is that the underlying loans are next.

“Either the public market is wrong or the private marks are,” a moderator said.

Crowding has reduced spreads—the additional interest lenders earn above base rates—from approximately seven percentage points in 2023 to just over five, a US credit manager said. Around 20% of new loans have no covenants, the terms that allow a lender to intervene early.

Software is another concern. Software companies make up approximately one-quarter of rated BDC assets, and the analyst who cited the 99% figure identified AI’s effect on them as the central question.

Asia has avoided all three problems so far, an Asian growth lender said. Outside Australia, few funds offer regular withdrawals, relatively few borrowers are software companies and lenders have not weakened their terms to win deals.

Indian direct loans continue to pay three to four percentage points more than comparable Western loans, one credit investor said.

• For LPs: Check that a fund’s withdrawal terms reflect how quickly its loans repay, and ask how much of the portfolio is exposed to software.

• For managers: Do not promise quarterly exits on loans that take years to repay.

Asia: with foreign money pulling back, exits depend on local buyers

Across Asia, foreign money has pulled back, so the next buyer is usually local. Each market is only as strong as its domestic buyer base. Investors now build their Asian exposure country by country.

India has the deepest pool. Indian households invest approximately $100 billion in the stock market each year, a mid-market buyout manager said. That flow allows private equity firms to sell through public markets rather than relying exclusively on other funds.

China has already shifted. More than 95% of its venture and private equity capital is now denominated in renminbi, China’s currency, a venture investor said. The number of China funds raising money in foreign currencies fell from approximately 1,100 in 2022 to 97 in 2025, according to another investor.

Hong Kong’s IPO market has revived, and many Chinese AI and robotics companies list there. One renminbi venture fund said it was now selling shares because it saw “bubbles” in parts of those sectors.

A listing does not pay early investors immediately. They must wait through a lock-up period, usually lasting several months, before they can sell. An open IPO window “doesn’t necessarily mean that the distribution window” is open, a deep-technology investor said.

Where local buyers are scarce, exits stall.

“Southeast Asia has disappointed on DPI for a number of decades,” a regional buyout founder said.

Limited exit routes are one reason investors increasingly want control. A buyout gives the investor control of the company and the timing of its sale. Buyouts increased from about 35% of Asian private equity between 2019 and 2023 to more than half, according to an investor that backs numerous private funds.

Southeast Asia has disappointed on DPI for a number of decades.
A regional buyout founder

Two points cut the other way.

A global fund manager argued that a fund invested in one country can go years without an exit, while a fund spread across multiple countries can continue returning cash.

Not all foreign capital has left, either. Gulf investors continue to invest in China, and they “can afford to not take sides”, one speaker said.

• For LPs: For each country, ask who will buy at exit: a local company, another fund or public-market investors.

• For managers: Name the local buyer in your pitch and demonstrate the rights, including control, that will allow you to sell to it.

Four things to watch in 2027

Each of these debates faces a practical test next year—a point at which the next buyer must name a price.

1. Data-center leases. Rating agencies count about $1 trillion of lease commitments by the largest cloud companies, much of it off their balance sheets. If AI spending slows, those leases, and the loans built on them, will feel it first.
2. Asian secondary discounts. If the gap with the US narrows, buyers have started to trust Asian managers’ marks.
3. Loan fund withdrawals. If individual investors keep asking for more money back than funds allow, the BDC discount will stay.
4. Hong Kong lockups. Lockups on many of 2026’s AI and robotics listings end in 2027. The selling will show whether public buyers still pay listing prices.

Where those tests go badly, the loss will land on whoever bought last.

NB: This is my attempt to identify the stories and narratives connecting four days of outstanding presentations. Any errors of interpretation are mine alone, and I hope this does justice to the quality of thinking in the room.