Asia-Pacific Private Markets Outlook 2027 — The Derived Demand Test

This is a forward-looking outlook written from a mid-2026 vantage point about a year that has not begun. Every statement below is a scenario or a monitoring question. None should be read as a forecast.

Asia-Pacific enters 2027 with its largest single exposure resolving somewhere else.

The region's AI exposure runs through the physical supply chain — semiconductors, memory, advanced packaging, components, servers, cooling and networking. The Asia-Pacific Investment Report 2024 describes this as owning a constraint rather than a thesis: revenue depends on the activity happening at all, not on which layer of the value chain captures the economics.

That is a strong position and a derived one. The demand for hardware derives from the capital expenditure decisions of a small number of very large buyers, and those buyers are not in the region. Their disclosures — capex guidance, depreciation, useful-life assumptions, utilisation — are the leading indicator for regional order books, and they are published quarterly, free, on a different continent.

This is analytically unusual and worth stating plainly: a regional investor's most important series is a foreign company's filing.

The capacity cycle amplifies whatever happens. Semiconductor and component capacity is lumpy: it requires large capital commitments with long lead times, so it arrives in increments and typically in excess. Capacity commissioned in response to shortage meets demand that may have moderated by the time it arrives. That is the standard cycle, it has not been repealed, and it means the resolution — in either direction — will be sharper than the underlying demand change.

Beneath that, the region's genuinely durable development continues: domestic capital growth, which is slow, structural and the thing most likely to matter over a decade.

Watching a foreign filing

The practical consequence of derived demand is a specific monitoring discipline, and it is worth setting out because it inverts the usual approach.

The usual approach to a regional allocation is to monitor regional indicators: regional growth, regional policy, regional funding, regional exchanges.

For the region's hardware exposure, those are lagging. By the time regional order books, revenue or capacity utilisation move, the decision that caused the move was taken quarters earlier in a capital expenditure meeting elsewhere.

The leading series, in the order they move:

• Capex guidance from the largest infrastructure buyers. Disclosed quarterly, forward-looking by construction, and the earliest available signal. A change in guidance precedes a change in orders by roughly two to four quarters.
• Useful-life assumptions in the same filings. Extending assumed asset life reduces annual depreciation and raises reported earnings without changing economics. A change here is disclosed and is one of the clearer indications of pressure on a return case — and it typically precedes a capex reduction.
• Semiconductor equipment billings, published free and monthly by SEMI. Equipment ships before capacity produces, so this leads capacity additions by roughly a year.
• Regional company order books and guidance, which is where the effect finally becomes visible regionally.

The discipline this implies: an investor holding regional hardware exposure should be reading the buyers' filings, not only the suppliers'. Funding announcements and regional commentary tell you what people believe; the buyers' capex line tells you what they have committed to.

The most important number for a Korean or Taiwanese supplier in 2027 is a line in an American company's cash flow statement. That is uncomfortable and it is the actual structure of the exposure.

Why the capacity cycle amplifies

The semiconductor capacity cycle is well understood and consistently underestimated in its severity, and 2027 is a point where it matters.

Why capacity is lumpy:

• Fabrication facilities require enormous capital and take years to build. Capacity cannot be added in small increments.
• The decision is made on expected demand at a horizon of several years, which is exactly the forecast most likely to be wrong.
• Once committed, the capacity arrives regardless. A facility part-built when demand moderates is generally completed, because the sunk cost is large and the marginal cost of completion is smaller than the write-off.

The resulting dynamic:

• Shortage produces high prices and high returns, which justifies capacity additions across the industry simultaneously.
• Everyone adds at once, because everyone sees the same shortage and the same returns.
• The additions arrive together, typically after the shortage has eased.
• Excess capacity collapses pricing, because the marginal cost of running a built facility is low and it will be run.

The amplification is the key point for 2027. If AI demand moderates, it meets capacity commissioned during the shortage. The price effect is therefore larger than the demand change, because supply is rising into falling demand rather than adjusting to it.

And the reverse holds. If demand continues, the capacity added may still be insufficient, since lead times mean the response lags — which extends the shortage and the pricing power.

Neither outcome is forecastable and both are observable in advance. SEMI's equipment billings show capacity being added roughly a year before it produces. Comparing that against buyers' capex guidance is the single most informative exercise available, and both series are free.

Domestic capital: the slow durable thing

Against the cyclical exposure above, the region's most durable development continues and deserves separate assessment.

What has been growing, per the Asia-Pacific Investment Report 2025: domestic institutional investors, corporate venture arms, family offices and government-backed funds, taking a rising share of regional private market funding.

Why it matters more than the amount suggests:

• Stability. Domestic capital has local liabilities and does not withdraw during a global risk-off episode. This directly addresses the vulnerability the 2015 and 2022 regional reports document, and the 2023 India and Latin America reports show being realised elsewhere.
• Different preferences. Domestic capital favours businesses serving local markets, where local knowledge is an advantage — categories foreign capital applying imported frameworks systematically underweighted.
• Longer horizons. Family office and corporate capital is not bound by a ten-year fund life, which expands what is fundable rather than merely changing who funds it.
• It develops alongside the exit route. The 2021 and 2023 India reports describe domestic listing bringing domestic institutions into venture-backed companies as public shareholders, which supports their participation as private investors. The public and private markets develop together.

What to watch, and it is not funding totals:

• Pension and insurance allocation to private markets by regional institution, disclosed in annual reports.
• Domestic investor participation share in venture and private equity rounds.
• Local currency bond market depth, which ADB's AsianBondsOnline tracks free and which is the foundation for domestic institutional investment generally.

The competitive implication for foreign investors intensifies. Capital alone is no longer a differentiated offer in most of the region's major markets. Participation increasingly requires cross-border expertise, access to international markets or specialist sector knowledge.

The unbundling question is settled

The Asia-Pacific Private Markets Outlook 2026 argues that the regional unit is the wrong level of analysis. By 2027 the evidence for that is unusually strong, and the useful discussion has moved on.

The evidence, accumulated across the archive's twelve-year regional sequence:

• 2015 — the dollar and commodity axes divided the region into four positions with opposite outcomes.
• 2020 — health response, fiscal capacity and economic composition produced dispersion within the region exceeding dispersion between regions.
• 2021 — one market repriced on regulation while the rest followed the global monetary cycle.
• 2022 — the commodity axis inverted, moving economies between quadrants without anything changing about them.
• 2025 — markets bifurcated along different axes, each revealing different selection criteria.

Five distinct demonstrations. The principle is not in serious doubt.

So the 2027 question is implementation, which the 2026 outlook addresses with a six-bucket middle version: North Asia manufacturing, Japan capital efficiency, India domestic demand, Southeast Asia, China, and Australia. Each has a coherent driver and each is investable through a manageable number of relationships.

The minimum viable version remains the same and is worth restating because most institutions will land here: hold a regional allocation but know what it contains. Check the country and sector weights, understand which drivers they expose you to, and treat the result as a position rather than as diversification.

What would falsify the whole framing, per the 2026 outlook: sustained convergence in returns and growth across the region's markets, with cross-sectional dispersion narrowing toward the dispersion within a single market. That is computable from free data and has not happened through 2026.

India and Japan: still uncorrelated

The two markets the 2023 and 2025 regional reports identify as diverging from the regional cycle continue to do so, and the reason is worth restating because it is the strongest practical argument for the unbundled approach.

Most of the region is exposed to a common set of drivers: the dollar cycle, trade and technology supply chain politics, and Chinese demand. Those drivers correlate across markets, which limits how much diversification a regional allocation actually provides.

India's drivers are largely domestic — demographics, internal consumption, public digital infrastructure, and a deepening domestic exit route. The India Venture Capital Outlook 2026 sets out the case and its principal caveat: the structural case is strong and substantially priced, which makes the entry point the binding question rather than the allocation.

Japan's driver is capital efficiency — a return source that does not require growth, arising from the gap between current and potential use of existing assets. The Japan Investment Report 2019 explains why the gap persisted for decades and why the catalyst had to be institutional. A thesis that does not require growth is uncorrelated with trade, technology and commodity cycles by construction.

Why this is the strongest argument for unbundling: a regional allocation averages these together with markets driven by entirely different variables. Preserving genuinely uncorrelated return drivers is what diversification is supposed to mean and rarely does, and a regional index actively destroys it.

The caveats for both remain live. India's valuation premium is the question; Japan's governance reform must continue converting into capital actually returned rather than plans published. Both are observable, both are free, and neither is a forecast.

What would change the picture

Evidence the derived demand is holding: capex guidance from the largest infrastructure buyers holding or rising; useful-life assumptions unchanged rather than extended; SEMI equipment billings continuing to rise.

Evidence it is not: capex guidance decelerating; useful-life extensions appearing in filings; equipment billings rolling over. The last of these leads regional revenue by roughly a year.

Evidence the capacity cycle is turning adverse: equipment billings rising while buyer capex guidance falls. That divergence is the specific signature of supply arriving into moderating demand, and it is visible from two free series.

Evidence domestic capital continues to deepen: regional institutional allocation to private markets rising; local currency bond markets deepening per ADB data; domestic investor share of rounds increasing.

Evidence India's structural case is converting: infrastructure and manufacturing execution against announced programmes; domestic listing depth improving; and on valuation, whether the premium narrows through earnings growth rather than through price decline.

Evidence Japan's thesis is real: the share of listed companies below book value falling; cross-shareholding unwinding continuing; and capital returned — buybacks and dividends — rising rather than plans being published.

Evidence the unbundling framing is wrong: cross-sectional dispersion of returns and growth across the region narrowing sustainably toward within-market levels.

What an allocator could act on

Read the buyers' filings, not only the suppliers'. For regional hardware exposure the leading indicator is a foreign company's capex guidance, published quarterly and free. Regional order books and revenue are lagging by two to four quarters, which means regional monitoring alone puts you last in the information chain.

Track the ordering: capex guidance, then equipment billings, then regional revenue. Each leads the next by roughly two to four quarters. SEMI publishes equipment billings free and monthly, which gives roughly a year of warning before capacity produces.

Watch for the divergence signature. Equipment billings rising while buyer capex guidance falls is supply arriving into moderating demand — the specific condition the capacity cycle amplifies. It is visible from two free series and it is the single most informative comparison available for this exposure.

Assess domestic capital on allocation data, not funding totals. Funding totals move for external reasons in both directions, as 2021 and 2023 demonstrated across every emerging market in this archive. Regional institutional allocation to private markets is the series that would show the structure changing, and ADB tracks the bond market foundation for it free.

Hold India and Japan as separate positions. Their drivers — domestic demand and capital efficiency — are uncorrelated with the trade, technology and dollar cycles that drive the rest of the region. A regional allocation averages exactly the diversification you were trying to obtain.

Implement the unbundled approach at whatever scale you can support. The full country-level version is expensive. The six-bucket middle version is workable. The minimum is holding a regional allocation while knowing what it contains — a regional position that has been examined is a legitimate choice; one that has not is an unexamined bet with a reassuring label.

Separate the constraint position from the thesis position. A supplier whose revenue depends on the activity happening is differently exposed from a company whose value depends on a particular layer capturing the economics. Both are AI exposure and they resolve differently in every scenario.