Corporate venture capital in Singapore: how to structure a CVC arm
Every large company eventually faces the same question: keep buying startups' products, keep acquiring them outright, or start investing in them early. Corporate venture capital is the third answer, and Singapore — with its startup density, treaty network and purpose-built fund vehicles — has become the natural base for Asian corporate venture programmes. The structuring decision, made early and often made casually, determines most of what follows: tax, governance, licensing, and whether the programme survives its first leadership change.
The short answer
A corporate building a venture arm in Singapore chooses between three structures: investing off the balance sheet (simple, but co-mingles venture risk with the core business), a dedicated fund vehicle — typically a VCC or a Singapore limited partnership — which ring-fences the programme and can access fund tax treatment, or LP positions in third-party VC funds, which buy exposure and deal flow without building a team. Mature programmes usually end up with the middle option, and the VCC fits the corporate case unusually well: its variable capital suits an evergreen mandate, its umbrella sub-funds can separate theses or business units, and outside co-investors can be admitted into one sub-fund later without rebuilding the structure. The two questions that need early, careful answers are tax (a single-corporate captive fund and the non-SFO third-party-capital expectation sit awkwardly together) and licensing (whose money will the team actually manage?).
CVC is not institutional VC in a corporate badge
The differences drive the structuring, so they are worth stating plainly:
- Objective. An institutional VC answers only to financial return. A CVC also answers to strategy — a window on technology, a pipeline of future acquisitions, early access to capabilities the parent will need. Good programmes are honest about the blend; programmes that pretend to be purely financial while reporting into a strategy function confuse everyone, founders included.
- Capital and horizon. A VC fund raises committed capital for a fixed life, typically ten years. A CVC usually runs on the parent's capital, evergreen — recycled as positions exit, expanded when conviction grows, cut when the cycle turns. That difference alone disqualifies structures built around fixed fund lives.
- Governance. A VC's LPs are passive by contract. A CVC's single "LP" is the parent's board, which is anything but — and managing that relationship is the real job.
The three routes, honestly compared
| Route | What it looks like | Where it wins | Where it hurts |
|---|---|---|---|
| Balance-sheet investing | The operating company (or a plain subsidiary) holds the stakes directly | Fast to start; no new entities; fine for a handful of strategic positions | Venture volatility lands in group accounts; no fund tax treatment; hard to incentivise a team; messy to unwind or scale |
| Dedicated fund vehicle (VCC or SG limited partnership) | A governed fund the parent capitalises; professional team appointed to run it | Ring-fenced risk; fund-grade governance and reporting; tax incentives possible; co-investors can join; survives reorganisations | Real setup and running costs; licensing and incentive conditions need designing, not assuming |
| LP in third-party funds | Commitments to one or more external VC funds, sometimes with co-invest rights | Instant exposure and deal flow; no team to build; a common first step | Strategy filtered through someone else's mandate; limited information rights; no capability built in-house |
The pattern we see most: start with LP positions to learn the market, add direct balance-sheet deals as conviction forms, then consolidate into a dedicated vehicle once the book is big enough that governance, tax and team incentives all argue for one. Nothing wrong with taking the steps in order — the mistake is reaching step three and structuring it as an afterthought.
Why the VCC fits the corporate case
Four properties line up with how corporates actually run venture programmes:
Variable capital matches an evergreen mandate. The parent injects capital when the programme expands and takes distributions at net asset value when positions exit — no capital-reduction procedure, no artificial fund life forcing exits. For a programme meant to outlive any single budget cycle, that is the load-bearing feature.
Sub-funds separate theses cleanly. An umbrella VCC can hold a climate-tech sub-fund, a fintech sub-fund and a logistics sub-fund — or one sub-fund per sponsoring business unit — each ring-fenced by statute, sharing one board, administrator and auditor. Business units get attribution; the group gets one vehicle.
Co-investors can come later. Corporates increasingly open a sleeve to outside capital — a sovereign partner, a customer, a family office. With an umbrella, that is a new sub-fund with its own investors, not a restructuring. (It does change the licensing analysis — below.)
The venture book leaves the operating balance sheet. Portfolio volatility, follow-on obligations and eventual write-offs sit in a governed fund the group consolidates deliberately, rather than surfacing as noise in operating results.
The Singapore limited partnership is the credible alternative where the programme genuinely is closed-end — a fixed pool, deployed and harvested — especially since 13OA extended the fund tax exemption to Singapore LPs. If the honest answer to "when does this fund end?" is "never, we hope," the VCC is the better-shaped tool.
Tax: where captive funds need real advice
The fund tax incentives were designed for funds with investors, and a wholly-captive corporate fund tests their edges. Three points to get right early:
- The third-party-capital expectation. On the non-SFO track, MAS expects 13O/13OA/13U funds to have third-party investors or a bona fide intention to raise them. A fund that will only ever hold the parent's money should take advice on how that condition applies before an award is assumed in the model. Where outside co-investors are part of the plan, the position is cleaner — and worth sequencing deliberately.
- The thresholds. Where a structure qualifies, 13U applies from S$50 million and 13O/13OA from S$5 million in designated investments on the non-SFO track, with the 2026 closed-end treatment available if the vehicle genuinely has a fixed life. Section 13H — the approved-venture-company scheme — is the VC-specific incentive worth assessing alongside. Confirm current conditions with MAS and IRAS; these schemes moved as recently as July 2026.
- Without an award, the vehicle is an ordinary Singapore taxpayer — which, given Singapore's absence of capital gains tax and its treaty network, is often still a perfectly workable outcome for an equity-holding venture book. Model both cases rather than treating the incentive as a given.
Licensing: whose money will the team manage?
The licensing question turns on one fact. A team investing only the parent group's own capital generally sits outside the regime that governs managing other people's money — confirm the current analysis, but captives are not what fund-management licensing exists to police. The moment outside investors join, the activity needs a licensed footing. From there the routes are: a venture capital fund manager (VCFM) licence — the lighter regime built for VC strategies; a full A/I LFMC licence; or appointing an established MAS-licensed manager to the vehicle, with the corporate's team in a defined sub-adviser role — the licensed manager runs the fund and carries the regulatory responsibility. Note the route that no longer exists: the RFMC regime was repealed on 1 August 2024, and any older planning built on it needs redoing.
The traps that actually sink CVC programmes
- Strategy drift at leadership change. A new CEO inherits a venture book they didn't build. A ring-fenced vehicle with a mandate document survives this; a scatter of balance-sheet positions rarely does.
- Conflicts with the parent. Portfolio companies sell to, compete with, or get acquired by the parent. Decide the information-barrier and conflicts rules before the first term sheet, not during it.
- Team economics. Venture talent expects carry-like upside; corporate pay bands don't offer it. A fund vehicle can carry a genuine incentive scheme; a cost centre cannot — and the 2026 announcement on taxing fund managers' performance profits makes Singapore-based structures more attractive for exactly this.
- Valuation into group accounts. Agree the valuation policy, frequency and audit approach with the group CFO on day one. A professional fund administrator striking NAV independently removes most of the friction.
- Follow-on discipline. Evergreen capital makes it easy to keep feeding losers. A reserve policy set at the vehicle level is the cheap protection.
Is a dedicated CVC vehicle right for you?
A rough screen: if the group expects to hold more than a handful of positions, wants a named team with real incentives, can commit capital across more than one budget cycle, and can imagine outside co-investors one day — structure the vehicle now. If the honest ambition is two or three strategic stakes watched by the corp-dev team, stay on the balance sheet and revisit in a year. Both are respectable answers; the expensive mistake is running the first ambition on the second structure.
Designing a corporate venture programme?
Tell us the parent's situation — how much capital, which theses, whether outside co-investors are in the picture, and where the team will sit. We'll walk you through the balance-sheet-versus-vehicle decision, the tax and licensing questions a captive fund raises, and connect you with MAS-licensed CMS fund managers where a regulated structure is the right fit.
Speak to a specialist →What is corporate venture capital (CVC)?
Corporate venture capital is a company investing in startups for a blend of strategic and financial return — access to technology, markets and talent that matter to the parent's business, alongside the investment upside. It differs from institutional VC in ownership (one corporate parent rather than many LPs), in horizon (often evergreen rather than a ten-year fund life), and in objective (a purely financial VC answers only to returns; a CVC also answers to the parent's strategy).
How do companies structure a CVC arm in Singapore?
Three routes. Direct balance-sheet investing — the operating company holds the stakes itself, simplest but co-mingles venture risk with the core business. A dedicated fund vehicle — a VCC or Singapore limited partnership that ring-fences the programme, professionalises governance and can access fund tax incentives. Or LP positions in third-party VC funds — exposure and deal flow without building a team. Many corporates sequence through all three as the programme matures.
Why use a VCC for corporate venture capital?
Because a CVC mandate is often evergreen, and the VCC's variable capital suits that: the parent can inject and withdraw capital at net asset value without capital-reduction procedures. An umbrella VCC also lets a corporate run ring-fenced sub-funds per thesis or business unit, keep the venture book off the operating company's balance sheet in a governed vehicle, and admit outside co-investors into a specific sub-fund later without restructuring.
Can a captive corporate fund get the 13O or 13U tax exemption?
Sometimes — but a single-corporate captive needs care. The non-SFO track of the 13O/13OA/13U schemes expects capital from third-party investors, or a bona fide intention to raise it; a fund holding only the parent's money should take advice on how MAS views that condition before relying on an award. Where the structure qualifies, 13U applies from S$50 million and 13O/13OA from S$5 million on the non-SFO track, and Section 13H remains the VC-specific incentive for approved venture companies. Confirm current conditions with MAS and IRAS.
Does a corporate venture arm need a fund management licence?
It depends on whose money is managed. A team investing only the parent group's own capital generally sits outside the licensing net that applies to managing third-party money, though the analysis should be confirmed against the current MAS position. The moment outside investors join, the management activity needs a licensed footing — a venture capital fund manager (VCFM) licence, an A/I LFMC, or appointing an existing MAS-licensed manager to the vehicle. The RFMC route no longer exists; it was repealed on 1 August 2024.