Foreign Capital Flows into Chinese Startups: What the Data Shows

It is tempting to read the decline in Western capital into Chinese startups purely as a sentiment shift — investors got nervous, so they pulled back. That is only half the story. On August 9, 2023, President Biden signed Executive Order 14105, directing the Treasury Department to restrict US investment into Chinese entities developing advanced semiconductors and microelectronics, quantum information technology, and specified AI systems relevant to military, intelligence, or surveillance applications. The rulemaking process that followed took longer than the initial signal suggested: an Advance Notice of Proposed Rulemaking in August 2023, a full Notice of Proposed Rulemaking in June 2024, and a final rule that only took legal effect on January 2, 2025 — meaning US funds operated under 18 months of proposed-but-not-final restriction, during which many chose to simply stop making new China commitments rather than navigate a moving compliance target.

The rule covers not just direct investment but, in many fund structures, LP commitments to vehicles that in turn invest in covered entities, which has made US institutional LPs (pension funds, university endowments, sovereign-adjacent US vehicles) broadly risk-averse to China-dedicated fund commitments even outside the three named technology areas. Separately, PCAOB audit-inspection access to Chinese ADR issuers was restored via an August 26, 2022 agreement (discussed in depth in a related fact), which technically resolved the Holding Foreign Companies Accountable Act delisting threat — but the resolution did not restore the pre-2021 confidence level among US public-market investors.

The commonly told version of this story is that Gulf sovereign wealth has stepped in cleanly to replace departing US capital. The real data is messier and worth correcting carefully, because getting it wrong leads directly to overconfident assumptions about who your capital-raising counterparts actually are in 2026. Gulf sovereign wealth funds as a group did post record global deployment in 2025 — the seven largest Gulf funds together committed roughly $126 billion globally, accounting for 43% of all capital invested by state-owned investors worldwide, a historical maximum, with Mubadala alone investing $12.9 billion in AI and digitalization globally in 2025 and Saudi Arabia's PIF committing $36.2 billion across its 2025 program. But their investment specifically into China moved in the opposite direction over the same period: sovereign wealth investment into China fell to an estimated $4.3 billion in 2025, down from $10.3 billion in 2024 — even as those same funds' overall global technology and AI spending hit records.

The resolution of this apparent contradiction is important: Gulf capital is genuinely more open to China than US institutional capital is (Mubadala and ADIA were part of an $8.3 billion consortium investing in China's largest mall manager in 2024, for instance), but 'more open than the US' and 'a rapidly growing, reliable pipeline into China specifically' are different claims. The correct 2026 picture is a foreign capital base that is smaller, more selective, and less US-dominated than 2018 — not one where a single new source of capital has cleanly stepped into the gap left by departing American LPs.

The practical consequence for a Gulf family office or a European growth fund entering this space now is that the institutional muscle memory that used to exist inside US-based China desks — built over a decade of Sequoia China, GGV, and Qiming-style operating experience — is not something most newer entrants have in-house, and the flow of new China deals with an existing foreign co-investor to learn from (via syndicate participation) has itself shrunk to 8.5% of rounds. That expertise gap is precisely the structural translation problem ZINVEO exists to close: the newer pools of foreign capital active in China today generally have less accumulated pattern-recognition on governance red flags, VIE risk, and political-exposure signals than the US funds that built that knowledge over the 2010s and have since scaled back, and they increasingly cannot simply learn by co-investing alongside a more experienced foreign syndicate partner, because there are fewer of those syndicate partners left in the room.

Breaking "foreign capital" down by investor type rather than geography sharpens the picture further. Traditional US pension and endowment LPs are the category that has retreated most completely, both for the outbound-investment legal reasons described above and for simple headline-risk aversion at institutions that answer to public boards or state legislatures. Family offices — Gulf, European, and increasingly Southeast Asian — have proven considerably stickier, in part because they answer to a single principal or family council rather than a broad institutional stakeholder base, and can make a more idiosyncratic risk-tolerance call than a public pension fund can. Corporate strategic investors (technology companies making venture-style bets adjacent to their core business) have also remained more active than pure financial LPs in specific sectors — a European automotive OEM investing in a Chinese battery-technology startup, for instance, is underwriting a strategic supply-chain relationship as much as a financial return, which changes its risk calculus relative to a pension fund's.

This matters for a new entrant because it suggests where to look for co-investors and reference points: a family office evaluating a Chinese deal today is more likely to find a useful comparison point in another family office's recent experience, or in a corporate strategic's diligence approach, than in the now largely-departed US pension fund playbook that dominated pre-2021 market commentary.

The $8.3 billion consortium investment by ADIA and Mubadala into China's largest mall manager in 2024, referenced above, is worth examining briefly as a template for the kind of deal Gulf capital has actually been willing to do in China even as its aggregate China allocation has fallen: a large, real-asset-backed, cash-flow-generating business with limited exposure to the specific technology sectors targeted by US outbound-investment restrictions, structured as a consortium (spreading both capital commitment and diligence burden across multiple sovereign investors) rather than a single fund acting alone. This pattern — large, real-asset-adjacent, consortium-structured, sector-selected specifically to sit outside the most geopolitically sensitive technology categories — is a more representative model of current Gulf appetite than a narrative of Gulf capital broadly replacing US venture capital across the technology sector would suggest.