Funding for Artificial Intelligence (AI)-related firms across Asia-Pacific is slowing, as private-equity (PE) and venture capital (VC) players are buffeted by headwinds in their attempts to raise new capital and exit previous investments despite dealmakers’ growing interest in unlocking and monetizing AI potential.

The APAC region reported 345 deals worth USD 5.06bn in the year to date (YTD) to 13 November, according to AVCJ data. Deal volume and deal count appear likely to fall well short of the levels recorded in 2022, when 481 transactions were valued at USD 7.97bn, and remain a far cry from the all-time high of USD 12.47bn generated across 518 deals in 2021.

The slowdown of activity in China – the world’s second-largest AI powerhouse behind the US – has been particularly pronounced. Volume has tumbled from USD 8.55bn in 2021 to USD 3.15bn in 2022 and USD 1.9bn in 2023 YTD.

The Middle Kingdom accounted for 68.6% of AI-related deal volume in Asia-Pacific in 2021, but only 39.5% in 2022 and 37.5% this year. The country’s market share in terms of deal count has also fallen, albeit less dramatically, from 43% in 2021 to 36.2% in 2022 and 31.9% in 2023.

China feeling geopolitical pinch

The PE and VC industry, already reeling from soaring interest rates and slowing economic growth across APAC, faces even greater challenges in China in the form of geopolitical and regulatory hurdles.

USD funds are beating a retreat following new regulations from the Biden administration aimed at blocking investment in China’s cutting-edge technologies, including AI, semiconductors and quantum computing.

Meanwhile, the US House of Representatives Select Committee on China is now probing legitimate investments made in Chinese high-tech sectors by VC firms GGV Capital, GSR Ventures, Walden International, Qualcomm Ventures and Sequoia Capital, which were finalized even before the new rules came into effect.

At least three of these VC firms – Sequoia Capital, GGV Capital and BlueRun Venture – split their China or Asia-related operations into a separate, dedicated legal entity to comply with new US regulations.

However, China-focused venture funds are finding it increasingly difficult to raise funds from Western investors, and USD fundraising is getting tighter and tighter. USD funds typically have a long investment maturity and higher risk tolerance than domestic yuan funds, which are less likely to invest in early-stage projects with no income.

Beijing battles back

Chinese authorities are cognisant of the problems they face and have adopted a series of measures to address the challenges. These include central government seeking to make state-capital guidance funds less risk averse, as well as encouraging the development of PE, VC and M&A-related financial products with greater risk tolerance.

Local governments including Guangdong province – China’s leading province in terms of GDP, and a key tech and manufacturing powerhouse – as well as Beijing, Shanghai, Shenzhen and Hangzhou have all rolled out AI technology-advancement plans to bolster the industry’s computing power, data and algorithmic capabilities.

Even amid these unprecedented challenges, China continues to attract APAC’s largest investment into companies whose core business is AI:

• AI applications software developer Baichuan Intelligence netted USD 300m in Series A1 funding in October, thanks to the participation of Alibaba, Tencent and Xiaomi, among others.
• Tencent-backed Enflame, a Chinese deep-learning chip company, raised CNY 2bn (USD 274.8bn) in Series D financing in September.
• AI solutions provider Minimax secured USD 250m from investors in its latest funding round in June, including an entity linked to Tencent.

LLMs seek targeted investment, ripe for consolidation

Certain AI areas may require more targeted investment and consolidation in China.

Yipin Ng, co-founder and managing partner of Yunqi Partners, told the AVCJ Private Equity & Venture Forum in Hong Kong that China needs to have its own set of AI champions. Yunqi Partners was an early investor in Minimax’s aforementioned fundraising.

However, Baidu CEO Robin Li recently warned that a rush to develop AI large language models (LLMs) – the underlying technology powering ChatGPT – risks wasting resources. As of October, 238 LLMs have been released, up from only 79 in June.

Several companies are fighting hard to grab a slice of the investment pie. The crowded contestant field includes the likes of Zhipu AI, Emotibot, 01.AI, Cloudwalk Technology, Mobvoi and Langboat Technology.