By Remy Beaumont. Updated September 2026.

Total raised in confirmed APAC startup funding rounds this week: approximately $800m+ across 15+ disclosed rounds, headlined by Anew Labs' $290m inaugural financing following its spin-out from ByteDance, and dominated by a wave of Chinese physical-AI and deep-tech deals.

The headline raise: Anew Labs, $290m inaugural financing

What they do: Anew Labs is an AI drug-discovery company spun out of ByteDance. The parent company retains a 56% stake after the financing. Anew combines machine learning with molecular modelling to identify and develop drug candidates, initially focused on oncology and central nervous system disease.

How much: $290 million at a $1.5 billion valuation.

Round stage: Inaugural external financing (first outside capital post-spinout).

Lead investors: HSG (formerly Sequoia China), IDG Capital, and Hillhouse Investment co-led, with 5Y Capital as co-lead. Gaorong Ventures, Primavera Venture Partners, Boyu Capital, SBP Group, and the Shanghai Future Industries Fund also participated.

HQ: Shanghai, China.

Why this matters: This is the most significant corporate AI spin-out in APAC this year. ByteDance giving its drug-discovery unit an independent capitalisation table, external investors, and governance structures built around biotech economics rather than consumer internet timelines is a strong signal. Investors are assigning $1.5bn in value to the scientific team, technology, and IP before Anew has operated as a standalone business. The hard test comes next: AI-discovered molecules still need to survive toxicology, clinical trials, and regulators.

How they announced it

Anew did not announce it. That is the story. Reuters broke the round through two sources familiar with the transaction. ByteDance and the investors did not respond to requests for comment. The company let the investor roster, the ByteDance pedigree, and the $1.5bn valuation speak through a controlled leak rather than a managed launch.

In China's AI ecosystem, where government attention to frontier AI companies is intense, a quiet close followed by a journalist-sourced reveal is a deliberate strategy. It creates the narrative without the exposure of a formal press conference. The investor names do the credibility work: HSG (formerly Sequoia China), IDG, and Hillhouse are among the most recognised growth-stage investors in the region.

What founders can steal from this

  1. A controlled leak can be more powerful than a press conference. Anew let Reuters break the story with source-confirmed details. The resulting coverage felt like investigative reporting, not marketing. That framing carries more credibility with sophisticated readers, institutional investors, and potential partners than a self-published blog post.

  2. Spin-outs need a narrative about independence, not just parentage. The critical detail in Anew's story is the dedicated cap table, external investors, and separate governance. If you are spinning out of a larger company, the announcement must answer "why is this a standalone business?" before anyone asks.

  3. In biotech-meets-AI, clinical milestones beat model benchmarks. Anew's real announcement will come when its candidates enter human trials. Founders building at the intersection of AI and science should remember that investors eventually need clinical, industrial, or regulatory proof, not just computational performance.

Other notable APAC rounds this week

  • CADDi (Tokyo) raised ¥17.7bn (~$120m) in Series D funding at a ¥182bn valuation to build an AI data platform for global manufacturing. New investors include Woven Capital (Toyota-backed), Salesforce Ventures, and Atomico.

  • Hello Robotaxi (Shanghai) raised ~$100m at nearly $3bn post-money for Level 4 autonomous driving, led by Shanghai Guotou's Pioneer Fund. The company plans a 10,000-GPU computing platform.

  • Space Epoch (Beijing) completed a Series B++ financing, bringing total B-series capital over six months to over RMB2.3bn (~$320m) for reusable rockets targeting first orbital launch by end of 2026.

  • Apex Intelligence (China) raised nearly RMB400m (~$55m) across Angel and Angel+ rounds less than three months after founding, for recursively self-improving AI research. Founded by a 27-year-old Tsinghua assistant professor.

  • ENCOS (Nanjing) raised >RMB300m (~$42m) Series B for robotic joint modules and humanoid-robot components.

  • Yincheng Intelligence (Shanghai) raised RMB100m ($14m) Series A for logistics-picking robots, alongside a reported RMB100m-class order from SF Express.

  • Kangwei Vision (Shenzhen) raised RMB100m ($14m) Pre-Series A for AI-driven semiconductor and PCB optical inspection equipment.

  • Nutshell Therapeutics (Shanghai) raised tens of millions in a C1 round for AI-driven drug discovery targeting difficult-to-drug protein conformations. Its p53 candidate is in Phase I.

  • Qubot (China) raised ~$13.8m Series A extension for magnetic surgical robotics.

The week's pattern

China dominated APAC funding this week with six of the ten largest rounds, spanning AI drug discovery, autonomous driving, commercial space, robotics components, logistics automation, and semiconductor inspection. Japan contributed the second-largest round through CADDi's manufacturing-AI play.

The theme connecting these deals is physical AI. Not chatbots. Not model APIs. Robots that pick packages, joints that move humanoid limbs, inspection systems that verify semiconductor boards, rockets that need to physically fly, and autonomous vehicles that must drive real roads. Chinese capital is simultaneously financing the AI research layer (Apex Intelligence) and the physical systems that AI needs to prove economic value in the real world.

For APAC founders, the lesson is clear: the most compelling funding stories combine AI capability with something models alone cannot replicate. CADDi owns decades of manufacturing data. Yincheng has a signed RMB100m order from SF Express. ENCOS makes the physical joints robots need to move. The moat is not the model. It is the proprietary data, the hardware, the customer contract, or the regulatory approval that sits around the model.