DataArc
数创弧光
DataArc (数创弧光), founded in January 2025 in Shenzhen and incubated by the Institute for Intelligent Computing (IDEA Research Institute), with academician Harry Shum serving as project advisor, focuses on synthetic data technology for large language models.
Position in SVTR
Snapshot 2026-10-22- Startup SignalsRankedof 2,297↓44
- DataRankedof 85–
About
DataArc (数创弧光), founded in January 2025 in Shenzhen and incubated by the Institute for Intelligent Computing (IDEA Research Institute), with academician Harry Shum serving as project advisor, focuses on synthetic data technology for large language models. The company uses synthetic data to systematically fill gaps where real data is unavailable, unshareable, or insufficient, covering stages including continual pre-training, supervised fine-tuning, and reinforcement learning.
Full tier & rationale — an SVTR Intelligence benefit
Capital graph
Company· 11
Investor· 5
Funding history
- 2026-07-01New funding round数千万元
Investors · Inno Angel Fund, 深智城, 深圳天使母基金, 君科丹木, 一村淞灵Source · SVTR Weekly #164 ↗
Comparable companies· by market
This company
BusinessLatest roundTotal raisedDate- DataArc
Synthetic data technology for large models, filling data gaps at the pre-training, fine-tuning and reinforcement learning stages for low-resource languages, privacy-sensitive and long-tail scenarios
2026-07
—
United States
BusinessLatest roundTotal raisedDate- Sepal AI
Training data, evaluation benchmarks and reinforcement learning environments for frontier models
— - Micro1
Human expert annotation, reinforcement learning evaluation and robot training data for frontier AI labs
Undisclosed · $71M · 2025-09 · Valuation $500M
$71MValuation $500M - Proximal
Long-horizon RL training environments and evaluation data for frontier AI labs, starting with coding agent training data
Seed · $15M · 2026-09 · Valuation $300M
$15MValuation $300M
China
BusinessLatest roundTotal raisedDate- 大衍科技
Spatial intelligence and synthetic data provider for embodied AI, with in-house tactile data collection devices and a tactile foundation model, supplying synthetic training data for robotics and autonomous driving.
Angel · 2026-07
— - MetAI
AI-based digital twins and synthetic data generation for manufacturing
Seed · $4M · 2025-01
$4M - Boden AI
Data annotation for 2D images, 3D/4D point clouds and medical imaging, serving autonomous driving and large-model companies
Series A · 2026-06
—
Europe
BusinessLatest roundTotal raisedDate- Aindo
Creates synthetic data for testing and training machine learning models
Series A · $6M · 2023-10
$6M - BeatpulseLabs
Converts multimedia content and domain expert knowledge into structured, labeled and validated training datasets for enterprise AI teams, addressing high-quality data supply for multimodal model training, fine-tuning and evaluation.
Pre-Seed · $2M · 2026-06
$2M - ICEYE
Builds a constellation of small synthetic aperture radar (SAR) satellites, providing government and commercial customers with day-and-night, all-weather Earth observation data
Series D+ · $1.0B · 2026-06 · Valuation $12B
$1.0BValuation $12B
Other / unspecified
BusinessLatest roundTotal raisedDate- Config
Training data infrastructure for robotics foundation models
Seed · $35M · 2026-05 · Valuation $200M
$35MValuation $200M - WEKA
Its AI data platform is designed to enable data portability across on-premises, cloud and edge infrastructure
Series D+ · $140M · 2024-05
$140M - Upstage
Develops enterprise AI tools and large language models that convert unstructured documents into structured data
Series B · $117M · 2025-08
$117M
Confirmed rounds only. Matched on subsector, sector, stage and funding scale — each verifiable on the company profiles.
Investors that may fit
Nanshan Strategic Emerging Industry Fund
2 Data deals in the last 24 months · 2 in South China
Request intro →Ranked by each firm’s deals in this sector over the last 24 months, weighted by nearby stage and shared geography. Indicative only — no rating judgment.