Betterdata

Market position

Positioned as an enterprise synthetic data infrastructure provider focused on regulated industries and government. Third-party material cites recognition by Gartner and an 'Outperformer' placement on a CB Insights ESP matrix for test data management relative to IBM and Oracle, and names Tonic AI and YData as competitors. The company also cites a Gartner estimate that synthetic data will overshadow real data in AI/ML training by 2030.

Stated differentiators include a breadth of proprietary model families covering tabular, relational and time-series data with a promptable foundation model for non-technical users; formal differential privacy guarantees and audit reporting; flexible deployment including air-gapped and on-premise installations via CIS-hardened packages; participation in authoring Singapore's synthetic data guidelines with the PDPC; and peer-reviewed research at NeurIPS and ICML.

Technology

A multi-model generation engine spanning a Tabular Foundation Model (TFM) pre-trained on more than one billion records, deep learning GANs, tree-based methods (SPN, ARF), and transformer/LLM-based models. Named proprietary models include TabTreeFormer, ARF, CTAB-GAN-DP and Tabula for tabular data; IRG-GAN and SPN for relational data with referential integrity across full database schemas; and Fractal-TSG, TS-V0 and TS-V1 for regular and irregular time-series data. Text-in-table is also supported. Privacy controls include integrated differential privacy guarantees, advanced anonymization, PII redaction, regulatory-grade audits, dataset versioning and locking, RBAC, and CIS-hardened packaging. The company states the stack was developed over five years and exceeds one million lines of code.

Go-to-market

Direct enterprise sales via a contact-sales and demo funnel, supported by published case studies, use-case pages, blogs and video tutorials. Seed proceeds were designated in part for hiring sales and marketing staff for business development. Credibility signals used in go-to-market include government deployments, participation in Singapore's PDPC synthetic data guidance, analyst mentions (Gartner, CB Insights) and academic publications.

Enterprises and government agencies operating under strict data regulation, particularly financial services (Tier 1 banks, trading, capital modeling, AML), healthcare, telecommunications, government/federal agencies, and Fortune 500 companies. Users span data science and engineering teams as well as non-technical users served through promptable models.

Geography

Headquartered in Singapore (PIXEL at One North, 10 Central Exchange Green). Customers and deployments span Asia (Tier 1 Asian bank, Kajima Corporation, Shanghai Pudong Development Bank), the United States (Department of Homeland Security) and Europe (luxury goods client). The company stated in 2023 an intention to expand across the Asia-Pacific region over the following one to two years.

History

Established in 2021 and headquartered at PIXEL at One North in Singapore. A $770K equity round was reported in October 2022, followed by an oversubscribed seed round announced in April 2023 (reported as US$1.7 million by The Business Times and as $2M by a startup database), used to publicly launch the product, extend support to single-table, multi-table and time-series data, and hire commercial staff. The company subsequently expanded into relational and time-series model families and a Tabular Foundation Model, contributed to Singapore's PDPC synthetic data guidance, and secured government and Fortune 500 deployments.

Risks & controversies

Sources conflict on key facts: the April 2023 seed round is reported as US$1.7 million by The Business Times and as $2M led by Khairu Rejal by a startup database (Khairu Rejal is identified in the press report as a principal at lead investor Investible). The same database lists an employee range of 1,001-5,000, which is inconsistent with the small team roster published on the company's own about page and with its disclosed funding. Most performance and customer claims (60% savings, 40% QA reduction, 5% conversion uplift, 70% faster data sharing) originate from the company's own marketing materials and are not independently verified in the sources provided. Several named customers are described only generically (Tier 1 Asian bank, Fortune 500 data storage provider, European luxury maison).