Duality, which uses homomorphic encryption to provide privacy-preserving data analysis tools, raises $16M led by Intel Capital
The ubiquity of APIs and cloud solutions have opened up a world of interesting ways for businesses to create a service without having to build every part of it themselves.
Context & Ripple Effects
This 2019 round was Intel Capital's early position in a company that has since compounded: Duality went on to raise a $30M Series B led by LG Technology Ventures in 2021, and by 2025 was applying its core technology to a private LLM inference framework that lets models answer fully encrypted prompts. The $16M is where that arc starts — a bet that homomorphic encryption could move from academic cryptography to commercial data-collaboration tooling.
The round also seeded a now-crowded category: Enveil raised a $25M Series B led by USAA on the same homomorphic approach, while Opaque took a different route through trusted execution environments, raising successive rounds that culminated in a $24M Series B at a $300M valuation in 2026. Intel Capital's check in 2019 preceded all of that validation.
First-order effects
- Duality gains $16M and Intel Capital as a strategic backer to commercialize its privacy-preserving data analysis tools for enterprises working on sensitive data.
- Intel Capital secures early equity in a privacy-computation startup years before the category drew larger rounds from LG Technology Ventures, USAA, and Walden Catalyst.
Second-order effects
- The round helps legitimize homomorphic encryption as a fundable product category, drawing parallel bets on the same approach (Enveil) and on the rival trusted-execution-environment approach (Opaque) — enterprises now face a choice between two competing technical routes to the same promise.
- Intel Capital's involvement ties Duality's tooling to a chip vendor's ecosystem at a time when Intel's venture arm is one of the most active foreign investors in China's AI and chip startups, giving the portfolio both strategic reach and geopolitical exposure.
Third-order effects
- If the pattern holds, regulated-industry data collaboration shifts from 'move the data to the compute' to 'compute on encrypted data' — Duality's 2025 encrypted-prompt LLM work shows the same 2019-era technology extending into AI inference, where privacy becomes a prerequisite for enterprise adoption rather than an afterthought.
- Privacy-preserving computation is consolidating into an infrastructure layer with rival camps (FHE versus TEEs), and the venture firms that backed both sides early — Intel Capital chief among them — are positioned to shape which approach becomes the default for sensitive AI workloads.
The trend: Privacy-preserving computation is maturing from a cryptographic research niche into enterprise AI infrastructure, with homomorphic encryption and trusted execution environments emerging as competing standards and early strategic investors like Intel Capital anchoring both.