SenseTime and SenseNova: How a National-Scale Chinese AI Platform Works

SenseTime (商汤科技) is one of the clearest examples in China of a full-stack AI company — meaning it owns significant pieces of the infrastructure, the foundation models, and the applied products built on top. For executives trying to understand how a national-scale Chinese AI platform is actually assembled, SenseTime is an ideal case.

The three layers in one picture

• SenseCore infrastructure. SenseTime''s AI compute and data infrastructure — large training clusters, data platforms, and the engineering that sits underneath the model stack.
• SenseNova (日日新) foundation models. A series of multimodal foundation models — text, vision, speech, and cross-modal capabilities — updated on a regular cadence and benchmarked publicly. SenseNova is used both internally and sold to enterprise customers.
• Applied product verticals. Smart auto (ADAS and cabin intelligence), smart city, healthcare, AR/content, and generative AI. Each vertical is a sales motion with dedicated engineering behind it.

This vertical integration matters because it lets SenseTime do things most Chinese and Western peers cannot: iterate faster on data ↔ model ↔ product, and offer enterprise customers a stack rather than a single API.

What a SenseTime visit typically includes

A productive on-site briefing covers:

1. A walkthrough of the SenseCore infrastructure and training roadmap.
2. A SenseNova model family overview — current capabilities, benchmarks, and enterprise-facing APIs.
3. Live demos of the headline verticals: driver monitoring / ADAS, medical imaging, AIGC/content generation, and smart-city visual analytics.
4. A conversation on international deployment — how the stack is packaged for customers outside China, including Africa and the Middle East.

Why this matters for African executives

Three specific angles are usually worth pushing on:

• Computer vision in mining, security and public sector. SenseTime''s vision capabilities are among the most deployed in China and translate well to African operational contexts — mine-site safety, port monitoring, retail analytics, smart-city CCTV.
• Cabin intelligence and ADAS. For fleet operators, logistics businesses, and transport regulators exploring AI-equipped vehicles, SenseTime is one of the most relevant vendors to know directly.
• Model localisation. Any serious deployment will need language, regulatory and data localisation. Asking how SenseNova localises — and how much engineering it takes — is a high-value question that few online briefings will answer truthfully.

How SenseTime fits the 2026 tour

SenseTime is a Day 2 Shanghai stop on the open cohort. In the broader arc of the week, it is the "platform" bookend: Lenovo provides the hybrid-AI frame, SenseTime provides the full-stack AI operator story, and the rest of the week tests both against industry realities.

---

The detailed Day 2 briefing and prep reading are included in the cohort brochure.