SenseTime Says Cost-Efficient AI Can Compete With Global Frontier Models

CNBC reported Tuesday that SenseTime is leaning into cost-efficient AI as its primary competitive weapon. The Hong Kong-listed firm believes lower-cost models can win meaningful market share even against more powerful rivals.

SenseTime’s Cost-Efficiency Argument

Co-founder and chief scientist Lin Dahua told CNBC that SenseTime’s latest model, SenseNova U1, runs at roughly one-tenth the cost of OpenAI’s image-generation tool. Lin acknowledged a quality gap still exists versus frontier Western models. However, he argued that most real-world tasks do not require top-tier performance. The multimodal model combines language and vision processing in a single unified system. That integration removes translation steps between data types, improving speed and cutting per-use expenses.

SenseTime has drawn inspiration from DeepSeek’s approach of squeezing strong results from tighter resource budgets. The company is also integrating capabilities from ByteDance’s AI video model Seedance into its own short-video product. That move allows SenseTime to combine external strengths with its own audio tools rather than building everything in-house.

Also Read: What DeepSeek’s Efficiency Push Means for AI Infrastructure Spending

Background: A Company Under Pressure

Founded in Hong Kong in 2014, SenseTime built its reputation on facial and image recognition technology. The company has since pivoted hard toward generative and multimodal AI to remain relevant. It has faced U.S. sanctions over allegations tied to surveillance of Uyghur minorities in Xinjiang, charges the firm denies. Despite those restrictions, SenseTime is actively pursuing overseas expansion, including in the Middle East, with no reported change to those plans.

Financially, the company narrowed its net loss by more than 58% last year. It also posted positive EBITDA in the second half of 2025 for the first time since its 2021 listing.

Also Read: China’s Big Tech Platforms Are Using Core Businesses to Fund AI Development

Pure-Play AI Firms Face a Structural Challenge

Analysts at Jefferies warned in a late-April note that standalone AI model companies face a difficult structural position. Low customer loyalty, a crowded field, and steep training costs create a tough revenue equation. Platform giants such as Alibaba, Tencent, and ByteDance hold a natural advantage. Their established user bases and cash flows allow them to subsidize AI development and cross-sell AI tools into existing products.

SenseTime is attempting to bridge that gap by combining models, applications, and infrastructure into a single offering. Targeting enterprise clients, who tend to pay more and switch less, gives the firm a more stable revenue base than consumer-facing peers.

Read Next: China’s AI Investment Surge Puts Pressure on U.S. Hardware Export Controls