Turning AI Innovation into Sustainable Profitability: Deepexi Technology (1384.HK) Delivers Surging Revenue, Quarterly Profitability, and a Differentiated Enterprise AI Platform

Hong Kong-based Deepexi Technology (1384.HK) has emerged as a rare example of an enterprise AI platform delivering both technological leadership and profitability. Their first-half 2026 results show a 115% year-over-year revenue surge, driven by a 209% increase in core AI revenue. This performance earned them "Buy" and "Overweight" ratings from brokerages, just one year since the public listing.

Deepexi's AI business generated RMB 226.0 million, a 209.2% year-over-year leap, accounting for 79.6% of total revenue. The company posted its first-ever quarterly net profit of approximately RMB 30.1 million in the second quarter of 2026, breaking even. This marks a significant milestone in Hong Kong's AI sector, where few pure-play enterprise AI companies achieve profitability, let alone quarterly net profit.

Deepexi's success is rooted in its AI-native platform, the DeepWorks Enterprise Agent platform. Unlike traditional SaaS vendors tied to human seat counts, Deepexi's consumption-based model aligns directly with the agentic AI future, offering enterprises AI productivity through autonomous agents. This model is less susceptible to the "SaaSpocalypse," where $2 trillion in B2B software valuations were erased in the first quarter of 2026 due to concerns over agentic AI displacing traditional licensing models.

Deepexi's success stems from its specialized "Data + AI" architecture, anchored by the FastData enterprise data intelligence platform and FastAGI agentic AI suite. The company's proprietary Deepology ontology dataset, containing over 2,000 vertical industry skills, powers these platforms, providing enterprise clients with domain-specific AI models that can be deployed without vendor lock-in.

This focus on enterprise-specific applications and proprietary data sets Deepexi apart from generic foundational models, ensuring its AI solutions truly meet real-world enterprise demands.