Deepexi Technology Co., Ltd. (1384.HK), a pure-play enterprise AI platform, has announced its first interim results since listing, revealing a 115% year-over-year revenue increase to RMB 284.0 million for the first half of 2026. Core AI revenue surged 209.2% to RMB 226.0 million, representing 79.6% of total revenue. The company also achieved a significant milestone: GAAP net profitability in the second quarter of 2026, with a net profit of approximately RMB 30.1 million. This marks a rare achievement for an enterprise AI firm in Hong Kong, particularly while maintaining high R&D investment.
The company's gross profit grew 120.5% to RMB 160.0 million, with gross margin expanding to 56.5%. Revenue concentration in industrial manufacturing exceeded 50%, while retail and consumer goods contributed 30%, with growing traction in healthcare and smart transportation. This vertical depth, combined with proprietary data accumulation, creates a competitive barrier that general-purpose AI vendors struggle to replicate.
The announcement comes amid what analysts call the "SaaSpocalypse," a market correction that erased approximately $2 trillion in B2B software market capitalization in early 2026. The iShares Expanded Tech-Software ETF (IGV) fell nearly 21% year-to-date, and enterprise software EV/Sales multiples dropped from 5.6x to 4.2x by mid-March. Investors have shifted toward AI-native infrastructure, rewarding platforms that manage data and autonomous systems. Deepexi's token-based, consumption-driven model aligns with this shift, avoiding the per-seat licensing vulnerabilities of traditional SaaS. As a Windsor Drake analysis notes, AI-native agentic platforms command 14x to 22x revenue, while legacy RPA trades at just 2.5x to 5x.
Deepexi's DeepWorks enterprise Agent platform sells AI productivity rather than software seats, enabling autonomous agents to replace repetitive administrative work. This structural insulation from seat compression is a key differentiator. The company's "Data + AI" dual-engine architecture, including the FastData platform and FastAGI suite, is anchored by the Deepology ontology dataset containing over 2,000 vertical industry skills. The FastData Foil fusion engine tokenizes complex data for large model training, while the platform's pluggable architecture supports open-source models like DeepSeek and Zhipu AI, ensuring vendor neutrality. DeepWorks integrates with workplace tools such as DingTalk, Feishu, and Tencent Meeting, embedding AI into daily operations.
In an enterprise AI market where 42% of organizations already have AI agents in production, Deepexi's governed workflows are becoming core infrastructure. The company's strategic partnership with Huawei Cloud, announced on August 12, addresses compute and data bottlenecks. By integrating with Huawei's Ascend AI clusters, Deepexi secures stable computing power and reduces token processing costs, enabling usage-based, token-metered pricing models that align with the agentic AI future.
Institutional confidence is evident through a successful placement of 14,286,000 new H-shares at HK$35.00 each, raising net proceeds of approximately HK$488.0 million. The company's strategic shareholders include Hillhouse, 5Y Capital, BAI, and Shanghai AI. The funds will support R&D in next-generation enterprise agent systems and token productivity platforms, targeting the projected $206.5 billion AI agent software market in 2026, expected to grow to $376.3 billion by 2027.
Deepexi's investment thesis highlights its AI-native architecture, rare profitability, model-agnostic approach, proprietary data moat, strategic compute partnership, and institutional backing. As the enterprise AI market shifts from conceptual exploration to commercial execution, Deepexi demonstrates that enterprise AI can deliver both transformative technology and sustainable profits.
