In a market where enterprise software valuations are being repriced by agentic AI, Deepexi Technology (1384.HK) has emerged as a counterexample, delivering both technology leadership and profitability. The company's first interim results post-listing reveal a 115% year-over-year revenue surge, powered by a 209% leap in core AI revenue, and a decisive turnaround to quarterly GAAP profitability in Q2 2026.
For H1 2026, Deepexi generated RMB 284.0 million in total revenue, up 115.0% year-over-year. AI revenue contributed RMB 226.0 million, a 209.2% surge, accounting for 79.6% of total turnover. Gross profit reached RMB 160.0 million, up 120.5%, with margins expanding to 56.5%. Notably, Q2 2026 saw a standalone net profit of approximately RMB 30.1 million, crossing the breakeven threshold.
This profitability is rare among Hong Kong's pure-play enterprise AI firms, underscoring Deepexi's operational efficiency and scalable software delivery. The company's vertical focus is strong: industrial manufacturing accounts for over 50% of revenue, with retail and consumer goods at 30%, plus growing presence in healthcare and smart transportation. In manufacturing, Deepexi translates complex engineering schematics, IoT sensor data, and maintenance logs into structured logic for AI models, creating a formidable competitive barrier.
The first quarter of 2026 witnessed the "SaaSpocalypse," where $2 trillion was erased from B2B software equities due to fears that agentic AI would cannibalize per-seat licensing. The IGV ETF plunged 21%, and EV/Sales multiples dropped from 5.6x to 4.2x. Investors now favor AI-native infrastructure, with AI-native agentic platforms clearing 14x to 22x revenue, while legacy RPA trades at 2.5x to 5x, according to a Windsor Drake analysis. Deepexi sits on the winning side, with a token-based, consumption-driven model that sells AI productivity rather than seats, structurally insulated from seat compression.
Deepexi's competitive edge lies in its "Data + AI" dual-engine architecture, anchored by the FastData platform, FastAGI suite, and Deepexi Enterprise Large Model Platform, supported by the Deepology ontology with over 2,000 vertical skills. The FastData platform tokenizes complex data for AI training, while FastAGI and DeepWorks provide a modular AI operating system. The platform is model-agnostic, connecting to open-source models like DeepSeek and Zhipu AI, eliminating vendor lock-in. DeepWorks integrates with DingTalk, Feishu, and Tencent Meeting, enabling employees to deploy AI copilots for automated reporting and coordination. With 42% of organizations having AI agents in production, Deepexi's governed workflows are core infrastructure.
On August 12, Deepexi announced a strategic alliance with Huawei Cloud to jointly launch enterprise data intelligence solutions. Huawei provides Ascend AI computing clusters, while Deepexi contributes data governance and vertical AI deployment expertise. This integration secures stable compute power and reduces token processing costs, enabling token-metered billing models. Falling AI costs are expected to accelerate agent adoption, positioning Deepexi for margin expansion as token volumes scale.
Institutional confidence is strong: on August 17, Deepexi successfully placed 14,286,000 new H-shares at HK$35.00, raising net proceeds of approximately HK$488.0 million. This placement, completed amid market volatility, reflects institutional recognition of Deepexi's model and profitability trajectory. Strategic investors include Hillhouse, 5Y Capital, BAI, and Shanghai AI. The funds will support R&D in next-generation enterprise agent systems, targeting the $206.5 billion AI agent software market projected for 2026, growing to $376.3 billion in 2027, per Gartner.
Investment thesis highlights: AI-native architecture aligned with agentic future, rare profitability at 56.5% gross margin with 95.5% revenue CAGR projected through 2027, model-agnostic platform, proprietary data moat with over 2,000 vertical skills, strategic Huawei Cloud partnership for cost-effective compute, and institutional validation. As the enterprise AI market shifts to commercial execution, Deepexi proves that AI can deliver both transformative technology and sustainable profits.
