Technology & AIAnalysis

Chinese AI Companies Shift to Profitability as Recurring Revenue Models Expand

Interim corporate earnings show enterprise software and AI vendors turning profitable through API and subscription billing.

Share
High-tech humanoid robot with LED face display, showcasing modern robotics and innovation.
Photo by Kindel Media on Pexels

The Brief

A cohort of Chinese software and artificial intelligence firms has reported sharp profit gains or swung into profitability in interim earnings, signaling an operational shift away from one-time project contracts toward recurring revenue streams. Reporting highlighted by People's Daily showed companies including SenseTime, 360 Security Technology, and 4Paradigm recording improved margins backed by subscription-based and API usage billing models. While heavy research spending persists across the broader sector, policy backing under Beijing's 'AI+' initiative is accelerating commercial deployment across enterprise workflows.

Why it matters

For years, commercial AI development faced skepticism over capital intensity, high cash burn, and slow monetization. The emergence of recurring revenue streams such as API calls and subscription add-ons demonstrates that enterprise demand is maturing from speculative experimentation into measurable operating expenditure, stabilizing vendor cash flows.

China context

China's state planners have prioritized the 'AI+' initiative since its inclusion in the 2024 government work report, followed by State Council implementation guidelines and targets embedded in the 15th Five-Year Plan. This policy framework encourages state-owned enterprises and traditional industries to integrate algorithmic tools directly into production lines, classrooms, clinics, and office software.

Editor's View

EDITOR'S VIEW — Analysis and inference, not factual reporting. The transition highlighted in these mid-year earnings illustrates a pragmatic shift in China's AI ecosystem: competing on parameter counts and venture funding headlines is yielding to monetization discipline. However, investors and industry observers should distinguish between core model creators and established enterprise software vendors layering AI features onto existing customer bases. While hybrid billing models like WPS 365 validate incremental monetization, broader sector profitability will depend on whether API consumption can remain durable without heavy platform discounting.

What to watch

  • Sustainability of annualized recurring revenue and API consumption growth across subsequent quarterly filings.
  • Operating cash flow trends among early-stage AI startups compared to established software vendors.
  • Enterprise retention rates for hybrid subscription-plus-usage pricing models in traditional business sectors.

Key Takeaways

  • 1Several Chinese AI and enterprise software companies posted sharp earnings increases or turned profitable in interim financial reports.
  • 2SenseTime posted its first IFRS net profit of 617 million yuan, with recurring revenue rising 124.4 percent to 1.145 billion yuan.
  • 3360 Security Technology and 4Paradigm both swung from losses to profitability, with 4Paradigm reporting an 860.8 percent increase in API revenue.
  • 4Vendors are transitioning from custom project contracts to recurring revenue, including subscription tiers and per-call API billing.
  • 5State support under the 'AI+' initiative and the 15th Five-Year Plan is facilitating broader adoption across enterprise and public-sector environments.
A wave of interim financial disclosures from Chinese software and artificial intelligence developers indicates that parts of the sector are beginning to achieve profitability, moving away from capital-intensive project contracting toward utility-style recurring revenue models, according to an analysis published by People's Daily. Financial results cited in the report reflect notable bottom-line improvements across several publicly traded technology firms. Kunlun Tech reported a net profit attributable to parent company shareholders of 1.088 billion yuan, representing a 227.17 percent increase year on year. Cybersecurity and cloud infrastructure firm Sangfor recorded net profit of 231 million yuan, up 201.58 percent, while 360 Security Technology returned to profitability with 216 million yuan in net income, reversing a 282 million yuan loss in the prior-year period. The turnaround was particularly evident among pure-play AI and productivity software vendors adjusting their billing architecture. SenseTime logged its first net profit under International Financial Reporting Standards since listing, posting 617 million yuan. Its recurring revenue reached 1.145 billion yuan—up 124.4 percent year on year—accounting for nearly 40 percent of total turnover. Enterprise productivity vendor Kingsoft Office saw WPS 365 subscription sales expand 60.84 percent to 497 million yuan as it introduced hybrid pricing combining base subscriptions with AI usage fees. Meanwhile, enterprise AI firm 4Paradigm swung into the black, driven by a 860.8 percent surge in application programming interface revenue to 464 million yuan. The People's Daily report attributed this operational transition to broader structural shifts. Previously, enterprise software engagements often relied on customized, one-off deployment contracts that carried high delivery overhead. In contrast, current growth increasingly stems from automated model calling and standardized API access, enabling software providers to charge clients on an ongoing consumption basis similar to public utilities. This commercial progression aligns with state policy directives. Following the formal introduction of the 'AI+' action campaign in the 2024 government work report, the State Council issued detailed implementation guidance aimed at embedding machine-learning tools into production workflows across diverse industries. The upcoming 15th Five-Year Plan further incorporates systematic rollouts for technology adoption. Nonetheless, the report emphasized that profitability remains uneven across the industry. Many firms still operate at a loss while directing substantial capital toward foundational technology research, algorithmic improvements, and business model restructuring.