Text/Yang Jianyong

SenseTime, once hailed as one of the "AI Four Little Dragons," is now on the brink of profitability, with an anticipated profit ranging from 500 million to 700 million yuan in the first half of 2026. This marks a stark contrast to the 1.489 billion yuan loss incurred during the same period last year, signaling a successful transformation from losses to gains.

Buoyed by these profit expectations, SenseTime's market value experienced a significant surge today, peaking at a 12% intraday increase and reaching a market capitalization of 65.5 billion Hong Kong dollars. However, this still represents a substantial decline from its initial market capitalization of 320 billion Hong Kong dollars at the time of listing, amounting to a reduction of 254.5 billion and an 80% shrinkage.

Since the advent of the generative AI era, the artificial intelligence industry has been presented with fresh avenues for growth. Leveraging the strengths of its AI-native infrastructure, SenseTime Core, SenseTime has capitalized on this market opportunity, rebounding to a high-growth trajectory and escaping the quagmire of persistent substantial losses.

In 2025, SenseTime's revenue soared to 5 billion yuan, marking a 33% year-on-year increase. With revenue hitting an all-time high, losses further contracted by 58.6%, culminating in an annual loss of 1.782 billion yuan. Notably, the EBITDA for the second half of 2025 stood at 380 million yuan, marking the first positive figure since the company's listing.

It's worth noting that, as the former frontrunner among the "AI Four Little Dragons," SenseTime previously grappled with exceptional challenges, struggling under the dual pressures of declining revenue and significant losses. From 2018 to 2024, cumulative losses amounted to 54.4 billion yuan. Although a substantial portion of these losses stemmed from fluctuations in the fair value of preferred shares, even after adjusting for this factor, cumulative losses remained in the tens of billions, underscoring the immense operational pressure.

The turning point for SenseTime arrived with the emergence of ChatGPT, which sparked a competitive race among various large models. Models such as ChatGPT, DeepSeek, GLM, and Kimi were launched in succession, with industries across the spectrum actively harnessing the transformative capabilities of AI large models to drive intelligent upgrades. Simultaneously, the deployment of AI large models has become widespread, extending from the cloud to the edge.

Confronted with the historic opportunities presented by the new wave of AI large models, SenseTime restructured its business in 2023, primarily focusing on generative AI. By the end of 2024, after redefining its strategic direction and core business, it established a new "1+X" architecture.

This strategic pivot enabled SenseTime to forge three new business models: generative AI, visual AI, and X innovative businesses. Among these, generative AI has exhibited sustained growth, emerging as the largest business segment. In 2025, revenue from generative AI reached 3.629 billion yuan, representing a 63.7% year-on-year increase, with its revenue share soaring to 72.4%.

The remarkable growth in generative AI can be attributed to the sustained explosive demand in the market for training, fine-tuning, and inference of generative AI models, a demand that has been met by SenseTime's construction of a robust AI infrastructure.

As of March 2026, the total operational computing power of SenseTime Core has reached 40,400 petaFLOPS, laying a solid foundation for the trinity strategy of large-scale infrastructure, models, and applications, and paving the way for the development of the generative AI business.

With the realization of scale effects in the generative AI business and changes in the fair value of AI ecosystem enterprises built by SenseTime, profitability is expected to improve further. The anticipated positive consolidated profit under IFRS in the first half of 2026 dispels market concerns about long-term losses and strengthens the foundation for future comprehensive profitability goals.

Finally, with continuous iterative breakthroughs in large model technology, industries across the board are rapidly integrating large model services. This demonstrates immense commercial potential for both C-end users and B-end enterprises, revolutionizing not only how the public accesses information but also accelerating the digital and intelligent upgrading of countless industries, directly driving a sustained surge in market token usage.

As an AI infrastructure, the average daily token service volume on the SenseTime Core platform has reached 2.42 trillion, with an expected 25-fold increase in service scale throughout 2026, providing a robust infrastructural foundation for the era of Agentic intelligent agents.

Overall, the booming generative AI market has fueled a comprehensive recovery in SenseTime's revenue growth, indicating that the strategy of transitioning to generative AI has borne fruit. Thus, SenseTime has successfully navigated its transition from traditional AI to generative AI, embarking on a new growth cycle.

Yang Jianyong, a contributor to Forbes China, expresses views that are solely his own. He is dedicated to providing in-depth interpretations of cutting-edge technologies such as AI large models, artificial intelligence, the Internet of Things, cloud computing, and smart home appliances.