Chinese artificial intelligence developer SenseTime has told investors it expects to swing to a net profit in the first half of 2026, a preliminary result that, if confirmed, would mark a notable turn for a company that reported a 1.489 billion yuan loss in the comparable 2025 period. The forecast, disclosed in a profit alert to the Hong Kong stock exchange, points to a first-half profit range of 500 million to 700 million yuan, compared with a loss of roughly 1.489 billion yuan in the comparable period of 2025.
According to SenseTime’s August 16 Hong Kong exchange profit alert, the company attributed the expected improvement to two main drivers: a narrowing of losses in its core business and fair-value gains on investments in what it described as AI ecosystem enterprises. The alert is a preliminary forecast rather than a finalized interim result, and investors will need to wait for the final interim report to confirm the exact figures and their composition.
A preliminary swing after years of red ink
The forecast range represents a sharp reversal in reported bottom-line performance. Moving from a 1.489 billion yuan loss to a projected profit of 500 million to 700 million yuan implies a headline year-over-year change of roughly 2.0 billion to 2.2 billion yuan. However, because part of the improvement is associated with fair-value gains on investments, the swing does not necessarily indicate that the underlying operating business has reached break-even on its own. Fair-value adjustments can move in either direction in subsequent periods and are sensitive to changes in the valuations of portfolio companies, so the quality of earnings will only become clearer when the company publishes its detailed interim disclosures.
Bamboo Works reported that SenseTime’s non-IFRS loss narrowed by 60% to 70% from a prior-year non-IFRS loss of 1.16 billion yuan. Non-IFRS measures typically strip out share-based compensation and certain one-off items, and they are often used by Chinese technology companies to give investors a view of operating performance that excludes accounting items management considers non-cash or non-recurring. A narrowing of that magnitude, if borne out in the final results, would suggest that the core-business improvement referenced by the company is more than a bookkeeping effect, although it would still leave the non-IFRS line in the red.
The distinction between the headline profit forecast and the non-IFRS loss deserves careful reading. On a reported basis, the company expects to be profitable. On a non-IFRS basis, according to the coverage from Bamboo Works, it still expects to record a loss, just a substantially smaller one. Both statements can be true simultaneously because the reported figure includes items such as investment revaluations that non-IFRS metrics are designed to exclude.
What has changed inside the business
SenseTime’s profit alert attributes the improvement to narrower core-business losses and fair-value gains on investments in AI ecosystem enterprises. Earlier reporting on the company’s strategy, discussed in this analysis of how SenseTime is betting on cost efficiency and overseas markets to stay relevant in China’s AI race, described a company narrowing its focus and pursuing more disciplined spending. The profit alert’s reference to reduced core-business losses is consistent with that direction, although the company has not, in this disclosure, quantified how much of the improvement came from cost reductions versus revenue changes.
The current profit alert does not attribute the expected result to any specific product line, and the company has not, in this disclosure, tied the numbers to any particular generative-AI offering.
The second contributor cited by SenseTime, fair-value gains on investments in AI ecosystem enterprises, points to the value of holdings the company has taken in adjacent AI businesses. As valuations across the Chinese AI startup landscape have shifted, mark-to-market adjustments on those stakes can produce material gains or losses in a given reporting period. Investors will want to see how much of the projected profit comes from this line item and whether those gains are realized or unrealized.
What to watch when the interim report lands
Several questions remain open until SenseTime publishes its full interim results. The first is the split between operating performance and investment gains. A profit driven primarily by revaluations would be more volatile and less repeatable than one driven by improving margins in its operating businesses. The second question is revenue trajectory. The profit alert focuses on the bottom line and does not, in the information disclosed, provide a top-line figure or a segment breakdown. Understanding whether the improvement was accompanied by revenue growth, flat sales, or contraction will be important for judging the durability of the turn.
A third area to watch is cash flow and the balance between research spending and cost discipline. Chinese AI developers face pressure to keep investing in models, chips, and infrastructure at a time when peers are scaling training runs and pushing into new product categories. Sustaining a return to profit while continuing that investment is a harder task than achieving it in a single half-year period.
Finally, market participants will look for management commentary on the second half of 2026 and on how the company sees its competitive position. The profit alert is a limited document by design, and it does not include forward guidance beyond the first-half range. For now, the disclosure indicates that SenseTime expects to report a profitable first half for 2026, subject to the finalization of the interim accounts and the verification of the investment-related gains that helped produce the projected result.