1H26 turnaround beats expectations; model upgrades may drive commercialization, but valuation has reflected some positives
AI summary card
1H26 turnaround beats expectations; model upgrades may drive commercialization, but valuation has reflected some positives
Goldman Sachs raises SenseTime's 12-month target price to HK$2.05 and maintains a Neutral rating, expecting improved cost efficiency and SenseNova upgrades to support earnings recovery.
- The company guided for 1H26 net profit of RMB500 million to RMB700 million, turning around from a net loss of RMB1.5 billion in 1H25 and exceeding Goldman Sachs' prior forecast of a RMB893 million net loss.
- Adjusted net loss is expected to narrow by 60% to 70% year-on-year, mainly driven by cost management, productivity improvements, and better-than-expected non-operating income.
- SenseNova 6.8 Flash Lite Preview adds long-context capabilities, dynamic sub-agent collaboration, and multimodal capabilities for documents, text, charts, and video; Token-based billing is planned for launch in July 2026.
- Goldman Sachs revises its expected net losses for 2026 and 2027 to RMB59 million and RMB336 million, respectively, and lowers its operating expense ratio assumptions.
- A two-stage DCF yields a 12-month target price of HK$2.05; the report believes the current share price implies 7.0x 2027E price-to-sales, with positive factors already largely priced in.
Report interpretation
Overview
SenseTime issued 1H26 earnings guidance, expecting net profit of RMB500 million to RMB700 million, a significant improvement from a RMB1.5 billion net loss in the prior-year period and materially better than Goldman Sachs' previous expectations. Goldman Sachs believes the company is improving profitability through cost controls, productivity gains, and expanded generative AI products, while maintaining a Neutral rating.
Core views
Goldman Sachs' core view is that SenseTime's integrated solution capabilities across computing power, AI models, and AI applications will help it continue expanding its generative AI product coverage. SenseNova model upgrades and a Token-based billing mechanism may drive customer adoption and scaled commercialization; however, the market has already partially reflected the benefits of improved earnings and AI upgrades, supporting a Neutral rating.
Analysis framework
The report revises earnings forecasts and operating expense ratios based on 1H26 profit guidance, uses a two-stage DCF as its primary valuation method, and assesses market pricing through comparison with 2027E price-to-sales multiples.
Methodology notes
Discounting future free cash flows and calculating terminal value
The target price is based on a two-stage DCF, with a WACC of 11.5% and a perpetual growth rate of 2%. Improved earnings expectations increase free cash flow, while WACC has been adjusted slightly from the prior assumption.
Assessing market valuation using forward revenue multiples
The report notes that the current share price implies 7.0x 2027E price-to-sales, while the target price implies 9.6x 2027E price-to-sales, which is used to assess the extent to which positive factors are reflected.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- SenseTime (0020.HK)Covered company
- Strengths
- Significantly improved 1H26 earnings guidance; integrated solutions spanning computing power, AI models, and AI applications; SenseNova upgrades strengthen long-context, sub-agent collaboration, and multimodal capabilities.
- Weaknesses
- Net losses are still expected in 2026 and 2027; earnings improvement partly relies on non-operating income; valuation has already reflected some positive expectations.
- Comparison
- The current share price implies 7.0x 2027E price-to-sales, below the 9.6x implied by the target price, but Goldman Sachs believes positive factors have already been largely priced in.
- Risks
- Slower-than-expected generative AI customer expansion, lower-than-expected customer spending, and intensified market competition.
Key data
- 1H26 Net Profit GuidanceRMB500 million to RMB700 millionA turnaround from a RMB1.5 billion net loss in 1H25.
- Change in Adjusted Net LossNarrowing by 60% to 70% year-on-yearBased on company guidance.
- 2026E Net LossRMB59 millionPrevious expectation was a RMB1.25 billion net loss.
- 2027E Net LossRMB336 millionPrevious expectation was a RMB357 million net loss.
- 12-month Target PriceHK$2.05Previously HK$2.00.
- WACC11.5%Previously 11.6%.
- Perpetual Growth Rate2%Unchanged.
- 2027E Price-to-Sales Ratio7.0xThe target price implies a 2027E price-to-sales ratio of 9.6x.
Impact & implications
If cost-efficiency improvements continue and SenseNova upgrades accelerate customer deployment and usage-based monetization, the company's earnings recovery could be further realized. On the other hand, the report believes valuation already reflects considerable positive expectations, so subsequent share-price performance will depend more on generative AI customer expansion and the realization of actual spending.
Risks
- Generative AI customer adoption and scaling may be slower than expected.
- Customer spending on AI products and services may be lower than expected.
- Market competition may be more intense than expected, potentially affecting pricing, market share, or profitability.
- Non-operating income may fall short of expectations, potentially weakening short-term earnings improvement.
What to watch
- Whether formal 1H26 results and adjusted loss improvement meet guidance.
- Customer adoption, usage, and commercialization progress of SenseNova's Token-based billing plan.
- The sustainability of operating expense ratios, cost management, and productivity improvements.
- Generative AI customer spending, order conversion, and changes in the competitive landscape.
- Whether free cash flow improvement supports DCF valuation assumptions.