AI adoption accelerates, but software budgets remain cautious: Saiyi Information’s computing-power business begins contributing profit
GF Securities believes downstream IT budgets are more likely to turn positive in the first quarter of 2027; Sealand Securities is focused on the recovery of Saiyi Information’s core business and the commercialization of its computing-power projects, with subsequent growth dependent on conditions including new-business expansion and cash flow.
Original source：广发证券、国海证券First published on Unhidea：
Key takeaway
There is still a time lag between the adoption of AI applications and an expansion in customers’ overall budgets. Saiyi Information has seen a recovery in its core business and profit contributions from computing power, but Sealand Securities’ high-growth expectations still depend on subsequent business execution.
AI applications are advancing quickly. Why has software demand not strengthened across the board?
China’s manufacturing and commercial distribution systems offer abundant use cases for AI, and downstream acceptance and the range of applications are expanding. However, customers remain restrained in their overall IT spending and are more conservative with spending on traditional IT applications while exploring AI. Accelerating AI penetration can therefore coexist with pressure on traditional software demand.
Liu Xuefeng of GF Securities believes downstream budgets are more likely to show positive changes in the first quarter of 2027. His expectation of a rebound in the software applications sector is also conditional on higher risk appetite and improved liquidity; his operational focus is on a small number of companies expanding rapidly in AI whose core businesses are expected to stabilize and improve.
What operational changes reflect Saiyi Information’s recovery?
In the first three quarters of 2026, Saiyi Information recorded revenue of 1.683 billion yuan, up 12.12% year on year, and net profit attributable to shareholders of 122 million yuan, up 342.87% year on year. Profit grew significantly faster than revenue. The materials also note that market demand for the core business continued to recover and that the first computing-power project began contributing revenue and profit.
As of the end of the third quarter of 2026, the company had delivered a cumulative 255 projects incorporating AI applications, accounting for more than 20% of all projects delivered. AI is entering the delivery process for traditional software implementation projects, with the business connection centered on integration with customers’ existing systems rather than the separate sale of a feature detached from those systems.
What changed as the computing-power business moved from validation to paid operations?
In the first three quarters of 2026, the first high-throughput advanced computing cluster project completed deployment steps including equipment procurement, phased delivery, data-center integration and joint system testing. It formally entered paid commercial operation and generated a stable contribution to revenue and profit in the third quarter. On that basis, Liu Xi of Sealand Securities is optimistic about the computing-power business moving from technical validation to practical deployment.
Around its Shanmou GPT enterprise AI platform, the company is developing applications, industrial models and computing infrastructure. Applications connect with customers’ business systems; industrial models are moving toward standardized products, starting with the PCB industry; and underlying computing capacity provides resources for applications and models. All three business layers have made progress at different stages, but each is at a different stage of development.
How high are Sealand Securities’ growth expectations, and what conditions constrain them?
Liu Xi of Sealand Securities forecasts that Saiyi Information’s revenue in 2026–2028 will be 2.977 billion, 5.959 billion and 8.127 billion yuan, respectively, and that net profit attributable to shareholders will be 220 million, 776 million and 1.046 billion yuan, respectively. The pronounced leap in the forecast is concentrated in 2027, and the growth required exceeds the current scale of operations in the first three quarters.
Paid operation of the first computing-power project provides a starting point for commercialization, while subsequent expansion remains constrained by the pace of the computing-power business’s development. Sealand Securities also highlights risks from customer concentration, rising labor costs, bad debts on receivables and cash flow: beyond revenue and profit growth, project collections and cost control also affect the quality of growth.
