Revised Full Text
Net profit for the first half surged by up to 288%, with Q2 net profit surging more than twofold quarter-on-quarter. Inspur Information's half-year earnings forecast signals that domestic AI servers have transitioned from concept hype to earnings realization.
On the evening of July 7, Inspur Information, a leading domestic AI server provider, disclosed its 2026 half-year earnings forecast, presenting impressive figures that have drawn significant market attention. The company expects its net profit attributable to shareholders for the first half to reach between 2.6 billion and 3.1 billion RMB, representing a staggering year-on-year increase of 226% to 288%. This performance not only indicates that the half-year profit surpasses last year's full-year figure of 2.413 billion RMB, but also sets a new record for the company's profitability in the same period since its listing.
If the overall first-half figures are impressive enough, the second-quarter performance alone can be described as explosive. Building on a first-quarter net profit of 605 million RMB, Inspur Information's second-quarter net profit is expected to reach between 1.995 billion and 2.495 billion RMB, marking a massive quarter-on-quarter surge of 229% to 312%. This leap from steady growth to soaring performance in a single quarter is by no means a random market blip, but a clear signal of concentrated demand release and the realization of economies of scale.
According to the company's announcement, the substantial earnings growth is primarily attributed to "seizing upward industry opportunities, continuously dedicating efforts to product technological innovation and enhancing customer satisfaction, thereby driving a significant increase in operating performance." Breaking it down, three key factors are at play: first, a dual explosion in global demand for AI large model training and inference, with cloud providers and internet companies continuously increasing investments in computing infrastructure, leading to rapid expansion in AI server procurement; second, the continuous enhancement of the company's product competitiveness, with an ever-improving product matrix of high-end AI servers and general-purpose servers, solidifying its leading position; third, the realization of economies of scale, where profitability has significantly improved alongside the expansion of production and sales volumes.
In fact, institutions had anticipated the boom in domestic AI servers well in advance. Caitong Securities analysis suggests that as domestic super nodes gradually scale up in volume, the value focus for complete machine solution providers is expected to shift upward toward full-stack integration encompassing "system-level architecture + high-speed interconnects + liquid cooling + large-scale delivery." Consequently, the profitability of leading manufacturers equipped with full-stack capabilities is expected to see significant improvement. As one of the few domestic manufacturers possessing full-stack R&D and large-scale mass production capabilities for super nodes, In August 2025, Inspur Information launched the Yuan Nao SD200 super node server supporting 64-way GPU chip interconnection and took the lead in its commercial deployment.
Inspur Information's outstanding performance marks a milestone for domestic AI servers transitioning from expectations to reality. As computing power demand shifts from concept-driven to order-driven, and as domestic substitution evolves from policy-driven to commercially driven, the sector’s growth story may only just be starting. After all, enterprises in China's AI server industry chain are benefiting not only from the development of the global AI industry but also from the accelerated dividends of domestic substitution.