markets.financialcontent.com, Oct. 08, 2025 –
Shanghai, China – October 8, 2025 – VeriSilicon Holdings Co., Ltd. (SHA: 688521), a leading platform-based, all-around, custom silicon solutions provider, has reported an astounding preliminary third-quarter 2025 revenue, more than doubling to 1.28 billion yuan (approximately US$179.7 million). This colossal 120% quarter-over-quarter surge, and a robust 78.77% increase year-on-year, unequivocally signals the insatiable global appetite for specialized AI computing power, cementing VeriSilicon's pivotal role in the burgeoning artificial intelligence landscape and the broader semiconductor industry. The company's exceptional performance underscores a critical trend: as AI models grow more complex and pervasive, the demand for highly optimized, custom silicon solutions is not just growing—it's exploding, directly translating into unprecedented financial gains for key enablers like VeriSilicon.
The dramatic revenue jump and a record-high order backlog of RMB 3.025 billion by the end of Q2 2025, continuing into Q3, are a direct reflection of intensified AI development across various sectors. VeriSilicon's unique Silicon Platform as a Service (SiPaaS) business model, combined with its extensive portfolio of processor intellectual property (IP), has positioned it as an indispensable partner for companies seeking to integrate advanced AI capabilities into their products. This financial triumph is not merely a corporate success story but a powerful indicator of the current state of AI hardware acceleration, highlighting the rapid pace at which the industry is evolving to meet the computational demands of next-generation AI applications, from edge devices to cloud infrastructure.
AI's Computational Engine: VeriSilicon's IP at the Forefront
VeriSilicon's recent financial disclosures paint a clear picture of AI as the primary catalyst for its phenomenal growth. A staggering 64% of new orders secured in Q3 2025 were directly attributed to AI computing power, with AI-related revenue comprising a significant 65% of all new orders during the same period. This highlights a strategic shift where VeriSilicon's deep expertise in custom chip design and IP licensing is directly fueling the AI revolution. The company’s comprehensive suite of six core processing IPs—Neural Network Processing Unit (NPU), Graphics Processing Unit (GPU), Video Processing Unit (VPU), Digital Signal Processing (DSP), Image Signal Processing (ISP), and Display Processing IP—forms the backbone of its AI strategy.
Specifically, VeriSilicon's NPU IP has been a cornerstone, now embedded in over 100 million AI chips globally, adopted by 82 clients in 142 AI chips as of 2024. This widespread adoption underscores its effectiveness in handling diverse AI operations, from computer vision to complex neural network computations. A notable advancement in June 2025 was the announcement of an ultra-low energy NPU capable of over 40 TOPS (Tera Operations Per Second) for on-device Large Language Model (LLM) inference in mobile applications, demonstrating a critical step towards ubiquitous, efficient AI. Furthermore, the company’s specialized AI-based image processing IPs, AINR1000/2000 (AI Noise Reduction) and AISR1000/2000 (AI Super Resolution), launched in February 2025, are enhancing applications in surveillance, automotive vision, cloud gaming, and real-time video analytics by leveraging proprietary AI pixel processing algorithms. This robust and evolving IP portfolio, coupled with custom chip design services, sets VeriSilicon apart, enabling it to deliver tailored solutions that surpass the capabilities of generic processors for specific AI workloads.
Reshaping the AI Ecosystem: Beneficiaries and Competitive Dynamics
VeriSilicon's surging success has profound implications for a wide array of AI companies, tech giants, and startups. Its "one-stop" SiPaaS model, which integrates IP licensing, custom silicon design, and advanced packaging services, significantly lowers the barrier to entry for companies looking to develop highly specialized AI hardware. This model particularly benefits startups and mid-sized tech firms that may lack the extensive resources of larger players for in-house chip design, allowing them to rapidly iterate and bring innovative AI-powered products to market. Tech giants also benefit by leveraging VeriSilicon's IP to accelerate their custom silicon projects, ensuring optimal performance and power efficiency for their AI infrastructure and devices.