Hangzhou's Nano-Core Chip (微纳核芯) announced a 1 billion yuan Series C1 on September 9 for chips designed to run AI models more efficiently. Its intended buyers include phone and PC manufacturers, server suppliers and other hardware companies adding AI capabilities. The company is developing an alternative to repeatedly moving large amounts of information between separate memory and processing components.
That movement is a substantial part of the work behind an AI answer. The company's architecture combines computation inside memory, vertically integrated memory close to computation, and programmable processing based on RISC-V, an open instruction-set standard. In plain language, it tries to do more work where the data already resides. Nano-Core presents this as a route to lower power use and cost without depending exclusively on the newest chip-manufacturing processes; the claimed performance advantage remains a company assertion.
The proposed customer workflow is to add an AI coprocessor alongside a device's main chip rather than replace the entire design. Nano-Core's company page describes that complementary approach, while its product catalog lists LP-CIM inference chips for phones, connected devices and servers. A device maker could use such a component to run a language model locally, with integration into its existing hardware and software. The product proposition is a chip-level capability for other manufacturers, not a consumer chatbot sold directly to users.
Commercial readiness is the next step to watch. In the financing announcement, Nano-Core says it has been completing chips, boards or servers and a full software stack, with customer sample testing across several uses planned during the year. It describes working with memory suppliers, model developers and device manufacturers, but does not identify a quantified book of customer orders. Those development and partnership claims should not be mistaken for established high-volume sales of the new AI chips.
The round includes financial investors such as Sunshine Life and Delian Capital, alongside strategic participants including Meig Smart and Huaqin Technology. Several existing shareholders also invested again. The company's release and independent reporting by the STAR Market Daily confirm the financing. The announced yuan amount is approximately $149 million using the September 10 reference exchange rate; the dollar figure is a conversion, not a separate financing.
Nano-Core's approach makes the relationship between memory, computation and software central to its business. Reducing data movement is useful only if manufacturers can run the models they need on reliable, affordable components. Customer testing will therefore need to establish more than a favorable chip benchmark: it must show that the coprocessor can fit a real product's power budget, software requirements and production schedule.