According to reports from multiple media outlets including Jiwei and ChinaVenture, AaltoSemi announced the completion of its Series A2 financing round, raising nearly CNY 1.5 billion.
Investors in this round include multiple industrial strategic investors such as Victory Giant Technology and Rockchip Electronics, state-owned capital platforms at both the Jiangsu provincial and municipal levels, and well-known financial investment institutions. Existing shareholders, including WuYueFeng Capital, Xiamen Lianhe Capital, Yuexiu Industrial Fund, Heli Capital, Teda Kechuang, Lingqing Investment, and Menghai Investment, continued to increase their investments. China Renaissance Securities, under China Renaissance Capital, served as the exclusive financial advisor for this financing round. To date, the company has received cumulative support from social capital exceeding CNY 5 billion.
Looking at the list of investing institutions, there are industrial players like Victory Giant Technology and Rockchip Electronics, state-owned capital from both Jiangsu Province and Nanjing City, as well as a group of existing shareholders who have followed the company for more than one round, including WuYueFeng Capital, Xiamen Lianhe, Yuexiu Industrial Fund, and Heli Capital.
| Where It Stands in the Industry Chain
Packaging substrates are referred to as the "foundation" of chips, responsible for transitioning the nanometer-scale circuits of the chip to the micrometer-scale world of the PCB. It may sound like a supporting role, but from the perspective of AI computing power, it has become the bottleneck.
High-end FCBGA substrates for AI server CPUs/GPUs and high-end mobile SoCs have long been dominated by manufacturers in Japan, South Korea, and the Taiwan region of China. The barrier to entry lies not in design, but in the long certification cycles, difficult yield ramp-up, and heavy capital expenditures.
AaltoSemi is tackling the most difficult tier. Its products cover FCBGA, FCCSP, and Coreless ETS, spanning three AI scenarios: cloud, edge, and endpoint. On the computing side, it serves AI server CPUs/GPUs and high-end SoCs; on the storage side, it provides substrates for server storage controller chips; and on the transmission side, it offers substrates for optical module chips. This forms a complete "compute, store, and transport" matrix, rather than a single-point product.
| The Pace Achieved in Five Years
According to the company's disclosure, the production line was operational in May 2023, the main factory building was topped out in 59 days, and core equipment was moved in six months after construction began.
On the customer front, according to the company's disclosure, samples passed the certification of leading domestic OSATs such as JCET, HT-Tech, and TFME seven months after submission. Ten months after sample submission, the company entered the ASE Group system, obtaining qualified supplier status for ASE and SPIL, and achieving mass production (the above timeline is currently mainly from the company's unilateral disclosure, and independent cross-verification is still lacking in public materials). It has now covered the top ten non-US global packaging manufacturers. The company claims that the export ratio will be nearly 70% in 2025 (this data also comes from the company's disclosure, with no third-party sources to corroborate it).
In terms of capacity, the company has planned a 415-mu production base with a total investment of over CNY 10 billion. Cleanroom equipment for the P2 factory is being installed, and the production line is expected to be operational in May 2027; the P3 factory will break ground in 2027 and begin mass production in 2028. The blueprint from P3 to P5 has also been released—essentially a pace of one new factory per year.
In terms of process nodes, 5-micrometer line width and spacing have passed technical verification, 7-micrometer samples have been sent to core customers, and 8-micrometer has entered large-scale stable mass production; current capacity is heavily shifted towards supporting substrates for 6nm, 4nm, and 3nm, and supporting substrates for 2nm are expected to enter mass production within 2026.
| Looking at the Investor Structure
Relying on patient capital from state-owned funds for long-cycle semiconductor projects has been a common practice in Jiangsu in recent years. Let us try to analyze, representing only personal views, why the two industrial players decided to invest.
Victory Giant Technology is a leading PCB manufacturer, standing on the demand and process sides of substrates; Rockchip Electronics is a chip design company, standing upstream in chip definition. One manages whether downstream applications can absorb it, and the other manages how upstream design can coordinate. The simultaneous equity investment by both suggests, in the author's view, that the company's substrates are no longer just samples "that can be made in a lab," but have entered the product definition processes of others. It should be noted that this is an inference based on industrial logic, not a verifiable fact.
Looking at other existing shareholders—Teda Kechuang, Lingqing Investment, Menghai Investment, etc.—all followed up with additional investments. I once consulted professionals in the investment circle, who said that if existing shareholders in the primary market are willing to add to their investments, it is usually because the commitments from the previous round were fulfilled. This is a rule of thumb in the investment circle, not an absolute rule. Another detail to note is that the company disclosed that employees have paid in over CNY 200 million with their own funds, which is uncommon in such asset-heavy companies.
| Where the Money Will Be Spent
The funds from this round will go to three destinations: expanding high-end FCBGA capacity, researching and developing glass substrates, and co-researching the domestic substitution of core materials.
None of the three directions can yield short-term results.
Capacity expansion requires heavy asset investment; glass substrates are the focus of competition for next-generation advanced packaging, and the world is still in the early stages; the domestic substitution of core materials is an even tougher challenge—the company claims to have completed the R&D and process verification of multiple high-end copper-clad laminates and build-up films in collaboration with domestic material manufacturers.
This also explains why it still needs CNY 1.5 billion at the Series A2 round, rather than negotiating an IPO.
According to Morgan Stanley research, the global high-end substrate market will remain roughly balanced overall in 2026, only entering a state of continuous supply shortage starting in 2027, and the gap will continue to widen up to 2030. AI substrates have much higher requirements for layer count, size, line precision, and warpage control than traditional products, while production line construction, certification, and yield ramp-up are all slow, making it difficult to expand effective capacity. Therefore, domestic manufacturers with stable mass production capabilities have become scarce.
| The Gap
The company itself admits that there is still a gap compared to international leaders such as Ibiden and Samsung Electro-Mechanics in terms of overall capacity scale and independent supporting of high-end materials.
It usually takes several years for FCBGA to go from certification to full production. Between obtaining customer qualifications and truly filling the production lines, there is still the hurdle of yield ramp-up. The "mass production yield comparable to the global leading level" mentioned in the press release currently lacks third-party data.
In five years, the company has accumulated over CNY 3 billion in equity financing, with over CNY 5 billion in available funds, two factories in operation, and three on the drawing board. A company founded only in 2021 dares to include the mass production timeline for 2nm substrates and the operational timeline for new factories in its public releases. This timeline itself is its trump card—whether it can be fulfilled will be known in a year.
Editor: Xiao Ming from Shishixinyu