--- title: "A Chinese analog chipmaker buys a clock-chip specialist at a 57.5% discount" date: 2026-11-09 category: Companies & Stack site: NeuroAI canonical: https://neuroai.site/a/na-corp-3peak-aura-clock-chip-deal language: en --- # A Chinese analog chipmaker buys a clock-chip specialist at a 57.5% discount > 3PEAK is paying ≈US$521M / HK$4.07B for 86.98% of Aura Semiconductor, priced 57.5% below the target's 2021 valuation and heavy on IP licensing. The buyer paid far less than the seller's last asking price, and both sides signed anyway. On 7 November 2026, 3PEAK (思瑞浦), a Suzhou-based analog and mixed-signal chip designer listed in Shanghai, disclosed the next step in an acquisition it has been assembling all year. Through a company controlled by an investment partnership it holds a stake in, it will take 86.9792% of Aura Semiconductor (宁波奥拉), a Ningbo clock-and-power chip house, for ¥3.698 billion — ≈US$521M / HK$4.07B. Delivery of the shares is under way. ## The structure, in plain terms 3PEAK is not buying Aura directly. It holds a stake in a partnership called Tongxin Jiayuan (同信嘉远), which set up a wholly owned vehicle, Tonghe Wanying (同和万盈), in August 2026. That vehicle signed share-transfer agreements with Aura's main shareholders and is funded by the partnership's own money plus an acquisition loan from a bank. Two consequences follow. First, the deal is all cash, no shares. 3PEAK said it is not currently considering issuing equity for the transaction, which avoids dilution but loads the structure with debt. Second, Aura will not be consolidated into 3PEAK's accounts for now. The listed company reports an investment result rather than line-by-line revenue and profit — a way to keep a large acquisition off the income statement while integration risk is still unknown. The price implies a 100% equity value of ¥4.252 billion — ≈US$599M / HK$4.68B. Aura carried a ¥10 billion (≈US$1.4B / HK$11B) pre-IPO valuation in 2021, so the new mark is about 57.5% lower. On 2025 net profit, the multiple is roughly 29 times, against about 67 times in the 2021 share transfers. ## Why the discount is not a distress signal Aura's revenue mix explains most of the gap. In 2025 it booked ¥604.9 million in revenue (≈US$85M / HK$665M) and ¥147.1 million in net profit (≈US$21M / HK$162M). Of that revenue, ¥402.8 million — about two-thirds — came from technology licensing (≈US$57M / HK$443M), with only ¥202.2 million from selling chips (≈US$28M / HK$222M). Licensing income is high-margin and lumpy, and it is valued differently from product revenue. 3PEAK's disclosure says the licensing business was priced only on cash flows from contracts already signed, discounting future royalties, while any deal still under negotiation was excluded. The chip business was valued on a going-concern basis with reference to comparable companies. It is a conservative approach, and a negotiating tactic: exclude the upside, argue about the downside, and settle on a number both boards can defend. The licensing book is real, and it has names attached. SiTime, the MEMS oscillator specialist, paid US$270 million in 2023 for a five-year licence to Aura's clock IP; onsemi paid US$144 million in 2025 for overseas sales rights to its multi-phase power designs, according to the Qianwan New Area government in Ningbo. Both are the long-term agreements cited in the filing. That is a rare thing in Chinese semiconductors: a domestic designer licensing technology outward to American incumbents rather than the other way round. It is also the part of the business most exposed to renewal risk, because a licence that expires takes its revenue with it. ## What 3PEAK is actually buying Clock chips and multi-phase power are the two gaps being filled. Aura's line covers jitter-attenuating clocks, clock generators, programmable oscillators, real-time clocks, multi-phase power controllers, envelope-tracking power chips and RF amplifiers. It says it was the first Chinese company to obtain full AMD product certification and enter AMD's reference designs for multi-phase power — a claim that matters because server and AI accelerator boards are designed around those reference platforms. Envelope tracking is the other prize: a power technique that improves amplifier efficiency, difficult to design, and supplied by very few Chinese vendors at volume. 3PEAK's own catalogue runs to more than 3,200 products across signal chain, power management and mixed-signal front ends. Adding clocking and multi-phase power lets it pitch a fuller board-level bill of materials into data-centre, communications and industrial customers instead of losing the socket to a competitor that has the missing part. ## The early evidence is mixed, and interesting Aura's 2026 numbers show the product business accelerating. In the first nine months of 2026 it reported revenue of ¥440.5 million (≈US$62M / HK$485M), with product sales of ¥286.9 million (≈US$40M / HK$316M) — already above the whole of 2025 — while licensing revenue fell to ¥153.6 million (≈US$22M / HK$169M) because fixed-fee licensing income is typically recognised in the fourth quarter. Net profit for the period was ¥66.2 million (≈US$9M / HK$73M), well below the 2025 full-year figure. That is what a mix shift looks like: more hardware, less royalty. For 3PEAK, that shift is the point. Product revenue recurs; licensing does not. ## The risks a reader should price in The performance commitments sit with an employee shareholding platform, not with the controlling shareholder, covering ¥900 million (≈US$127M / HK$990M) across the licensing business. When the seller does not back the numbers with its own balance sheet, the buyer is accepting more of the downside. Integration is the second risk. Two analog companies with different design flows, different customer lists and a licensing culture meeting a product culture do not merge on a spreadsheet. The third is the debt. An acquisition loan inside a partnership structure works while rates are stable and cash flows arrive on schedule. A licensing renewal that slips by two quarters would test it. ## Key takeaways - 3PEAK will acquire 86.9792% of Aura Semiconductor for ¥3.698 billion (≈US$521M / HK$4.07B) through a partnership vehicle, implying a ¥4.252 billion (≈US$599M / HK$4.68B) equity value. - The price is about 57.5% below Aura's 2021 pre-IPO valuation of ¥10 billion (≈US$1.4B / HK$11B), on roughly 29x 2025 profit versus about 67x in 2021. - Aura earned ¥604.9 million (≈US$85M / HK$665M) in 2025, two-thirds from IP licensing to SiTime and onsemi; the deal fills 3PEAK's gaps in clock chips and multi-phase power. ## Honest limitations The transaction terms come from 3PEAK's official disclosure filed on 7 November 2026, as published by China Securities Journal and mirrored on CNINFO, cross-checked against Securities Times and a 科创板日报 interview with the company's securities department. The 2025 and 2026 figures for Aura are explicitly marked unaudited in the filing. The 57.5% discount and the 29x and 67x multiples are arithmetic I performed on disclosed numbers, not company-stated ratios, and depend on which profit figure is used. The claim about AMD reference-design entry is Aura's own. I have not verified whether the SiTime and onsemi licences have renewal dates inside the performance-commitment window. Currency conversions use ¥7.1 to US$1 and ¥1 to HK$1.1 and are rounded. ## What readers can do now - If you source analog parts, ask both companies' distributors whether clock and multi-phase lines will be cross-sold or kept separate; the answer determines whether this is a bundle or two catalogues under one roof. - If you follow semiconductors, watch Aura's licensing renewals rather than its product shipments — a single non-renewed contract moves more profit than a year of design wins. - If you evaluate the deal, note that non-consolidation hides the acquired business inside one investment line; read the notes, not the headline revenue. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-corp-3peak-aura-clock-chip-deal Free to quote with attribution and a link to the original.