香港, 2026年4月22日 - (亞太商訊 via SeaPRwire.com) - 4月20日，雲頂新耀宣佈在2026年美國癌症研究協會（AACR）年會上公佈其自主研發的mRNA個性化腫瘤治療性疫苗EVM16單藥治療及聯合PD-1抑制劑（替雷利珠單抗）治療晚期實體瘤的首次人體臨床試驗（FIH）數據。結果顯示出良好的安全性和耐受性、顯著的免疫原性及積極的初步療效信號，支持進一步臨床開發的推進。據悉，該臨床試驗由北京大學腫瘤醫院與復旦大學附屬腫瘤醫院聯合發起（EVM16CX01，NCT06541639），並於2025年3月完成首例患者給藥。EVM16CX01是一項評估EVM16注射液單藥及聯合PD-1抑制劑治療對晚期或復發實體瘤患者的安全性、耐受性、免疫原性和初步療效的劑量遞增及擴展研究，該試驗採用3+3 劑量遞增設計，包括三個劑量組：0.1 mg、0.3 mg和1.0 mg。入選患者為經標準治療失敗的晚期或復發性實體瘤患者，且至少有一個可測量靶病灶。患者將接受2次EVM16單藥治療（每兩週一次），隨後接受EVM16聯合替雷利珠單抗治療。截至2025年12月7日，共入組9例患者。安全性和耐受性方面，EVM16未觀察到劑量限制性毒性（DLT）。所有患者均報告至少一次與研究藥物相關的不良事件，均為2級及以下，且均可自行緩解。 EVM16在9例患者中有8例誘導出強烈的新抗原特異性T細胞免疫反應，並顯示出劑量依賴性趨勢。初步臨床療效方面，一例既往接受過三線系統治療失敗的胃食管結合部癌患者達到確認的部分緩解（PR），無進展生存期（PFS）為126天。另有2例患者達到疾病穩定(SD)，其中一例既往接受過三線系統治療失敗的非小細胞肺癌患者PFS為88天，另一例既往接受過三線系統治療失敗的食管鱗癌患者截至數據截止日已隨訪112天，尚未觀察到疾病進展。北京大學腫瘤醫院消化腫瘤內科主任、中國臨床腫瘤學會胃癌專家委員會主任委員沈琳教授表示：「EVM16作為一款新型mRNA個性化腫瘤治療性疫苗，在首次人體臨床試驗中展現出良好的臨床開發潛力。數據顯示，該疫苗在晚期實體瘤患者中整體安全性與耐受性良好，未觀察到劑量限制性毒性。在免疫學層面，EVM16能夠誘導新抗原特異性T細胞應答，並呈現一定劑量依賴趨勢，提示其具有較強的免疫啟動能力。在療效探索方面，臨床試驗在既往接受多線標準治療失敗的患者中觀察到初步抗腫瘤活性信號，包括部分緩解（PR）及疾病穩定（SD），顯示其免疫學作用機制具有改善臨床獲益的潛力。總體而言，該臨床試驗初步驗證了EVM16通過新抗原驅動T細胞應答實現腫瘤免疫啟動的臨床價值，為其後續臨床開發提供了重要依據。當前，儘管免疫檢查點抑制劑已顯著改善部分腫瘤的治療格局，但總體獲益人群仍有限，尤其是在多線治療失敗的晚期實體瘤患者中仍存在顯著未滿足需求。EVM16所代表的mRNA個性化腫瘤治療性疫苗治療路徑，有望在現有免疫治療基礎上進一步拓展受益人群，並帶來更多治療選擇。」雲頂新耀首席執行官羅永慶表示：「非常高興看到EVM16的首次人體臨床試驗數據在AACR大會上公佈。此次公佈的在晚期實體瘤患者中展現出良好的安全性與耐受性、療效及作用機制等方面的積極結果，初步驗證了EVM16的治療潛力和雲頂新耀自研mRNA平台的臨床價值。對於晚期實體瘤患者而言，尤其是多線治療失敗人群，仍然缺乏有效且具有針對性的治療選擇。EVM16作為公司自主研發的mRNA個性化腫瘤治療性疫苗，具備誘導免疫啟動並產生新抗原特異性T細胞應答的潛力，有望預防腫瘤轉移或復發，代表了腫瘤免疫治療的重要探索方向。該項目基於公司自主研發的EVER-NEO-1『妙算』腫瘤新抗原人工智慧演算法系統。該系統可基於患者腫瘤特異性突變進行分析，識別具有較高免疫原性的腫瘤新抗原，並生成相應的mRNA疫苗設計方案，且具備自我迭代優化能力，以持續提升新抗原預測與篩選效率。依託國際領先的mRNA平台，我們將加速推進EVM16後續臨床開發工作，推動該創新療法惠及更多患者。」雲頂新耀已成功將經過臨床驗證的自研AI+mRNA平台本地化，構建了一個涵蓋從抗原設計、mRNA序列優化、脂質納米顆粒（LNP）靶向遞送技術到產業化生產的端到端全產業鏈平台，具備了在全球範圍內解決巨大未滿足醫療需求的潛力。EVM16是公司自主研發的mRNA個性化腫瘤治療性疫苗，根據每位患者特有的腫瘤細胞突變，通過EVER-NEO-1「妙算」系統識別具有較高免疫原性的腫瘤新抗原，並設計編碼數十種腫瘤新抗原的mRNA治療性疫苗。EVM16通過LNP靶向遞送系統在體內進行高效的抗原呈遞，啟動患者自身的新抗原特異性T細胞免疫，進而達到殺傷腫瘤細胞和治療癌症的目的。從投資與市場維度來看，個體化腫瘤疫苗尚處早期突破期。根據Grand View Research的預測，全球個性化癌症疫苗市場規模2030年將達14.5億美元，2025-2030年複合增長率高達44.86%。麥高證券指出，mRNA腫瘤疫苗有望實現泛癌種覆蓋，同時兼具高可及性、現貨化與個性化優勢。作為極具潛力的新型腫瘤免疫療法，其可通過廣泛聯用，以輔助療法切入臨床，逐步打開數百億美元的市場空間。在此背景下，雲頂新耀在該領域的佈局正加速落地，逐步構建具備全球競爭力的研發管線。依託自主AI+mRNA平台與國內率先佈局的臨床身位，雲頂新耀正處於產業化與估值成長的關鍵節點。本次EVM16首次人體臨床試驗（FIH）數據的讀出，有望進一步釋放臨床與商業化潛力，進一步吸引市場及投資者持續聚焦，助力公司在腫瘤免疫賽道實現長足的價值跨越。 Copyright 2026 亞太商訊 via SeaPRwire.com. All rights reserved. www.acnnewswire.com

HONG KONG, Apr 22, 2026 - (ACN Newswire via SeaPRwire.com) - Apr 20, Everest Medicines announced that the first-in-human (FIH) clinical trial data of EVM16, a proprietary personalized mRNA cancer vaccine, were presented at the 2026 American Association for Cancer Research Annual Meeting (AACR 2026). The data include results from EVM16 as a monotherapy and in combination with the PD-1 inhibitor tislelizumab in patients with advanced solid tumors. The results demonstrated that EVM16 has a favorable safety and tolerability profile, robust immunogenicity, and promising preliminary efficacy in patients with advanced solid tumors, supporting its further clinical development.The clinical trial (EVM16CX01, NCT06541639) was conducted jointly by Peking University Cancer Hospital and Fudan University Shanghai Cancer Center, with the first patient dosed in March 2025. EVM16CX01 is a first-in-human (FIH), dose-escalation, and expansion study evaluating the safety, tolerability, immunogenicity, and preliminary efficacy of EVM16 injection as a monotherapy and in combination with tislelizumab in patients with advanced solid tumors. The study follows a 3+3 dose-escalation design across three dose levels (0.1 mg, 0.3 mg, and 1.0 mg). Eligible patients are those with advanced or recurrent solid tumors who have failed standard of care and have at least one measurable target lesion. Patients receive 2 doses of EVM16 monotherapy once every two weeks, followed by combination therapy of EVM16 and tislelizumab. As of December 7, 2025, a total of 9 patients had been enrolled.No dose-limiting toxicities (DLTs) were observed. All patients experienced at least one investigational product (IP) related adverse event (AE), all of which were Grade ≤ 2 and resolved spontaneously. EVM16 elicited strong neoantigen-specific T cell responses in 8 of 9 patients, which showed a dose-dependent trend. A gastroesophageal junction cancer patient who failed 3 prior lines of systemic therapy achieved a confirmed partial response and a PFS of 126 days. Another 2 patients achieved stable disease. A non-small cell lung cancer patient who failed 3 prior lines of systemic therapy achieved a PFS of 88 days, and an esophageal squamous cell carcinoma patient who failed 3 prior lines of systemic therapy has been followed up for 112 days to date and has not experienced disease progression.“As a novel personalized mRNA cancer vaccine, EVM16 has demonstrated encouraging clinical development potential in its first-in-human clinical trial,” said Professor Shen Lin, Director of the Gastrointestinal Oncology Department at Beijing Cancer Hospital and Chair of the Gastric Cancer Expert Committee of the Chinese Society of Clinical Oncology. “The data show a favorable safety and tolerability profile in patients with advanced solid tumors, with no dose-limiting toxicities (DLTs) observed. Importantly, EVM16 stimulated neoantigen-specific T cell responses with a dose-dependent trend, highlighting its consistent immune activation capability. Preliminary anti-tumor activity signals, including partial response (PR) and stable disease (SD), were observed in patients who had failed multiple lines of standard-of-care therapies, suggesting potential clinical activity. These findings provide early clinical validation of EVM16 and reinforce its differentiated mechanism as a personalized mRNA cancer vaccine. While immunotherapies, including checkpoint inhibitors, have transformed the treatment landscape in selected tumor types, most patients, particularly those with heavily pretreated advanced solid tumors, still represent a significant unmet medical need. Personalized mRNA cancer vaccines represented by EVM16 may help expand the population benefiting from immunotherapy and provide additional treatment options.” “The favorable safety, tolerability, immunogenicity, and preliminary efficacy results for EVM16 provide initial evidence of its therapeutic potential and support the clinical value of our in-house mRNA platform,” said Mr. Rogers Yongqing Luo, Chief Executive Officer of Everest Medicines. “Patients with advanced solid tumors, particularly those who have failed multiple lines of therapy, continue to suffer from limited treatment options. As our first in-house developed personalized mRNA cancer vaccine, EVM16 demonstrates the ability to induce immune activation and generate neoantigen-specific T cell responses, which may help reduce the risk of tumor metastasis or recurrence. This further underscores its potential as a novel approach in cancer immunotherapy. This project is powered by EVER-NEO-1, Everest’s proprietary AI-based neoantigen prediction algorithm. It analyzes patient-specific tumor mutations to identify highly immunogenic neoantigens and support the design of personalized mRNA vaccine candidates. The algorithm incorporates iterative learning capabilities designed to continuously improve neoantigen prediction accuracy and selection efficiency. Leveraging our leading mRNA platform, we will continue to advance the clinical development of EVM16 and bring this innovative therapy to more patients.”Everest Medicines has localized its clinically validated proprietary AI+mRNA platform, establishing an end-to-end integrated platform spanning antigen design, mRNA sequence optimization, lipid nanoparticles (LNP) targeted delivery and scalable manufacturing, with the potential to address significant unmet medical needs globally.EVM16 is a novel personalized therapeutic mRNA cancer vaccine in-house developed by Everest. It contains neoantigens with high immunogenicity potential, predicted based on the unique tumor mutations of each patient using Everest’s proprietary AI-based neoantigen prediction algorithm, EVER-NEO-1. The vaccine is designed to encode dozens of tumor neoantigens. The vaccine uses a lipid nanoparticle (LNP) delivery system to efficiently deliver neoantigen-encoded mRNA in vivo, activating neoantigen-specific tumor-killing T cells and inhibiting tumor growth.From an investment and market perspective, personalized cancer vaccines are in an early stage of breakthrough. According to Grand View Research, the global personalized cancer vaccine market is projected to reach $1.45 billion by 2030, with a staggering compound annual growth rate (CAGR) of 44.86% from 2025 to 2030. Myguide Securities points out that mRNA cancer vaccines have the potential to achieve pan-cancer coverage while offering the advantages of high accessibility, “off-the-shelf” availability, and personalization. As a highly promising new form of immunotherapy, it can enter clinical use as adjuvant therapies through extensive combinations, gradually unlocking a market space worth tens of billions of dollars.Against this backdrop, Everest Medicines is accelerating its strategic positioning in this field, progressively building a research and development pipeline with global competitiveness. Leveraging its proprietary AI+mRNA platform and its leading clinical position in China, Everest Medicines is currently at a critical juncture for industrialization and valuation growth. The upcoming readout of first-in-human (FIH) clinical trial data for EVM16 is expected to further release clinical and commercial potential, continuing to attract focus from the market and investors, and helping the company achieve a significant leap in value within the immunotherapy sector. Copyright 2026 ACN Newswire via SeaPRwire.com. All rights reserved. www.acnnewswire.com

香港, 2026年4月22日 - (亞太商訊 via SeaPRwire.com) - 在創新藥研發競爭日趨激烈的當下，具備自主研發能力與全產業鏈佈局的中國生物製藥企業正加速走向國際資本市場。邁威（上海）生物科技股份有限公司（以下簡稱"邁威生物"）於4月20日開啟公開招股，擬在香港主機板上市，進一步拓展全球研發與商業化網路。2025年，邁威生物在業務拓展領域接連取得里程碑式突破，先後完成多項覆蓋小核酸、IL-11靶向療法及長效G-CSF等前沿技術領域的授權合作，累計潛在總金額超過16億美元。這一系列合作不僅驗證了公司平臺的研發實力，也為其後續全球臨床開發與商業化提供了充沛的資金支持。正是憑藉這些重磅合作的持續落地，邁威生物的全球商業化能力逐步凸顯，其技術價值正加速轉化為可量化的商業回報。對外授權成果顯著，驗證全球研發價值邁威生物的研發實力已獲得國際製藥巨頭的真金白銀認可，其卓越的商業拓展能力成為驅動公司價值增長的核心引擎。2025年，公司總收入達到6.59億元人民幣，同比增長超過2.3倍。其中，對外授權收入達4.09億元，同比激增超過6倍，這主要得益於與多家國際知名生物科技公司的重磅合作。2025年6月，公司就9MW3811（一款靶向IL-11的人源化單抗）與穀歌系公司Calico達成獨家授權合約，授權其在除大中華區外的全球範圍內開發和商業化。Calico長期專注於"抗衰老"及年齡相關疾病的生物學研究，在該領域擁有深厚的技術積累，此次合作將加速9MW3811的全球臨床開發，進一步提升其在纖維化疾病及癌症治療領域的國際影響力。9MW3811靶向的IL-11是一種關鍵的促纖維化細胞因數，在肺纖維化、皮膚瘢痕、腎纖維化、肝纖維化等多種病理進程中扮演核心角色。目前，9MW3811不僅已獲批開展特發性肺纖維化臨床試驗，更于2025年底成為全球首款進入病理性瘢痕（包括增生性瘢痕和瘢痕疙瘩）治療臨床階段的IL-11靶向藥物，為其未來拓展更多纖維化適應症打開了空間。同年10月，邁威生物進一步拓展技術邊界，與專注於NewCo孵化的專業機構Aditum Bio就新型雙靶點siRNA候選藥物達成獨家授權合約。Aditum Bio由前諾華高管創立，以其"精准篩選具有巨大潛力的資產，快速推進至臨床概念驗證，將項目打造成大型製藥企業青睞的收購目標"的成熟模式而聞名。此次合作不僅為邁威生物在小核酸藥物領域的前瞻性佈局引入了國際頂級專業資本的背書，更重要的是，Aditum Bio高效的臨床推進能力將大幅縮短該管線的驗證週期，有望加速其臨床轉化進程。深耕相關領域的合作方往往推進意向明確。邁威生物與Disc Medicine就9MW3011（一款靶向TMPRSS6的單克隆抗體）達成的合作正是這一邏輯的體現。雙方於2023年1月簽署獨家授權合約，合同總價值高達4.1億美元。9MW3011通過調節鐵調素水準控制鐵超載，為真性紅細胞增多症這一尚無有效改良療法的罕見血液病提供了全新治療思路。Disc已於2025年9月啟動該產品治療真性紅細胞增多症的II期臨床試驗，至此公司累計收到首付款及I、II期里程碑付款合計2,500萬美元。合作的高效推進不僅驗證了靶點價值，也為後續更高金額的開發及商業化里程碑奠定了基礎。與聚焦早期研發管線的授權合作不同，邁威生物與齊魯製藥圍繞已上市產品邁粒生(R)（長效G-CSF）達成大中華區獨家授權合作，借助成熟藥企的商業化能力，實現優勢互補與風險共擔。齊魯製藥在中國腫瘤領域擁有強大的商業化網路和市場滲透能力，雙方聯手既能加速邁粒生(R)的市場放量，又能豐富齊魯製藥的產品組合，最大化共用市場紅利。技術築底：源頭創新驅動，ADC與TCE雙引擎發力密集的BD成果背後，是邁威生物強大的源頭創新能力。公司已建立起以自主智慧財產權的定點偶聯技術IDDC™為核心的ADC藥物開發平臺，以及一體化高效抗體發現平臺和T細胞銜接器（TCE）雙/三特異性抗體開發平臺。其中，IDDC™平臺依託四大核心技術DARfinity(R)（定點偶聯）、IDconnect™（穩定連接子分子）、Mtoxin™（新型毒性分子）和LysOnly™（條件釋放結構），能夠開發出更均勻、更穩定、純度更高、潛在療效更好、更安全的優化ADC藥物。以核心產品9MW2821（靶向Nectin-4 ADC）為例，該平臺通過定點偶聯工藝，使產品DAR值均一性超過95%，顯著優於傳統隨機偶聯技術，更均一的分子結構帶來了更好的穩定性和安全性。值得關注的是，根據美國FDA官網及權威公開資料，已上市的同靶點藥物Padcev帶有黑框警告，主要涉及嚴重皮膚反應；而根據邁威生物招股書披露的臨床資料，9MW2821在臨床試驗中展現出良好的安全性，未觀察到同類的嚴重不良事件。療效方面，9MW2821在尿路上皮癌、三陰性乳腺癌等適應症上均表現出優於Padcev的客觀緩解率（ORR）趨勢，展現其"同類最佳"潛力。2024年以來，該產品已獲得國家藥監局突破性療法認定及FDA多項快速通道與孤兒藥資格，國際監管認可度持續提升。邁威生物ADC平臺的另一個代表性產品是7MW4911（靶向CDH17 ADC）。 該產品針對的適應症是國內高發的胃腸道腫瘤，患者基數龐大，現有治療手段仍存在耐藥等局限。為應對這一挑戰，7MW4911不僅結合創新單抗Mab0727、新型連接子，還採用了專有的Mtoxin™載荷，這是一種專為克服ABC轉運蛋白介導的多藥耐藥機制而設計的強效DNA拓撲異構酶I抑制劑，能有效殺傷耐藥腫瘤細胞。目前，7MW4911單藥療法已於2025年11月在中國啟動I期臨床試驗，並於2026年1月在美國啟動I/II期臨床試驗。在TCE領域，邁威生物也取得了實質性進展。公司的6MW5311（靶向CD3/LILRB4）已于近日向國家藥監局提交IND申請並獲受理，同時其美國臨床試驗申請目前處於pre-IND階段，計畫於2026年第二季度向FDA正式遞交申請。該管線用於治療復發/難治性急性髓系白血病、慢性粒單核細胞白血病以及復發或難治性多發性骨髓瘤，其獨特優勢在於：在無腫瘤環境下對T細胞的結合活性極低，而在腫瘤與T細胞共存的微環境中則對腫瘤細胞表現出強效殺傷作用，從而在保障療效的同時顯著提高了安全性。TCE與ADC在腫瘤治療中機制互補，前者啟動T細胞免疫，後者精准殺傷腫瘤細胞。展望未來，邁威生物有望探索二者的聯用方案，在腫瘤領域構建起更為廣闊和立體的治療矩陣，釋放巨大的管線協同效應。結語 全球腫瘤藥物市場預計2032年將突破5,400億美元，ADC、TCE等新一代抗體類藥物正重塑精准治療格局，罕見病與纖維化等領域仍存在巨大未滿足需求。憑藉扎實的技術平臺、前瞻性的管線佈局以及卓越的全球商業拓展能力，隨著"A+H"兩地上市落地，邁威生物將擁有更靈活的資本工具，加速核心產品的全球多中心臨床試驗、擴充生產產能、深化國際合作。展望未來，公司不僅有望在尿路上皮癌、宮頸癌、三陰性乳腺癌等領域實現關鍵上市，更將在纖維化疾病、罕見血液病、核酸藥物等新賽道持續佈局，成長為輻射全球的創新型生物製藥領軍企業。 Copyright 2026 亞太商訊 via SeaPRwire.com. All rights reserved. www.acnnewswire.com

SUNNYVALE, CA, Apr 22, 2026 - (ACN Newswire via SeaPRwire.com) - Graid Technology, the pioneer in GPU-accelerated NVMe storage, today announced its Agentic AI Storage Portfolio: a purpose-built family of KV cache solutions designed to eliminate the storage bottleneck that stalls "always-on" production AI. The portfolio spans three deployment tiers: KV Cache Server, KV Cache Rack, and KV Cache Platform, all built on SupremeRAID™ technology. KV Cache Platform, the portfolio's highest tier, is purpose-aligned to NVIDIA's STX reference architecture, with native BlueField-4 DPU execution on the roadmap for H2 2026.As agentic AI moves from experimentation to production, the infrastructure assumptions that underpinned single-shot inference have broken down. Models running continuous multi-step tasks and maintaining context across hours of operation generate KV cache demands that overwhelm GPU HBM. The result: latency spikes up to 18x, GPU utilization as low as 50%, and model-level failures, including hallucinations and reasoning degradation, that are difficult to detect and costly to recover from.SupremeRAID™ addresses this directly, aggregating up to 32 NVMe drives into a single 280 GB/s virtual pool, bypassing the CPU via GPU Direct Storage, and delivering KV cache reads at 1.3ms- 77x faster than standard NVMe. The three portfolio tiers bring this capability to every deployment scale:KV Cache Server - single-node NVMe acceleration for individual inference servers and edge AI deployments. Available now.KV Cache Rack - rack-scale, partner-validated solutions co-engineered with leading server OEM partners for enterprise multi-GPU clusters. Available now.KV Cache Platform - Purpose-built for NVIDIA's STX reference architecture, with native BlueField-4 DPU execution and rack-scale storage expansion on the roadmap."A year ago, at GTC 2025, Jensen Huang predicted that storage would become GPU-accelerated for the first time. This year, NVIDIA turned that concept into an architecture with STX and CMX," said Leander Yu, CEO of Graid Technology. "Our KV Cache Portfolio is built for precisely this moment, delivering the storage performance that agentic AI demands, at storage-tier economics."For enterprises and infrastructure teams evaluating agentic AI deployments, the full deployment architecture, technical specifications, and NVIDIA STX compatibility details are available in the solution brief: Graid Technology Agentic AI Storage Portfolio: Purpose-built KV Cache Solutions for Inference at Scale.To learn more about Graid Technology's AI offerings, visit graidtech.com/ai.Media Inquiries:Andrea Eaken, Sr. Director of Marketing, Americas & EMEAandrea.eaken@graidtech.comAbout Graid TechnologyGraid Technology is building the storage backbone for the future of AI, enterprise, and high-performance computing. As the creator of SupremeRAID™, the world's first and only GPU-based RAID, and the global steward of Intel® Virtual RAID on CPU (Intel® VROC), Graid Technology delivers flexible RAID solutions that maximize NVMe performance while ensuring resilient, scalable data protection for modern data infrastructure. Headquartered in Silicon Valley with global operations and R&D in Taiwan, Graid Technology is advancing RAID innovation for the next generation of data-intensive workloads. To learn more, visit graidtech.com.SOURCE: Graid Technology Inc. Copyright 2026 ACN Newswire via SeaPRwire.com. All rights reserved. www.acnnewswire.com

TSUKUBA, Japan, Apr 22, 2026 - (ACN Newswire via SeaPRwire.com) - Discovering and characterizing new materials is important for unlocking advances in fields like clean energy, advanced manufacturing, and improved infrastructure. Researchers use machine learning and other computational tools to help them, but the trial-and-error nature of the process creates specific challenges. The research produces large amounts of experimental and computational data, and scientists need tools that can track and store not only the results but also the chain of reasoning behind them.A newly developed system tracks and stores not only the results but also the chain of reasoning behind them, allowing researchers to review the decision making process for a greater transparency and reproducibility in materials science research.A new system called pinax, published in the journal Science and Technology of Advanced Materials: Methods, provides precisely those features. Developed by engineers at Japan’s National Institute for Materials Science (NIMS), pinax captures the entire process of developing new materials, including machine learning workflows and decision-making processes. “By formalizing both successful and unsuccessful trial-and-error processes, pinax enhances reproducibility, accountability, and knowledge sharing while maintaining strict data governance,” says Satoshi Minamoto of NIMS, the study’s lead author.The new pinax system consists of three layers: the core machine learning infrastructure (bottom), the provenance recording and tracking that visualises the reasoning behind final results (middle), and the advanced feature layer for materials development (top). Credit: STAM-MMachine learning models are playing an ever-larger role in materials discovery and characterization. While the models are powerful tools, the reasoning processes they use are generally opaque. Researchers don’t know what considerations and trial-and-error processes went into their final predictions. “The system introduced in this study visualizes these invisible processes. This enables others to review, verify, and build upon the path to the conclusions,” says Minamoto.Minamoto highlights the importance of such access in applications where safety, reproducibility, and accountability are important, saying that this work “demonstrates how transparent AI systems can transform scientific discovery into a more reliable, efficient, and socially responsible endeavor.”The team tested pinax using two case studies: one on predicting steel properties and another using transfer learning to predict the thermal conductivity of polymers. The system made it possible to link the model’s performance predictions to the specific data or model aspects that influenced them, and to reproduce complex, multi-stage workflows. “In particular, the transfer-learning example highlights pinax’s ability to track how information flows between intertwined datasets and models, making every step in the reasoning process explicitly traceable,” says Minamoto.The engineers plan to expand pinax towards an autonomous, closed-loop materials discovery system. By integrating pinax’s tracking capabilities with automated experimental and simulation systems, they aim to create a loop that can use data generation, machine learning models, and decision-making systems to systematically and independently carry out the entire research cycle.Further information Satoshi MinamotoNational Institute for Materials Scienceminamoto.satoshi@nims.go.jpPaper: https://doi.org/10.1080/27660400.2026.2629051 About Science and Technology of Advanced Materials: Methods (STAM-M)STAM Methods is an open access sister journal of Science and Technology of Advanced Materials (STAM), and focuses on emergent methods and tools for improving and/or accelerating materials developments, such as methodology, apparatus, instrumentation, modeling, high-through put data collection, materials/process informatics, databases, and programming. https://www.tandfonline.com/STAM-MDr Kazuya SaitoSTAM Methods Publishing Director SAITO.Kazuya@nims.go.jpPress release distributed by Asia Research News for Science and Technology of Advanced Materials. Copyright 2026 ACN Newswire via SeaPRwire.com. All rights reserved. www.acnnewswire.com

加州桑尼維爾, 2026年4月22日 - (亞太商訊 via SeaPRwire.com) - 身為 GPU 加速 NVMe 儲存技術先驅的 Graid Technology 今日宣布推出其「Agentic AI 儲存產品組合」：這是一系列專為消除阻礙「全天候運行」生產級 AI 的儲存瓶頸而設計的 KV 快取解決方案。該產品組合涵蓋三個部署層級：KV Cache Server、KV Cache Rack 以及 KV Cache Platform，均基於 SupremeRAID™ 技術打造。作為產品組合中最高階的 KV Cache Platform，其設計專為 NVIDIA 的 STX 參考架構而優化，並已將原生 BlueField-4 DPU 執行功能納入 2026 年下半年產品路線圖中。隨著自主型 AI 從實驗階段邁向生產環境，支撐單次推論的基礎架構假設已不復成立。執行連續多步驟任務並在數小時運作中維持上下文的模型，會產生遠超 GPU HBM 處理能力的 KV 快取需求。其結果是：延遲飆升至最高 18 倍，GPU 利用率低至 50%，並引發難以偵測且修復成本高昂的模型層級故障，包括幻覺和推理能力退化。SupremeRAID™ 直接解決了此問題，可將多達 32 顆 NVMe 硬碟彙整成單一 280 GB/s 的虛擬儲存池，透過 GPU Direct Storage 繞過 CPU，並以 1.3 毫秒的速度執行 KV 快取讀取，速度比標準 NVMe 快 77 倍。這三個產品層級讓這項功能能適用於各種部署規模：KV Cache Server —— 專為單一推論伺服器及邊緣 AI 部署設計的單節點 NVMe 加速解決方案。現已上市。KV Cache Rack —— 與領先的伺服器 OEM 合作夥伴共同設計、經合作夥伴驗證的機架級解決方案，適用於企業級多 GPU 叢集。現已上市。KV Cache Platform —— 專為 NVIDIA 的 STX 參考架構打造，具備原生 BlueField-4 DPU 執行能力，並已將機架級儲存擴充納入發展藍圖。「一年前，在 GTC 2025 大會上，黃仁勳預言儲存將首次實現 GPU 加速。今年，NVIDIA 透過 STX 和 CMX 將這個概念轉化為具體的架構，」Graid Technology 執行長 Leander Yu 表示。「我們的 KV Cache 產品組合正是為此時刻而生，能在符合儲存層成本效益的前提下，提供代理式人工智慧所需的儲存效能。」對於正在評估代理式 AI 部署的企業及基礎設施團隊，完整的部署架構、技術規格以及 NVIDIA STX 相容性詳情，請參閱解決方案簡報：《Graid Technology 代理式 AI 儲存產品組合：專為大規模推論設計的 KV 快取解決方案》。如欲進一步了解 Graid Technology 的人工智慧解決方案，請造訪 graidtech.com/ai。媒體洽詢：Andrea Eaken，美洲及歐洲、中東與非洲地區行銷資深總監 andrea.eaken@graidtech.com關於 Graid TechnologyGraid Technology 正致力於為人工智慧、企業級應用及高效能運算的未來打造儲存基礎架構。作為全球首創且唯一的 GPU 基於 RAID 技術 SupremeRAID™ 的開發者，以及 Intel® Virtual RAID on CPU (Intel® VROC) 的全球技術維護者，Graid Technology 提供靈活的 RAID 解決方案，在確保現代資料基礎架構具備彈性且可擴展的資料保護能力之同時，亦能將 NVMe 效能發揮至極致。Graid Technology 總部位於矽谷，業務遍及全球，並在台灣設有研發中心，致力於推動 RAID 技術創新，以滿足下一代數據密集型工作負載的需求。欲了解更多資訊，請造訪 graidtech.com。來源：Graid Technology Inc. Copyright 2026 亞太商訊 via SeaPRwire.com. All rights reserved. www.acnnewswire.com