輝達釋出免費 AI 模型與阿里巴巴自研晶片,全球算力競賽加劇Nvidia Releases Free AI Model and Alibaba Unveils Proprietary Chip, Intensifying Global Computing Power Race

輝達釋出免費 AI 模型與阿里巴巴自研晶片,全球算力競賽加劇

Nvidia Releases Free AI Model and Alibaba Unveils Proprietary Chip, Intensifying Global Computing Power Race

輝達預計年底釋出 Nemotron 4 模型,阿里巴巴則發表自研 AI 晶片「真武 V900」。隨全球 AI 支出預測上修至 2.67 兆美元,亞太地區正處於技術自主與國際結盟的戰略轉折點。

Nvidia is expected to release the Nemotron 4 model by the end of the year, while Alibaba has announced its self-developed AI chip, the "Zhenwu V900." With global AI spending forecasts revised upward to $2.67 trillion, the Asia-Pacific region stands at a strategic turning point between technological autonomy and international alliances.

全球人工智慧產業正進入高度競爭與資本密集化的新階段。根據高德納諮詢公司(Gartner)最新報告,2026 年全球 AI 領域總支出預計將攀升至 2.67 兆美元,較 2025 年成長 49.5%,其中 AI 基礎建設支出佔比高達 56%,約 1.5 兆美元。

The global artificial intelligence industry is entering a new phase of intense competition and capital intensification. According to the latest report from Gartner, total global AI spending is projected to climb to $2.67 trillion by 2026, a 49.5% increase from 2025, with AI infrastructure spending accounting for as much as 56%, or approximately $1.5 trillion.

在此背景下,科技巨頭紛紛採取不同策略,輝達(Nvidia)計畫於今年底前開放 Nemotron 4 模型免費下載,旨在鞏固其硬體生態系,而阿里巴巴集團則於 9 月 22 日在杭州雲棲大會上發表「真武 V900」AI 晶片,展現中國追求算力自主的企圖。

Against this backdrop, tech giants are adopting varying strategies. Nvidia plans to make the Nemotron 4 model available for free download by the end of this year, aiming to consolidate its hardware ecosystem. Meanwhile, Alibaba Group unveiled its "Zhenwu V900" AI chip at the Apsara Conference in Hangzhou on September 22, demonstrating China's ambition to pursue computing power autonomy.

輝達釋出 Nemotron 4 的決策,被視為深化與現有硬體客戶連結的關鍵手段。與此同時,阿拉伯聯合大公國(UAE)等國的科技策略出現顯著轉向,為了確保能持續取得輝達的先進技術,該國國有 AI 公司 G42 已於兩年前出售與中國科技公司的股份,顯示美中科技競爭已迫使部分亞太與中東國家進行戰略選邊。

Nvidia's decision to release Nemotron 4 is seen as a key move to deepen ties with existing hardware customers. Simultaneously, the technology strategies of countries such as the United Arab Emirates (UAE) have shifted significantly; to ensure continued access to Nvidia's advanced technology, the state-owned AI company G42 sold its stakes in Chinese tech firms two years ago, indicating that US-China tech competition has forced some Asia-Pacific and Middle Eastern nations to choose sides strategically.

另一方面,阿里巴巴宣佈將投入 530 億美元推進算力基建,目標在 2032 年前將阿里雲全球資料中心容量擴充至 20 百萬瓩(GW),試圖建立輝達晶片以外的自主替代方案。

On the other hand, Alibaba announced it will invest $53 billion to advance computing infrastructure, with a goal to expand Alibaba Cloud's global data center capacity to 20 gigawatts (GW) by 2032, attempting to establish an autonomous alternative to Nvidia chips.

目前各方資訊顯示,全球 AI 發展路徑存在明顯分歧。輝達透過軟硬體整合鞏固市場霸權,而中國企業則透過大規模資本支出與自研晶片進行抗衡。值得關注的是,技術應用的廣度亦在擴張,國科會轄下國家實驗研究院於 9 月 21 日至 22 日在臺北舉辦「2026 年亞洲突變鼠資源聯盟暨亞洲實驗鼠表型分析聯盟年會」,聚焦 AI 驅動的模式動物分析與基因編輯技術,顯示亞太地區在生醫領域的 AI 應用正深化跨國學研合作。

Current information indicates a clear divergence in global AI development paths. Nvidia is consolidating market hegemony through hardware and software integration, while Chinese companies are countering with large-scale capital expenditure and self-developed chips. Notably, the breadth of technological application is also expanding. The National Applied Research Laboratories (NARLabs) under the National Science and Technology Council hosted the "2026 Asian Mouse Mutagenesis and Phenotyping Consortium Annual Meeting" in Taipei from September 21 to 22, focusing on AI-driven model animal analysis and gene editing technology, demonstrating that AI applications in the biomedical field in the Asia-Pacific region are deepening cross-border academic and research cooperation.

然而,阿里巴巴「真武 V900」的實際運算效能是否足以撼動輝達在高階市場的地位,目前尚無第三方評測資料證實,仍待觀察其後續在雲端服務中的應用表現。

However, whether the actual computing performance of Alibaba's "Zhenwu V900" is sufficient to shake Nvidia's position in the high-end market remains unverified by third-party evaluation data, and its subsequent performance in cloud service applications remains to be seen.

此波 AI 投資浪潮對亞太地區的影響深遠。對於依賴輝達硬體生態的企業而言,Nemotron 4 的免費釋出將降低開發門檻;然而,對於追求供應鏈自主的經濟體,阿里巴巴的策略反映了地緣政治對技術架構的強制重塑。高德納的資料上修顯示,基礎建設仍是未來兩年的核心,這將持續推動亞太地區在資料中心與算力資源上的大規模資本投入,但也同時加劇了技術規格分流的風險。

This wave of AI investment has profound implications for the Asia-Pacific region. For companies reliant on Nvidia's hardware ecosystem, the free release of Nemotron 4 will lower development barriers; however, for economies pursuing supply chain autonomy, Alibaba's strategy reflects the forced reshaping of technological architecture by geopolitics. Gartner's upward revision of data shows that infrastructure remains the core for the next two years, which will continue to drive large-scale capital investment in data centers and computing resources in the Asia-Pacific region, while simultaneously exacerbating the risk of technical specification fragmentation.

儘管市場對 AI 支出前景樂觀,但分析師提醒,過度解讀單一廠商的技術突破或市場預測可能忽略了宏觀經濟波動的風險。AI 支出的爆發性成長雖然反映了市場需求,但基礎建設的建設速度、能源供應限制以及國際貿易出口管制政策,均是影響未來發展的關鍵變數。特別是在亞太地區,學研合作與產業應用雖有進展,但如何平衡技術自主與全球供應鏈的互通性,仍是各國政府與企業面臨的長期挑戰。

Although the market is optimistic about the outlook for AI spending, analysts warn that over-interpreting a single vendor's technological breakthrough or market forecast may overlook the risks of macroeconomic volatility. While the explosive growth in AI spending reflects market demand, the speed of infrastructure construction, energy supply constraints, and international trade export control policies are all key variables affecting future development. Particularly in the Asia-Pacific region, while there has been progress in academic cooperation and industrial application, balancing technological autonomy with global supply chain interoperability remains a long-term challenge for governments and enterprises.

後續觀察重點在於:第一,輝達 Nemotron 4 開放下載後,對開源 AI 社群與雲端服務市場的實際影響力;第二,阿里巴巴「真武 V900」在阿里雲實際部署後的算力表現,以及是否能吸引中國以外的市場採用;第三,高德納預測的 2.67 兆美元支出中,各區域市場的具體分配比例,以及亞太各國在生醫、製造等垂直領域的 AI 落地成效。

Key points for follow-up observation include: first, the actual impact of Nvidia's Nemotron 4 on the open-source AI community and the cloud service market after its release; second, the computing performance of Alibaba's "Zhenwu V900" after its actual deployment on Alibaba Cloud, and whether it can attract adoption in markets outside of China; third, the specific distribution ratio of the $2.67 trillion in spending forecast by Gartner across regional markets, and the effectiveness of AI implementation in vertical fields such as biomedicine and manufacturing across Asia-Pacific countries.