NVIDIA GTC 2026 聚焦人工智慧與數位孿生技術的最新發展
NVIDIA GTC 2026 highlights latest in AI and digital twin technology
於 2026 年 3 月 16 日至 19 日在聖荷西舉行的 NVIDIA GTC 2026 大會，被譽為「AI 界的胡士托音樂節」，象徵著邁向[實體 AI|技術]時代的重大轉變。
NVIDIA GTC 2026, held in San Jose from March 16–19, was celebrated as the "Woodstock of AI," signaling a major shift toward the era of [Physical AI|technology].
此次大會突顯了 NVIDIA 從晶片製造商演變為提供完整[工業級堆疊|基礎建設]的供應商，涵蓋了模型、硬體與機器人技術。
The conference highlighted NVIDIA's evolution from a chipmaker into a provider of a complete [industrial-grade stack|infrastructure], covering models, hardware, and robotics.
會議重點在於轉向「代理型 AI」(agentic AI)，即自動化系統能在現實環境中進行規劃、推理與採取行動。
A key focus was the transition to "agentic AI," where autonomous systems can plan, reason, and act in real-world environments.
為此，NVIDIA 推出了 OpenClaw 與 NemoClaw，這些架構旨在幫助企業部署智慧代理。
To support this, NVIDIA introduced OpenClaw and NemoClaw, frameworks designed to help enterprises deploy intelligent agents.
此外，該活動展示了透過 Omniverse 在[數位孿生|模擬]領域的進展，使公司能夠構建高擬真的虛擬工廠。
Furthermore, the event showcased advancements in [digital twins|simulation] through Omniverse, enabling companies to build high-fidelity virtual factories.
為了解決機器人數據短缺的問題，NVIDIA 發布了「實體 AI 數據工廠」(Physical AI Data Factory)，利用 Cosmos 基礎模型生成合成訓練數據。
Addressing the data shortage for robotics, NVIDIA unveiled the Physical AI Data Factory, which uses the Cosmos foundation models to generate synthetic training data.
隨著 Vera Rubin 平台的首度亮相，NVIDIA 加強了對大規模推理的承諾。
With the debut of the Vera Rubin platform, NVIDIA reinforced its commitment to massive-scale inference.
正如執行長黃仁勳所指出，AI 不再僅是數位新奇事物，它已成為關鍵的[工業基礎建設|系統]層。
As CEO Jensen Huang noted, AI is no longer just a digital novelty; it has become a critical [industrial infrastructure|system] layer.
從與比亞迪和現代汽車等汽車製造商的合作，到互動式機器人的演示，NVIDIA 正為智慧代理在全球無縫運作的未來奠定基礎。
From partnerships with automakers like BYD and Hyundai to demonstrating interactive robots, NVIDIA is setting the stage for a future where intelligent agents operate seamlessly across the globe.
