公司指南

圖AI人形機器人

Figure AI develops humanoid robots and the Helix family of vision-language-action models.

聆聽此指南AI 語音敘述 · ~2 min

本頁閱讀時間約2分鐘
  1. 概述
  2. 深入探討
  3. 現實世界的實施
  4. 圖AI人形機器人的未來
  5. 付諸實踐
  6. 不斷探索
  7. 常見問題

概述

Figure describes these systems as aiming for general-purpose physical tasks; capabilities depend on the specific robot generation, model, and tested environment.

深入探討

Figure AI was founded by Brett Adcock in 2022. In 2024, Figure 02 was tested at BMW Group Plant Spartanburg, where BMW and Figure described trials placing sheet-metal parts into fixtures. Figure introduced its Helix vision-language-action model in February 2025, describing a System 2 for higher-level reasoning and a fast System 1 for visual-motor control. The original Helix announcement focused on upper-body control, including hands and fingers. In January 2026, Figure introduced Helix 02, which it describes as extending learned control to the full humanoid body. These are company- and partner-reported deployments and demos, not evidence of universal autonomy.

現實世界的實施

Understanding the difference between Figure’s 2024 BMW pilot and later commercial demonstrations

Studying how Helix connects visual perception, language instructions, and robot action

Comparing the original upper-body Helix system with Helix 02’s full-body control research

Evaluating a factory task such as sheet-metal placement under controlled robot trials

圖AI人形機器人的未來

Figure continues to develop humanoid hardware and Helix models for household and industrial tasks. New demos should be evaluated by task duration, degree of autonomy, human intervention, and the robot generation used. Commercial deployment requires separate reliability and safety validation.

付諸實踐

為什麼這很重要

  • 供應商策略. 供應商路線圖會影響您的團隊接下來可以建立的功能。
  • 成本與預算. 商業條款和部署選項會影響長期成本和風險。
  • 風險與安全. 公司激勵措施塑造了產品預設、安全態勢和開放性。

留意

  • 發佈公告可能會超過實際生產工作流程的穩定性。
  • API 定價或政策轉變可能會在一夜之間打破假設。
  • 單一供應商依賴性增加了鎖定和遷移成本。

後續步驟

  1. 使用您自己的任務和資料集評估提供者。
  2. 在整合之前查看隱私、安全和法律條款。
  3. 維護跨模型或供應商的後備計劃。
  4. 監控發行說明,以便路線圖的變更不會讓團隊感到意外。

不斷探索

免費電子報

獲取每日 AI 簡報

每個工作日上午提供三則經過驗證的 AI 新聞,以淺顯易懂的英文撰寫。永久免費,無廣告。

每個工作日一封電子郵件。一鍵取消訂閱。我們絕不出售或分享您的地址。

測試一下自己

Take the Figure AI Humanoid Robots quiz

每個答案皆有即時回饋,通過課程後可獲得具備可驗證 ID 的可分享證書。

開始測驗

支持免費的 AI 教育。AI Understanding 是一個 501(c)(3) 非營利組織——絕無廣告,絕無付費牆。 捐款

常見問題

什麼是Figure AI人形機器人?

Figure AI develops humanoid robots and the Helix family of vision-language-action models. Figure describes these systems as aiming for general-purpose physical tasks; capabilities depend on the specific robot generation, model, and tested environment.

What is the name of Figure AI’s in-house vision-language-action model introduced in 2025?

Figure announced Helix in February 2025 as its vision-language-action model for humanoid control.

At which company’s factory did Figure 02 conduct manufacturing trials in 2024?

BMW reported trials of Figure 02 at its Spartanburg plant, including placing sheet-metal parts into fixtures.

圖人工智慧是誰創立的?

Brett Adcock 於 2022 年創立了Figure AI;他之前曾創辦過其他企業,包括Archer Aviation。

Figure 的人形機器人的「通用」目標主要意味著什麼?

通用目的是一個單一的硬體平台,它透過軟體學習許多任務,而不是為每個工作建立專門的機器。

哪種設計理念最能描述 Helix 如何控制機器人?

Helix 使用雙系統方法:較慢的高階推理來規劃,加上快速網路即時產生連續的馬達指令。