Sanctuary AI Sanctuary AI זמינות מוגבלת

סנקטוארי AI פיניקס

שנת דגם 2023 עודכן

Phoenix הוא הרובוט ההומנואיד המסחרי של Sanctuary AI, שמתאפיין ב-Carbon — מערכת AI קניינית שמשתמשת במודלי שפה ליצירת התנהגות רובוטית, המייצגת גישה אדריכלית שונה מרוב המתחרים.

לוגיסטיקה קמעונאות
אין תמונה עדיין
7.9 ציון עריכה
/ 10

מפרט טכני

Manufacturer יצרן Sanctuary AI Sanctuary AI
Height גובה ~1.70 מ' (5'7")
Weight משקל ~70 ק"ג (154 פאונד)
Payload משקל נושא ~10 kg ~10 ק"ג
Max Speed לא נחשף
חיי סוללה לא נחשף
דרגות חופש לא נחשף במלואו
פלטפורמת AI Carbon — proprietary LLM-based task planning system
מחיר ארגוני (לא ציבורי)
חיישנים מצלמות RGB, חיישני עומק, חישת כוח

תרחישי שימוש

ניהול מלאי קמעונאי
מיון ואחסון מוצרים
איתור ושליפת פריטים
משימות לוגיסטיקה חצי-מובנות

שירות ותפעול

יוצר ב ונקובר, קולומביה הבריטית, קנדה. דוח עיתונאי 2026-05-12 מקור מקור חלש. מטה בלבד; אתר הייצור לא אומת.

באז, 7 ימים אחרונים 0 אזכורים

לא נמצאו חדשות אחרונות עבור רובוט זה.

נייטרלי Buzz rankings →

Sanctuary AI Phoenix — Latest Videos

Sanctuary AI מוסיפה חיישני מגע לרובוטים הומנואידיים מדגם Phoenix

Robotics247 24,802 צפיות 6 במרץ 2025

Sanctuary’s New PHOENIX 7 AI Robot Demo Stuns Entire Industry With This Tech (ASTRIBOT S1, VIDU)

חדשות AI 11,974 צפיות Apr 29, 2024

Sanctuary AI's Phoenix: World's First Learning AGI Humanoid Robot

AI Tech Pro 1,073 צפיות Apr 27, 2024

Sanctuary AI is a Vancouver-based company pursuing what it calls “human-like intelligence for general-purpose robots.” Phoenix, its commercial platform, is the physical expression of that ambition. But what distinguishes Sanctuary from competitors is its AI architecture, not its hardware.

Carbon is Sanctuary’s proprietary AI system, built on the premise that the path to general-purpose robot behavior runs through language model reasoning connected to motor execution. Where most competitors train neural networks to map sensor observations directly to motor commands, Carbon uses an LLM-based planning layer that interprets tasks in natural language, breaks them into sub-goals, and sequences learned motor primitives to execute them.

The approach has a compelling theoretical advantage: language models encode vast amounts of task knowledge from internet text, giving Carbon a head start on understanding what “sort these items by color” or “check the expiration date on this product” means without task-specific training data. In practice, bridging language understanding to reliable physical execution remains a hard problem.

Phoenix has been deployed in commercial pilots with several retail and logistics customers. Mark’s Work Wearehouse (a Canadian retailer) was among the first disclosed customers, using Phoenix for inventory management tasks including sorting, stocking, and item location. These are representative of the kind of semi-structured tasks Carbon is designed to handle.

Phoenix stands approximately 170 cm and weighs 70 kg. Hardware is capable but not class-leading — the differentiation is intentionally in software. Sanctuary uses a teleoperation-first approach to data collection: human operators demonstrate tasks via a control interface, generating labeled data that trains Carbon’s task models.

יתרונות מרכזיים: Carbon LLM-based AI architecture, natural language instruction, retail/logistics pilots, teleoperation data collection flywheel.

Frequently Asked Questions About Sanctuary AI Phoenix

What is Sanctuary AI Phoenix used for?

Phoenix is designed for retail and logistics tasks including inventory sorting, stocking shelves, and item location in semi-structured environments. Mark’s Work Wearhouse, a Canadian retailer, uses Phoenix for inventory management. The robot handles tasks that require interpreting natural language instructions like sorting items by category or checking product details.

Who manufactures Sanctuary AI Phoenix?

Sanctuary AI, a Vancouver-based company, manufactures Phoenix. The company focuses on developing what it calls human-like intelligence for general-purpose robots. Sanctuary AI differentiates through its Carbon AI architecture rather than hardware specifications, using a teleoperation-first approach to data collection where human operators demonstrate tasks to train the system.

How much does Sanctuary AI Phoenix cost?

Pricing is not publicly available as of 2026. Phoenix is offered through enterprise contracts for commercial pilot programs. There is no consumer or retail purchase option. Interested organizations can contact Sanctuary AI through sanctuary.ai to discuss pilot deployments and enterprise pricing for retail and logistics applications.

What are the key specifications of Sanctuary AI Phoenix?

Phoenix stands approximately 170 cm (5 feet 7 inches) tall and weighs roughly 70 kg (154 lbs). It has an estimated payload capacity of about 10 kg. The robot uses RGB cameras for perception. Detailed specifications for speed, battery life, and degrees of freedom have not been fully disclosed by Sanctuary AI publicly.

Where is Sanctuary AI Phoenix currently deployed?

Phoenix has been deployed in commercial pilots with several retail and logistics customers. Mark’s Work Wearhouse, a Canadian retailer, was among the first disclosed customers, using Phoenix for inventory management tasks. These deployments focus on semi-structured retail environments requiring sorting, stocking, and item identification capabilities.

How does Sanctuary AI Phoenix compare to other humanoid robots?

Phoenix differentiates through its Carbon AI software architecture rather than hardware specifications. While its physical capabilities are capable but not class-leading, the LLM-based Carbon system enables natural language task instruction, which most hardware-focused competitors lack. This software-first approach targets task generalization over raw physical performance metrics.

What AI capabilities does Sanctuary AI Phoenix have?

פיניקס מפעיל את Carbon, מערכת תכנון המשימות הקניינית של Sanctuary AI המבוססת על LLM. Carbon משתמשת בשכבת הסקת מסקנות של מודל שפה המחוברת לביצוע מוטורי, מפרשת הוראות בשפה טבעית, מחלקת אותן לתת-מטרות, ומרצפת פרימיטיבות מוטוריות נלמדות. נתוני האימון מגיעים מהפעלה מרחוק שבה מפעילים אנושיים מדגימים משימות, ויוצרים גלגל נתונים מסומנים לשיפור מתמשך.

חוזקות

Carbon LLM architecture for generalization
Natural language task instruction
Retail deployment pilots
Unique AI differentiator

מגבלות

Specs less impressive than hardware-focused rivals
LLM-to-motor gap remains hard
Smaller company, limited resources
Narrow current deployment scope