返回新聞
產業AI Understanding 簡報

DealStreetAsia 報導 Sharpa 籌集約 6.69 億美元並開設自主機器人餐廳

根據 DealStreetAsia 報導,禾賽科技聯合創始人創立的 Sharpa 籌集了超過 45 億元人民幣,並在上海開設了一家餐廳,由人形機器人製作冰雪皇后冰淇淋。

6 min readRead the linked source
Source-provided image accompanying DealStreetAsia reports Sharpa raises about $669 million and opens autonomous robot restaurant
來源參考來源記錄
出版商
dealstreetasia.com
來源連結
dealstreetasia.comhttps://www.dealstreetasia.com/stories/ai-robotics-startup-sharpa-funding-493670
來源類型
連結來源-主要來源狀態尚未確定。
背景60 秒內了解這一點

從這裡開始

關鍵術語

機器學習(ML)
允許系統從數據中學習模式並隨著時間的推移進行改進的方法。
自治系統
一種可以在有限或沒有直接人類控制的情況下即時做出決策和採取行動的系統。
測試一下自己AI 代理測驗

發生了什麼事

DealStreetAsia reports that Sharpa raised more than 4.5 billion yuan, or about $669 million, from investors including Alibaba, Meituan, Tencent, JD.com, Transsion, HongShan, Qiming Venture Partners, Luminous Ventures and Long-Z Investments. The startup also opened a restaurant in Shanghai where it says humanoid robots autonomously prepare Dairy Queen Blizzard ice cream using standard commercial equipment.

DealStreetAsia reports that Sharpa, an artificial-intelligence robotics startup founded in 2024, secured more than 4.5 billion yuan, equivalent to about $669 million. The company was founded by David Li, Xiang Shaoqing and Sun Kai, the co-founders of Hesai Technology. According to DealStreetAsia, the investor group includes Alibaba Group, Meituan, Tencent Holdings, JD.com and Transsion Holdings, along with HongShan, Qiming Venture Partners, Luminous Ventures and Meituan’s corporate venture arm, Long-Z Investments. The report says the round valued Sharpa at 22 billion yuan, or about $3.3 billion, after the investment, but does not provide a breakdown of the financing or its terms.

DealStreetAsia reports that Sharpa announced the financing one day before the opening of a robot-operated Dairy Queen restaurant on Wujiang Road in Shanghai’s Jing’an District. The restaurant began operating on August 29, 2026, and is scheduled to run daily from 10 a.m. to 10 p.m., according to the report. Sharpa said the site is a “zero-modification” operation: its humanoid robots use Dairy Queen’s standard commercial equipment and tools rather than specially altered machinery. The supplied report does not independently confirm the restaurant’s opening, its hours or the absence of human intervention.

At the restaurant, DealStreetAsia reports, Sharpa’s robots carry out a 55-step process for making Blizzard ice cream. The sequence reportedly covers order receipt, preparation, handoff and the customary upside-down presentation of the product. Sharpa described the site as the first instance in which humanoid robots have operated fully autonomously in a real-world commercial food-and-beverage environment. That “first” claim is attributed to Sharpa through DealStreetAsia and is not independently established in the supplied material. Sharpa said the financing will support research and development, hiring and the move from technical validation to practical applications.

DealStreetAsia reports that Sharpa is headquartered in Singapore, operates a business center in California and has a manufacturing research-and-development center in Shanghai. The company says its current products serve original-equipment manufacturers, research institutions and food-and-beverage clients, while its longer-term ambition includes household robots. The report provides background on the founders’ earlier work at Hesai, which raised private capital before listing on Nasdaq in 2023 and Hong Kong in 2025. Those corporate-history details are reported by DealStreetAsia and are not independently verified here.

來源詳情: dealstreetasia.com ↗

為什麼這很重要

The reported financing is unusually large for a young robotics company, while the restaurant offers a concrete commercial setting for testing humanoid robots outside controlled demonstrations. Sharpa’s claims remain dependent on the company’s own disclosures as presented by DealStreetAsia; the supplied source contains no independent technical or financial verification.

The significance of the announcement lies in the combination of capital and deployment. DealStreetAsia’s report describes a large funding round alongside a robot operating in an ordinary commercial venue, with equipment that was not modified for the experiment. If the company’s description is accurate, the restaurant is intended to test whether a general-purpose humanoid can repeat a multi-step physical workflow under everyday operating conditions rather than complete a single task in a laboratory or staged demonstration.

That distinction matters because food service involves several linked requirements: handling tools, following a sequence, responding to orders, preparing a product consistently and transferring it safely to a customer. Sharpa’s reported 55-step process provides a more concrete test of commercial reliability than a narrow demonstration. However, the source does not report throughput, quality measurements, failed attempts, downtime, staffing levels, customer complaints, injury records or the extent to which workers monitor or intervene in the process. Without those details, the deployment’s practical performance cannot be assessed.

The financing also signals investor interest in embodied AI, a field that connects machine learning with robots acting in the physical world. DealStreetAsia identifies participation by several major technology and consumer platforms, but does not explain their strategic roles or whether they will become customers, partners or simply financial investors. The reported valuation therefore indicates substantial market confidence, not proof that the technology is commercially viable. The supplied source also does not establish Sharpa’s revenue, cash position, ownership structure or prior deployment record.

For the public, the immediate implications are limited because the report concerns one restaurant in Shanghai. Still, the project could become relevant if Sharpa expands into additional food-service locations or other settings where physical automation affects jobs, workplace safety and customer expectations. Any broader conclusion should wait for evidence about how the robots perform over time, how workers are involved and what safeguards apply when the system encounters an unfamiliar object, order or operating condition.

Interactive Mechanism

互動機制:它實際上是如何運作的

以互動方式探索這項發展背後的基礎技術。

Agent Lifecycle Stage:
1
User Intent & Planning: "Audit customer refund request #4092 and settle payment."
2
Tool Calling: Emits structured JSON call crm_get_transaction(id='4092').
3
Guardrail & Verification:🛡️ Paused: High-value action requires human operator sign-off.
4
Final Settlement: Refund recorded, email receipt dispatched, and audit log stored.
Core takeaway: An AI agent is not just a language model—it is a closed loop of planning, tool invocation, and environment feedback. Production systems require self-healing retries and strict human approval guardrails.
互動式概念檢查+10 Points
AI Agents Quiz

An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?

接下來看什麼

Key unknowns include the financing structure, investor contributions, operating results, error rates, human supervision, safety procedures and the restaurant’s longer-term performance. Further evidence will be needed to determine whether Sharpa’s deployment can scale beyond one site and whether its robots can perform reliably across other businesses and environments.

The first priority is independent verification of the financing. DealStreetAsia attributes the amount, investor list and valuation to Sharpa’s disclosure in a WeChat post, but the supplied source contains no financing documents, investor statements or regulatory filings. Reporting that clarifies whether the round consists of equity, debt or another structure—and whether all named investors participated directly—would make the financial claims easier to evaluate.

The restaurant itself warrants sustained observation rather than a one-day launch report. Useful follow-up evidence would include operating data over weeks or months: the number of orders completed, average preparation time, failed or repeated tasks, maintenance needs, periods when staff take over and the ratio of robot work to human supervision. It would also be important to establish whether the restaurant remains open on the reported schedule and whether the system handles more than the specific Blizzard-making workflow described by DealStreetAsia.

Safety and labor practices are another major unknown. The source does not say how the robots are isolated from customers, what happens when equipment malfunctions, who has authority to stop them or whether workers remain nearby throughout operation. It also does not describe training, insurance, incident reporting or changes to staffing. Those details will determine whether the project represents useful automation, a tightly supervised pilot or a more with unreported limitations.

Finally, watch for evidence that Sharpa can generalize beyond one branded food-service setting. The company says it wants to develop general-purpose robots and eventually household robots, but DealStreetAsia provides no independent testing across tasks, locations or equipment. New deployments, third-party evaluations, customer contracts and transparent performance measurements would help distinguish a commercially repeatable platform from a highly customized flagship demonstration. Until then, the most defensible conclusion is that Sharpa has reported a major funding round and a notable restaurant deployment, while important financial, technical and operational facts remain unknown.

相關指引和測驗

人工智慧代理人工智慧模型解釋AI 的未來AI 倫理測試你所知道的—嘗試免費的人工智慧測驗在我們的詞彙表中尋找人工智慧術語關注 AI 資金追蹤器
覺得有用嗎?