
xAI
Founded Year
2023Stage
Series C | AliveTotal Raised
$12.137BValuation
$0000Last Raised
$6B | 3 mos agoMosaic Score The Mosaic Score is an algorithm that measures the overall financial health and market potential of private companies.
+81 points in the past 30 days
About xAI
xAI focuses on artificial intelligence, specifically in the domain of language learning models. The company's main product, Grok, is designed to answer questions and suggest potential inquiries, functioning as a research assistant that helps users find information online. xAI primarily caters to the AI research community and the general public seeking AI tools for information retrieval and understanding. It was founded in 2023 and is based in Burlingame, California.
Loading...
Loading...
Research containing xAI
Get data-driven expert analysis from the CB Insights Intelligence Unit.
CB Insights Intelligence Analysts have mentioned xAI in 14 CB Insights research briefs, most recently on Jan 30, 2025.

Jan 30, 2025 report
State of AI Report: 6 trends shaping the landscape in 2025
Jan 7, 2025 report
State of Venture 2024 Report
Oct 29, 2024 report
State of AI Q3’24 Report
Oct 4, 2024
The 3 generative AI markets most ripe for exitsExpert Collections containing xAI
Expert Collections are analyst-curated lists that highlight the companies you need to know in the most important technology spaces.
xAI is included in 3 Expert Collections, including Generative AI.
Generative AI
1,298 items
Companies working on generative AI applications and infrastructure.
Unicorns- Billion Dollar Startups
1,261 items
Artificial Intelligence
7,222 items
xAI Patents
xAI has filed 1 patent.

Application Date | Grant Date | Title | Related Topics | Status |
---|---|---|---|---|
2/13/2024 | Application |
Application Date | 2/13/2024 |
---|---|
Grant Date | |
Title | |
Related Topics | |
Status | Application |
Latest xAI News
Feb 20, 2025
馬斯克與 Grok-3:引爆 AI 市場的超強新勢力? 億萬富翁馬斯克旗下的 xAI 日前正式發表全新 AI 模型 Grok-3,聲稱在數學、科學、程式等多項標準測試中超越競爭對手,同時展現了前所未有的運算威力與高速學習潛能。在實況發表會後,引爆了科技圈與產業觀察家的一陣吹捧或貶低:既有人讚賞這款新模型的強悍推演能力,也有人質疑其實際運用效果,認為 xAI 此次展示雖然陣容華麗,但實際功能可能仍有與先行者「看齊」的成分。馬斯克才成立不到兩年的 xAI 公司,究竟目前是否能與市場前輩「看齊」? Grok-3 的「特殊之處」似曾相識嗎? xAI 直播內容,Grok-3 訓練環境整合 20 萬顆 Nvidia GPU,這個被暱稱為「Colossus」的超級電腦叢集,122 天內完成了對 10萬顆 GPU 的同步訓練,接著又花了 92 天才全面升級到 20 萬顆 GPU。xAI 聲明,甚至考慮運算資源擴展 5 倍,打造更強大的資料處理能力。馬斯克自豪地表示,建立如此龐大且高效的基礎設施,挑戰程度不亞於開發 AI 模型本身,並預告下一階段仍將持續強化硬體層級,希望在未來取得更強勁的成長動能。 另一方面,xAI 也同步推出 Deep Search 功能,標榜「下一代搜尋引擎」。開發團隊展示如何讓 Grok-3 利用 Deep Search 搜尋廣泛網路內容,迅速生成彙整報告,這種更深層的研究代理人模式被 xAI 稱做可在「幾分鐘內完成過去需要數小時的研究工作」。xAI 官方資料,Deep Search 可問答過程反覆搜尋並精煉,能大幅縮短專業人員查找資料的時間。 此功能雖被部分使用者譽為突破,但也有人指出,ChatGPT 與 Gemini 其實早已有類似功能上線,甚至中國 AI 公司 DeepSeek 也推出過基於網路搜尋的研究代理模型,因此 xAI 的這次發表在許多實際操作細節上依舊有「後來者追趕先行者」的意味。 值得注意的是,馬斯克在 2024 年於杜拜舉辦的 The World Governments Summit 上便曾預告,Grok-3 將是「嚇人地聰明」。他指出 Grok-3 透過大量「合成資料」進行學習,還能反覆省思並修正錯誤,以提高邏輯一致性。xAI 也聲稱在內部測試中,Grok-3 不論在數學還是程式題目上,都達到比 OpenAI 及 DeepSeek 更高的測試水準。以 Chatbot Arena 上進行的盲測為例,早期版 Grok-3(代號「Chocolate」)平均 ELO 評分勝過其他參與對手,代表無論哪個用戶比較,都更常選擇 Grok-3 為回答較優秀的模型。 先前在 OpenAI 任職的計算機科學家 Andrej Karpathy 在 X 平台寫道,他認為 Grok-3 thinking 的推理運算實力表現與 o1-Pro 相當、同時領先於 DeepSeek R1 (但他表示這是他個人看法,還沒詳細測試。 且其 Reasoning Beta 版本利用內部的「chain-of-thought」推理流程並搭配額外運算資源,在 AIME 2025 這類數學測試可達 93% 的答題正確率,超越市面上大多數對手。也有開發者在測試後認為,Grok 3 mini Reasoning Beta 在部分情況下甚至表現更佳,這是因為小型模型經過更長時間的訓練,顯示 Grok-3 未來仍有進一步成長空間。 I was given early access to Grok 3 earlier today, making me I think one of the first few who could run a quick vibe check. Thinking ✅ First, Grok 3 clearly has an around state of the art thinking model (“Think" button) and did great out of the box on my Settler’s of Catan… pic.twitter.com/qIrUAN1IfD 訂閱價格跟特斯拉定價一樣混亂 馬斯克旗下 AI 公司 xAI 發表最新 AI 模型 Grok-3 幾小時後,X 便大幅調漲 Premium+ 方案價格,讓用戶取得最新 AI 模型的使用權,月費近 50 美元。X 官方支援頁面顯示的訂閱方案價格,美國地區的 Premium+ 月費調整為 50 美元,年度訂閱費升至 350 美元。 這已是 X 第二次調漲 Premium+ 價格。去年 12 月月費從 16 美元升至 22 美元。照最新價格計算,意味 Premium+ 訂閱者支付費用漲價兩倍。xAI 將 Grok-3 部分功能放入獨立 SuperGrok 訂閱,需經由 Grok 應用程式單獨購買。因此,Premium+ 用戶只能使用部分功能,如「深度搜尋」與「推理」能力。 有人抱怨價格波動過大,也有人提到 X 在近期短短幾個月內已兩度調漲 Premium+ 費用,引述自多家國際媒體的追蹤報導,這種反覆漲價策略可能會壓縮中小用戶對 Grok-3 的興趣,甚至影響平台黏著度。但按照馬斯克在經營特斯拉時的做法──價格調整是即時且浮動的,而這種做法也導致特斯拉的二手價格有時極度混亂。 馬斯克的個人風格會影響 xAI 發展? Grok-3 在商業應用領域同樣備受期待。馬斯克特別提到未來將為這套模型增加語音功能,不只單純轉換文字,而是打造「能夠理解上下文並給出自然表達」的 AI 語音助理。這項語音功能預計在短期間內釋出,並於後續開放企業級 API 連線,讓開發者得以將 Grok-3 的核心推理與語音能力整合到自己的應用程式或服務裡。 不過,對於 xAI 與 OpenAI 的紛爭仍是業界一大話題。馬斯克與 OpenAI 首席執行長 Sam Altman 早先在 2015 年共同創立 OpenAI,但近年雙方在管理哲學與企業策略上漸行漸遠。日前馬斯克與投資人團隊向 OpenAI 的非營利母體提出 974 億美元的收購提案 遭到拒絕 ,更讓雙方關係撲朔迷離。據稱馬斯克對 OpenAI 轉為營利性機構與其技術方向存有質疑,甚至提起法律訴訟指控 OpenAI 偏離初衷。OpenAI 則發表聲明強調若接受外部收購,將與最初「為人類福祉服務」的目標衝突。如今 xAI 強勢推出 Grok-3,也被視為馬斯克親自向 OpenAI 宣示競爭。 ▲ 直播時華人面孔也引發中國媒體討論。(Source:XAI) Grok-3 的推出,無疑是 xAI 在今年度最受矚目的事件之一。馬斯克在正式發表時曾強調「這只是個 beta」,表示模型目前難免存在不完善之處,但會幾乎每天都不斷更新修補,希望使用者能對其未來成長保持期待。之後,xAI 也計劃逐步開發「AI 遊戲工作室」等新領域應用,讓 Grok-3 不僅文字生成與問題解決,而能運用在更廣泛、更具創意的產業需求。 Grok-3 的超大型叢集與多項測試表現已經展現出世界頂尖的運算火力,還能夠將語音、搜尋等應用融合。雖然然現在看來部分能力與競爭對手有所重疊,但還需要實際應用才知道 Grok-3 的實際水準大概到哪個等級。但 Grok-3 為 xAI 的旗艦 AI 模型,已經為市場投入一顆足以攪亂整池水的石子。從 DeepSeek R1 出現後,AI 的競爭進入更為白熱化的階段,OpenAI 似乎也準備好推出 GPT-4.5 或其他更新版本,DeepSeek、Google、Anthropic 等對手也可能在未來幾個月內發表更強大的 AI 模型。Grok-3 能否在下一波競賽中憑藉更強的 GPU 叢集及推理功能脫穎而出,關鍵還在於實際落地應用的體驗與持續改進的速度。畢竟,用戶所關心的不只是華麗漂亮的算力數據,而是 AI 能帶給使用者與產業多大程度的實用價值──價格與實用性都是用戶評估的重點。
xAI Frequently Asked Questions (FAQ)
When was xAI founded?
xAI was founded in 2023.
Where is xAI's headquarters?
xAI's headquarters is located at 216 Park Road, Burlingame.
What is xAI's latest funding round?
xAI's latest funding round is Series C.
How much did xAI raise?
xAI raised a total of $12.137B.
Who are the investors of xAI?
Investors of xAI include Valor Equity Partners, Kingdom Holding Company, Andreessen Horowitz, VY Capital, Sequoia Capital and 20 more.
Who are xAI's competitors?
Competitors of xAI include Anthropic and 8 more.
Loading...
Compare xAI to Competitors

OpenAI is an artificial intelligence research and deployment company that focuses on ensuring that AI benefits all of humanity. The company's main offerings include developing AI technologies with a commitment to safety, alignment with human values, and broad societal benefits. OpenAI's products and services are designed to address global challenges and promote the equitable distribution of AI advantages. It was founded in 2015 and is based in San Francisco, California.

Adept is a machine learning research and product lab that focuses on developing general intelligence within the technology sector. The company specializes in creating artificial intelligence models that understand and execute tasks on a computer based on natural language commands. Adept's technology is designed to facilitate collaboration between humans and computers, enhancing productivity and enabling informed decisions. It was founded in 2022 and is based in San Francisco, California.

DeepSeek focuses on the development of artificial general intelligence (AGI) within the technology sector. The company offers a platform that facilitates easy API calls for integrating AI capabilities and a chat interface for exploring AGI applications. DeepSeek's products are designed to serve as efficient AI assistants for various user needs. It was founded in 2023 and is based in Hangzhou, China.

Inflection is an artificial intelligence (AI) studio. The studio offers a personal AI named Pi, designed to be supportive and empathetic, providing users with a new class of digital experiences. Inflection primarily serves individuals seeking personal AI interactions. It was founded in 2022 and is based in Palo Alto, California.

AI21 Labs operates as an artificial intelligence (AI) lab and product company. The company offers a range of AI-powered tools, including a writing companion tool to assist users in rephrasing their writing and an AI reader that summarizes long documents. It also provides language models for developers to create AI-powered applications. It was founded in 2017 and is based in Tel Aviv, Israel.

Pat focuses on enhancing conversational artificial intelligence (AI) through advanced natural language understanding in various business sectors. Its platform provides human-like interactions by understanding context, synonyms, and conversation history, aiming for high accuracy in conversation regardless of language or domain. It was formerly known as Thinking Solutions. It was founded in 2015 and is based in Palo Alto, California.
Loading...