View Financial HealthMininglamp Technology 配当と自社株買い配当金 基準チェック /06Mininglamp Technology配当金を支払った記録がありません。主要情報n/a配当利回り-2.7%バイバック利回り総株主利回り-2.7%将来の配当利回りn/a配当成長n/a次回配当支払日n/a配当落ち日n/a一株当たり配当金n/a配当性向n/a最近の配当と自社株買いの更新更新なしすべての更新を表示Recent updatesお知らせ • Apr 16Mininglamp Technology Officially Open-Sources Mano-P 1.0 ModelMininglamp Technology has officially open-sourced Mano-P 1.0, a self-developed GUI-aware agent model capable of executing complex cross-platform tasks entirely through pure vision. By seamlessly controlling desktop software, web interfaces, and complex graphical workflows, Mano-P breaks the "see but not do" limitation of AI, empowering users to build personalized AI while guaranteeing data sovereignty and ushering in the era of privatized personal AI. Unlike traditional automation constrained by API calls or HTML parsing, Mano-P utilizes pure visual understanding to operate without external interfaces. It serves as a critical execution foundation for the Agent ecosystem, integrating seamlessly into agents like OpenClaw. This breakthrough eliminates manual intervention bottlenecks, enabling autonomous execution in complex commercial scenarios such as 3D applications and professional tools that were previously inaccessible to standard agents. The model redefines performance ceilings, achieving overwhelming State-of-the-Art (SOTA) results across 13 authoritative multimodal benchmarks. Operating on a dual-version architecture, the 72B model ranked first globally in the OSWorld benchmark with a 58.2% task success rate, leading the runner-up by 13.2 percentage points, while also dominating systems like ScreenSpot-V2 and MMBench. Furthermore, the highly efficient 4B quantized model supports local execution on Apple M4 Pro devices, achieving up to 476 tokens/s prefill speeds with just a 4.3GB memory footprint. As AI integrates into core workflows, data privacy is paramount. Mano-P ensures zero data uploads to the cloud by running locally on Mac devices or via a dedicated USB 4.0 compute stick. This "pure visual understanding plus local execution" architecture guarantees physical isolation from external networks, allowing AI to securely drive complex business processes and self-correct even in offline, high-privacy enterprise environments where data sovereignty is a non-negotiable requirement. Released under the Apache 2.0 license, Mano-P provides out-of-the-box usage modes, supporting commercial use and secondary development without complex API configurations. Mininglamp Technology is initially open-sourcing the Mano-CUA core skill for easy integration into existing workflows like OpenClaw or Claude Code, with local models and SDKs following within the month. Future releases will include underlying training methods and token pruning technology, accelerating the global development of a personalized AI ecosystem.Reported Earnings • Mar 29Full year 2025 earnings: EPS misses analyst expectationsFull year 2025 results: CN¥137 loss per share (down from CN¥0.18 profit in FY 2024). Revenue: CN¥1.43b (up 3.2% from FY 2024). Net loss: CN¥6.41b (down CN¥6.42b from profit in FY 2024). Revenue was in line with analyst estimates. Earnings per share (EPS) missed analyst estimates significantly. Revenue is forecast to grow 25% p.a. on average during the next 2 years, compared to a 24% growth forecast for the Software industry in Hong Kong.お知らせ • Mar 26Mininglamp Technology, Annual General Meeting, Jun 12, 2026Mininglamp Technology, Annual General Meeting, Jun 12, 2026.お知らせ • Mar 16Mininglamp Technology to Report Fiscal Year 2025 Results on Mar 26, 2026Mininglamp Technology announced that they will report fiscal year 2025 results on Mar 26, 2026Board Change • Nov 03No independent directorsFollowing the recent departure of a director, there are no independent directors on the board. The company's board is composed of: No independent directors. 5 non-independent directors. Executive Director, Senior VP & Chief Client Officer Jie Zhao was the last director to join the board, commencing their role in 2024. The company's lack of independent directors is a risk according to the Simply Wall St Risk Model.決済の安定と成長配当データの取得安定した配当: 2718の 1 株当たり配当が過去に安定していたかどうかを判断するにはデータが不十分です。増加する配当: 2718の配当金が増加しているかどうかを判断するにはデータが不十分です。配当利回り対市場Mininglamp Technology 配当利回り対市場2718 配当利回りは市場と比べてどうか？セグメント配当利回り会社 (2718)n/a市場下位25% (HK)2.7%市場トップ25% (HK)6.9%業界平均 (Software)2.9%アナリスト予想 (2718) (最長3年)n/a注目すべき配当: 2718は最近配当金を報告していないため、配当金支払者の下位 25% に対して同社の配当利回りを評価することはできません。高配当: 2718は最近配当金を報告していないため、配当金支払者の上位 25% に対して同社の配当利回りを評価することはできません。株主への利益配当収益カバレッジ: 2718の 配当性向 を計算して配当金の支払いが利益で賄われているかどうかを判断するにはデータが不十分です。株主配当金キャッシュフローカバレッジ: 2718が配当金を報告していないため、配当金の持続可能性を計算できません。高配当企業の発掘7D1Y7D1Y7D1YHK 市場の強力な配当支払い企業。View Management企業分析と財務データの現状データ最終更新日（UTC時間）企業分析2026/05/26 19:31終値2026/05/26 00:00収益2025/12/31年間収益2025/12/31データソース企業分析に使用したデータはS&P Global Market Intelligence LLC のものです。本レポートを作成するための分析モデルでは、以下のデータを使用しています。データは正規化されているため、ソースが利用可能になるまでに時間がかかる場合があります。パッケージデータタイムフレーム米国ソース例会社財務10年損益計算書キャッシュ・フロー計算書貸借対照表SECフォーム10-KSECフォーム10-Qアナリストのコンセンサス予想+プラス3年予想財務アナリストの目標株価アナリストリサーチレポートBlue Matrix市場価格30年株価配当、分割、措置ICEマーケットデータSECフォームS-1所有権10年トップ株主インサイダー取引SECフォーム4SECフォーム13Dマネジメント10年リーダーシップ・チーム取締役会SECフォーム10-KSECフォームDEF 14A主な進展10年会社からのお知らせSECフォーム8-K* 米国証券を対象とした例であり、非米国証券については、同等の規制書式および情報源を使用。特に断りのない限り、すべての財務データは1年ごとの期間に基づいていますが、四半期ごとに更新されます。これは、TTM（Trailing Twelve Month）またはLTM（Last Twelve Month）データとして知られています。詳細はこちら。分析モデルとスノーフレーク本レポートを生成するために使用した分析モデルの詳細は当社のGithubページでご覧いただけます。また、レポートの使用方法に関するガイドやYoutubeのチュートリアルも掲載しています。シンプリー・ウォールストリート分析モデルを設計・構築した世界トップクラスのチームについてご紹介します。業界およびセクターの指標私たちの業界とセクションの指標は、Simply Wall Stによって6時間ごとに計算されます。アナリスト筋Mininglamp Technology 1 これらのアナリストのうち、弊社レポートのインプットとして使用した売上高または利益の予想を提出したのは、 。アナリストの投稿は一日中更新されます。1 アナリスト機関Liping ZhaoChina International Capital Corporation Limited
お知らせ • Apr 16Mininglamp Technology Officially Open-Sources Mano-P 1.0 ModelMininglamp Technology has officially open-sourced Mano-P 1.0, a self-developed GUI-aware agent model capable of executing complex cross-platform tasks entirely through pure vision. By seamlessly controlling desktop software, web interfaces, and complex graphical workflows, Mano-P breaks the "see but not do" limitation of AI, empowering users to build personalized AI while guaranteeing data sovereignty and ushering in the era of privatized personal AI. Unlike traditional automation constrained by API calls or HTML parsing, Mano-P utilizes pure visual understanding to operate without external interfaces. It serves as a critical execution foundation for the Agent ecosystem, integrating seamlessly into agents like OpenClaw. This breakthrough eliminates manual intervention bottlenecks, enabling autonomous execution in complex commercial scenarios such as 3D applications and professional tools that were previously inaccessible to standard agents. The model redefines performance ceilings, achieving overwhelming State-of-the-Art (SOTA) results across 13 authoritative multimodal benchmarks. Operating on a dual-version architecture, the 72B model ranked first globally in the OSWorld benchmark with a 58.2% task success rate, leading the runner-up by 13.2 percentage points, while also dominating systems like ScreenSpot-V2 and MMBench. Furthermore, the highly efficient 4B quantized model supports local execution on Apple M4 Pro devices, achieving up to 476 tokens/s prefill speeds with just a 4.3GB memory footprint. As AI integrates into core workflows, data privacy is paramount. Mano-P ensures zero data uploads to the cloud by running locally on Mac devices or via a dedicated USB 4.0 compute stick. This "pure visual understanding plus local execution" architecture guarantees physical isolation from external networks, allowing AI to securely drive complex business processes and self-correct even in offline, high-privacy enterprise environments where data sovereignty is a non-negotiable requirement. Released under the Apache 2.0 license, Mano-P provides out-of-the-box usage modes, supporting commercial use and secondary development without complex API configurations. Mininglamp Technology is initially open-sourcing the Mano-CUA core skill for easy integration into existing workflows like OpenClaw or Claude Code, with local models and SDKs following within the month. Future releases will include underlying training methods and token pruning technology, accelerating the global development of a personalized AI ecosystem.
Reported Earnings • Mar 29Full year 2025 earnings: EPS misses analyst expectationsFull year 2025 results: CN¥137 loss per share (down from CN¥0.18 profit in FY 2024). Revenue: CN¥1.43b (up 3.2% from FY 2024). Net loss: CN¥6.41b (down CN¥6.42b from profit in FY 2024). Revenue was in line with analyst estimates. Earnings per share (EPS) missed analyst estimates significantly. Revenue is forecast to grow 25% p.a. on average during the next 2 years, compared to a 24% growth forecast for the Software industry in Hong Kong.
お知らせ • Mar 26Mininglamp Technology, Annual General Meeting, Jun 12, 2026Mininglamp Technology, Annual General Meeting, Jun 12, 2026.
お知らせ • Mar 16Mininglamp Technology to Report Fiscal Year 2025 Results on Mar 26, 2026Mininglamp Technology announced that they will report fiscal year 2025 results on Mar 26, 2026
Board Change • Nov 03No independent directorsFollowing the recent departure of a director, there are no independent directors on the board. The company's board is composed of: No independent directors. 5 non-independent directors. Executive Director, Senior VP & Chief Client Officer Jie Zhao was the last director to join the board, commencing their role in 2024. The company's lack of independent directors is a risk according to the Simply Wall St Risk Model.