Qihoo 360 / 奇虎 360

Overview

Qihoo 360 / 奇虎 360 is the Chinese internet and security company whose evidence arc moves from free PC security, browser traffic, search, and platform conflict into enterprise security, AI search and browsers, specialized models, and agent infrastructure.

Current Profile

The historical source frames 360 through a paradox: it used anti-rogue-software trust to answer the kind of browser-plugin behavior associated with 3721, then converted free security into a large traffic platform. That PC-era advantage proved difficult to carry into mobile, content, finance, and new distribution systems, and Zhou later describes the company as having overexpanded after 3Q.

The newer source presents an application-led AI reset. Rather than train the largest general model, 360 retains smaller and vertical-model capability around reasoning, coding, intent, routing, and security, while using browsers, search, 纳米 AI, AI video, an agent factory, and a multi-agent engine as application and workflow layers. The direction is coherent with 360’s entry-point and security history, but product reach, quality, adoption, and commercial outcomes remain founder-reported.

Key Characteristics

• Consumer security company built through free antivirus and mass-market PC utility distribution.
• Browser, navigation, and search entry-point owner with a history of aggressive platform competition.
• Post-PC company that moved toward enterprise and government security after mixed mobile and content expansion.
• Application-led AI builder emphasizing multi-model access, routing, specialized models, agents, and browser execution.
• Security operator applying models and agents to defense while acknowledging AI-enabled attacks and prompt injection.
• Organization using AI video and agent construction as source-scoped tests of multi-agent coordination.

Evidence

• Consumer-security platform: No.216 ⚔️ 「十步杀一人」：红衣大炮周鸿祎的三十年全面战争史 | 中国互联网故事 27 covers Security Guard, free antivirus, browser/navigation distribution, search, listing history, and platform conflicts.
• Strategic limits and repositioning: No.216 ⚔️ 「十步杀一人」：红衣大炮周鸿祎的三十年全面战争史 | 中国互联网故事 27 describes mobile and content difficulties and the “big security” turn; 周鸿祎×罗永浩！周鸿祎深度谈 AI！近四小时高密度输出 adds service to enterprise digitization and traditional-industry AI.
• AI architecture: 周鸿祎×罗永浩！周鸿祎深度谈 AI！近四小时高密度输出 describes smaller vertical models, model routing, browser and search integration, tool use, agent factories, and multi-agent systems.
• Product cases: 周鸿祎×罗永浩！周鸿祎深度谈 AI！近四小时高密度输出 uses 纳米 AI and long-form AI video as the main application examples.
• AI security: 周鸿祎×罗永浩！周鸿祎深度谈 AI！近四小时高密度输出 connects security models and agents to faster analysis while identifying phishing, malicious tooling, hallucination, and prompt injection as new risks.

Qualifications

Market shares, revenue, legal summaries, model integrations, agent counts, video duration, product quality, staff adoption, customer benefit, and security accuracy remain source-scoped. The newer evidence is a founder account and does not independently establish competitive advantage or validate the claim that avoiding a frontier general model is the best long-run strategy.

What Changed

• Extended 360 from PC and enterprise security into an application-led AI and agent strategy.
• Added Nano AI, AI browser/search, model routing, agent factories, and AI video as source-scoped product branches.
• Added AI’s dual defensive and offensive security role.

Relationships

• 周鸿祎 / Zhou Hongyi - founder and public strategic face of the company.
• Nano AI / 纳米 AI - application-layer AI product spanning search, browser, routing, and agent construction.
• Free Security Software Strategy / 免费安全软件策略 - core consumer-security growth tactic.
• Chinese Internet Entry-Point Competition / 中文互联网入口竞争 - recurring browser, navigation, search, and AI-interface role.
• 3Q War / 3Q 大战 - defining conflict and later overexpansion reference point.
• Vertical Agent Architecture - specialized model, tool, routing, and domain-context approach.
• Agent Organizational Culture - coordination problem exposed by multi-agent product work.
• Human-in-the-Loop Agent Governance - approval and traceability boundary for agent execution.