Key Points:
- Chinese security giant 360 Security Technology launched a suite of artificial intelligence tools named “Yitian Tulong” to establish a domestic alternative to Western cybersecurity software.
- The suite includes “Tulongfeng,” an AI tool designed to automatically scan and identify software bugs, specifically rivaling Anthropic’s powerful Mythos system.
- Along with defensive systems, the firm introduced “Yitianzhen,” a sibling tool focused on automating network protection and responding immediately to digital threats.
- The release comes as geopolitical tensions rise, with the United States imposing severe restrictions on advanced AI exports to prevent foreign entities from utilizing systems that could impact global computer networks.
The global arena for artificial intelligence has shifted rapidly from writing creative essays to managing national security infrastructure. During a major cybersecurity conference in Beijing, the prominent Chinese technology firm 360 Security Technology introduced a new suite of artificial intelligence tools designed to counter advanced Western models. Named “Yitian Tulong” after a classic martial arts novel, the package represents a direct attempt to match the sophisticated cyber capabilities developed by American firms. The launch highlights a growing consensus among international technology experts that defensive and offensive computer security will soon run almost entirely on automated, intelligent systems.
The Chinese package consists of two distinct software engines designed to handle opposite sides of the cybersecurity spectrum. The first tool, called “Tulongfeng,” focuses on the automated discovery of software vulnerabilities. During its initial trials, this tool successfully scanned complex computer programs and identified more than 3,400 software vulnerabilities. National agencies have already officially verified 105 of those security flaws. The second tool, “Yitianzhen,” automates system defense and immediate incident response. Together, these systems manage the constant cycle of identifying software weaknesses and patching them before malicious actors can exploit the openings.
This dual-tool release directly targets the capabilities of Anthropic’s Mythos, a highly advanced artificial intelligence model that made its public preview debut earlier this year. Developed by the San Francisco-based AI developer Anthropic, which recently reached a private market valuation of $965 billion, Mythos surprised the industry with its ability to find and exploit complex software bugs autonomously. The U.S. startup chose to restrict access to Mythos, granting entry to only a few dozen select organizations to prevent the technology from falling into the hands of hostile states or criminal groups. The exceptional power of Mythos even prompted the United States government to step in, placing strict export limits on the technology due to fears over global digital safety.
Chinese technology leaders view the lack of domestic access to these advanced Western systems as a significant strategic vulnerability. The founder of 360 Security, Zhou Hongyi, argued during his presentation that China could not afford to exist in a state of “one-way transparency” where foreign governments possess automated hacking and defense systems while domestic enterprises remain empty-handed. He described these advanced AI tools as critical national strategic assets. According to Zhou, such powerful technology changes the very nature of cyber warfare, and building a local alternative is necessary to safeguard the country’s public utilities, banks, and critical infrastructure networks.
Developing a comparable alternative has forced Chinese engineers to innovate around significant hardware bottlenecks. Severe trade restrictions and export controls prevent Chinese firms from acquiring the latest high-performance semiconductor chips necessary to train massive, monolithic models. To overcome this limitation, 360 Security utilized a specialized multi-agent architectural framework. Instead of relying on one massive, resource-heavy model, this approach links together several smaller, highly specialized AI “agents” that collaborate to solve complex programming tasks. Tech analysts report that this collaborative agent design allows the company to achieve capabilities comparable to Western systems like Mythos without requiring the same level of raw computing power.
The financial stakes surrounding this technological race remain incredibly high. Listed on the Shanghai Stock Exchange under the ticker 601360, 360 Security Technology holds a market capitalization of approximately 67.06 billion CNY, which converts to roughly $9.2 billion USD. The company operates a massive security network across Asia and reported a trailing 12-month revenue of $1.19 billion USD. As investors recognize the potential of AI-driven compliance and defensive tools, the company’s stock has experienced notable volatility, trading within a wide 52-week range of 8.88 CNY to 14.68 CNY. Financial analysts point out that the rising demand for domestic AI solutions could significantly alter the revenue models of enterprise security firms over the next fiscal year.
The development of these tools is part of a broader Chinese movement to build independent artificial intelligence systems. For instance, the domestic AI research firm Zhipu recently rolled out its GLM 5.2 language model, which independent evaluators say approaches the performance of other leading Western models. Observers at European cybersecurity research centers, including the Center for Security Studies at ETH Zurich, have noted that while some claims made by tech firms are naturally elevated for marketing purposes, the underlying progress is undeniable. Western industry experts previously estimated that Chinese developers would lag six to twelve months behind their American counterparts in replicating these capabilities, a window that appears to be closing much faster than expected.
Ultimately, the transition to autonomous cybersecurity tools introduces new challenges for the global tech industry. When automated systems can scan, discover, and weaponize software vulnerabilities at machine speed, traditional human-led defense methods become obsolete. If an automated tool can find thousands of zero-day exploits in widely used operating systems within hours, security teams will have to rely on opposing AI engines to write and deploy patches before attacks happen. This reality turns the geopolitical technology race into a contest of speed, where the stability of global digital commerce depends on which automated system acts first.
