By Jianli Yang

Copyright thediplomat

China’s surveillance and propaganda industries are often depicted as a seamless extension of state power, directed by the Chinese Communist Party (CCP) from the top down. Yet the latest leaks from two firms, Geedge Networks and GoLaxy, reveal something more complex: a commercial ecosystem in which private companies compete for contracts, leverage academic ties, and build sophisticated products to satisfy both ideological demands and market pressures. Their stories show how repression in China is both a political imperative and a profitable business, one that increasingly crosses borders. Together, Geedge and GoLaxy illustrate how China is not only perfecting digital authoritarianism at home but also packaging it for export – posing deep challenges for democracies in the global race for AI supremacy. Geedge Networks became visible when more than 100,000 internal files were leaked to a consortium of journalists, technologists, and human rights groups. These files included technical specifications, source code, client contracts, and marketing materials that laid bare how Geedge sold itself as a cybersecurity company while actually building censorship and surveillance infrastructure. GoLaxy, meanwhile, was exposed when researchers at Vanderbilt University uncovered nearly 400 pages of internal planning documents – pitch decks, brochures, sales targets, and even complaints from disgruntled employees. Together these leaks have provided an unprecedented window into the mechanics of China’s censorship and propaganda apparatus. The products they sell differ, though they complement each other within the larger control system. Geedge is an infrastructure builder. Its flagship product, the Tiangou Secure Gateway, is essentially a turnkey firewall in a box. It performs deep packet inspection, blocks VPNs and other circumvention tools, fingerprints devices, analyzes metadata, and even offers prototype “reputation”-based access controls. To make these tools accessible to non-technical officials, Geedge layers them with user-friendly dashboards that can show which users are connected, where they are located, and which applications they are running. In effect, Geedge provides the hardware and software to throttle, surveil, and deny information flows across entire networks. GoLaxy operates further up the stack, in the realm of perception and influence. Its systems ingest open-source social media data, map relationships among political actors and influencers, and use artificial intelligence to generate content for orchestrated campaigns. Dashboards allow operators to monitor discourse around sensitive topics – Taiwan, Hong Kong, Xinjiang, U.S. politics – and plan interventions by seeding narratives or amplifying favored voices. Where Geedge builds the pipes for information control, GoLaxy provides the tools to flood those pipes with content aligned to government priorities. These product differences shape their business models and client bases. Geedge’s model revolves around infrastructure contracts with governments. It sells bundled systems – hardware, software, training, and maintenance – designed to be slotted into telecom backbones or provincial networks. Domestically, its clients include provincial public-security bureaus, state-owned telecoms, and regional governments, with pilot deployments documented in Xinjiang and other provinces. Geedge has also reached abroad, marketing its “Great Firewall in a box” to regimes that want Chinese-style control. Reports show deployments in Kazakhstan, Pakistan, Ethiopia, and Myanmar. In Pakistan, Amnesty International traced the country’s “Web Monitoring System 2.0” to Geedge technology, noting how it was integrated with Western components to monitor all incoming and outgoing traffic. In Myanmar, Justice for Myanmar has documented how Geedge’s systems enable the junta to censor the internet, facilitating arrests and torture of dissidents. These cases illustrate that Geedge is not merely a domestic contractor – it is a global exporter of authoritarian infrastructure. GoLaxy’s client roster is narrower but politically significant. The Vanderbilt documents show that its main customers are CCP propaganda departments, state security services, the military, and other government bureaus tasked with “public opinion guidance.” Its business is explicitly designed to serve the party’s narrative goals. Corporate records reporting also suggests ownership ties to state-linked supercomputing and research institutions, ensuring that GoLaxy operates under both commercial and political mandates. Unlike Geedge, GoLaxy does not yet appear to export its tools widely to foreign regimes; instead, it focuses on strengthening the CCP’s domestic propaganda system and building the capacity for global influence campaigns. Its exports are narratives rather than hardware, pushing disinformation or pro-Beijing storylines into international discourse through coordinated social media activity. These client patterns reveal how both firms embody the party-state’s dual strategy: to perfect control at home and to promote its model abroad. Geedge’s exports to regimes like Pakistan and Myanmar show how China’s censorship technology is becoming a commercial product on the international market, offering a ready-made solution for governments eager to monitor and suppress dissent. GoLaxy, meanwhile, represents the CCP’s ambition to shape narratives globally, not by selling software directly to foreign governments but by running its own influence campaigns and embedding its narratives into the international information space. In both cases, the clients they serve – provincial security bureaus, authoritarian allies, CCP propaganda organs – show how intertwined political ideology and commercial opportunity have become. Both companies also maintain close ties to the Chinese Academy of Sciences (CAS), China’s premier research institution. CAS-affiliated labs have been documented visiting Geedge’s facilities, while GoLaxy’s staff often hold dual roles at CAS. These connections provide credibility, talent pipelines, and technical resources. They also blur the line between academic research and commercial repression, ensuring that cutting-edge AI techniques developed in state labs flow quickly into products that serve censorship and propaganda. The leaks underline how ideology and business mutually reinforce one another in China’s digital-control industries. Ideology generates demand: the CCP’s obsession with “cyber-sovereignty,” “public opinion guidance,” and the suppression of “hostile forces” creates a constant market for new tools. Business competition then supplies the means: companies like Geedge and GoLaxy package censorship and propaganda into modular, exportable products, backed by sales targets and marketing pitches. Provincial governments shop for turnkey censorship systems; party departments shop for influence dashboards. Competition among firms drives iterative improvement, from more user-friendly interfaces to more sophisticated AI models. The result is an ecosystem in which political goals and commercial incentives align to produce rapid innovation in repression. This has broad implications for understanding China’s AI ecosystem. First, it is state-guided but market-driven. Political priorities define what problems need solving, but companies compete to deliver solutions, pushing technological development forward. Second, it is deeply integrated: academia, private enterprise, and the state share personnel and resources, collapsing boundaries between research and deployment. Third, it is adaptive: when Western components are cut off by sanctions, firms like Geedge find ways to repurpose old hardware or substitute domestic alternatives, ensuring continuity. Finally, it is entrepreneurial: both firms demonstrate the ability to productize surveillance and propaganda, turning repressive capacity into revenue streams. For the United States, these cases highlight the multidimensional challenge of competing with China in AI. It is not just about large language models or semiconductor supply chains. It is about the infrastructure and narratives that shape the global information environment. Geedge’s export of censorship systems shows how China is spreading its model of digital authoritarianism abroad, building dependencies and normalizing practices of mass surveillance. GoLaxy’s focus on narrative operations demonstrates how AI can be industrialized to manage perception, interfere with discourse, and undermine democratic resilience. Democracies face asymmetric constraints: where China can fuse academia, business, and state to deploy tools rapidly, the United States must operate within frameworks of privacy, transparency, and accountability. Yet those very constraints are what preserve the values that China’s system seeks to erode. To meet the challenge, the U.S. must broaden its conception of the AI race. Competing on compute and models is necessary but insufficient. Washington and its allies must also invest in secure, rights-respecting alternatives to Chinese network gear, support civil society tools to resist censorship, and build coalitions to counter disinformation. They must demonstrate that openness and innovation, not surveillance and propaganda, offer the more sustainable path for societies. The Geedge and GoLaxy leaks make one thing clear: China has turned repression into an export industry, and unless democracies respond with both technological innovation and principled leadership, the global balance in the digital age may tilt decisively toward authoritarianism.