In the spring of 2026, China’s AI companies were suddenly pulled into an unexpected “blitzkrieg” by an open-source project called OpenClaw.
In just two weeks, big tech companies such as ByteDance, Tencent, Zhipu, and Alibaba seemed to have received the same challenge letter, and one after another unveiled “weapons” named ArkClaw, WorkBuddy, AutoClaw, and Copaw, respectively. What looked like a tightly packed wave of product launches was, in fact, an intense battle for position centered on the same open-source foundation—and a fierce scramble for the right to define the ecosystem.
OpenClaw has already been widely regarded in the industry as the “Android of AI Agents.” When every player is feverishly building its own skyscraper atop the same open-source blueprint, the real nature of the contest has long since moved beyond technology itself——it is a collective rehearsal, on the eve of an explosion in next-generation human–computer interaction entry points, over traffic, data, and ecosystem sovereignty.
Lightning Rally
For commercial archrivals to make major strategic bets at nearly the same time, in the same direction, requires an unusually strong consensus.
OpenClaw was that consensus.
The modular “digital employee” foundation it provides (message routing, a reasoning engine, a plugin system) showed the giants a shortcut to rapidly building AI-native capabilities.
Behind the speed lies deep anxiety——anxiety about missing the chance, as in the early days of the mobile internet, to define the ecosystem. As a result, a “forking race” based on open source yet aimed at constructing closed, proprietary advantages quickly intensified.
Zhipu moved first. With AutoClaw, it puts “data never leaves the local environment” at the center—hitting the most sensitive security nerve in the enterprise market head-on. Tencent’s WorkBuddy, meanwhile, took a deeper path: it doesn’t just package OpenClaw; it also uses its in-house multi-agent architecture to reshape the interaction logic, embedding it deeply into the body of WeCom and QQ to build an ecosystem moat. ByteDance’s ArkClaw, backed by Volcano Engine, emphasizes cloud-native design and out-of-the-box usability, aiming to become a “new artery” connecting Lark and Douyin.
At the same time, sharp vertical players are also opening up flank battlefields: Moonshot AI is using KimiClaw to dig deep into complex, long task workflows; NetEase Youdao’s Lobster AI focuses on educational documents; and some players are even rewriting in Rust in pursuit of extreme performance.
The opening act peaked immediately.
The first chapter of this race already revealed a brutal truth: on top of the “common language” provided by open-source frameworks, the giants are striving to write their own distinctive, exclusive “dialects.” Open source levels the playing field for technology, but ecosystems build defenses through “forking.”
Why Reinvent the Wheel Again?
A fundamental question then emerges: if the code is open source, why are the giants spending heavily to reinvent the wheel? Behind this seemingly contradictory behavior are three deep games that intertwine with one another.
First is the battle for the entry point.
In the giants’ strategic blueprints, AI Agents are the smarter, more proactive “next-generation traffic hub” after the super app. They may become the core interface through which users access every service. For Tencent, this is the lifeline for consolidating its social and workplace territory; for ByteDance, it is a new bridge linking content with collaboration scenarios.
Fighting for dominant influence over the OpenClaw ecosystem is, in essence, fighting for the “meta-power” to define how users will interact in the future.
Second is the battle for security, which is especially evident in the enterprise market.
Here, security and compliance are a harder currency than efficiency. OpenClaw’s open-source nature—and its support for on-premises deployment—makes it an ideal foundation for meeting “data sovereignty” demands. Each company’s so-called “in-house developed” version is, in essence, built atop that foundation: stacking on the highest levels of security certification and the most granular access governance, turning open-source technical advantages into a “letter of trust” that large government and enterprise clients simply can’t refuse.
This is an arms race over “who is more reliable.”
And at the very heart of it lies a battle over the ecosystem.
Open-source frameworks are the “commons” of infrastructure, but commercial success depends on cultivating a thriving “private garden.” The giants’ strategies are strikingly consistent: use OpenClaw to connect everything, but “lock in” value through their own ecosystems. Tencent ties it to WeCom; ByteDance links it through Feishu; Alibaba anchors it in DingTalk. The open-source portion wins developer mindshare and technical legitimacy, while the closed, deep ecosystem integrations form the real commercial moat.
This is a refined strategy of “open-source as the base, closed-loop profit on top.”
Capital markets’ reaction underscored its strategic value. After related products were released, Zhipu and Tencent saw their share prices rise significantly. This wasn’t just optimism about a single product—it was a bet on the long-term narrative that, by capturing the Agent entry point, they could consolidate or expand their broader ecosystem empires.
Major Trends
Today’s free-for-all will eventually settle into a new industry order. Several major trends are already coming into view.
Competition is rapidly escalating from “feature delivery” to “ecosystem definition”. In the future, the decisive factor won’t be who builds the best individual Agent, but who can provide developers and users with fertile ground capable of incubating massive numbers of AI-native applications. The giants’ various moves around OpenClaw are, at their core, an arms race in extending ecosystem capabilities.
Market trajectories will also diverge more rationally: “cloud-based agility” and “on-premise fortresses” will coexist for the long haul, serving different customer segments. ByteDance’s cloud-native route covers efficiency-driven SMEs and mainstream users; meanwhile, versions that emphasize private, on-premise deployment will dominate large government and enterprise markets with high data sensitivity. The two are not substitutes; they are parallel paths that expand the market frontier together.
As general capabilities converge through open source, deep vertical-industry know-how becomes the ultimate moat. “Agent-ification” in education, law, healthcare, finance, and other fields requires a profound grasp of industry logic and the ability to deliver closed-loop services. This will be a vast battleground where challengers can stage comebacks and specialist players can take the crown.
If we widen the lens, the OpenClaw phenomenon marks a pivotal shift in mindset for China’s AI industry: a collective trial of moving from “open-source consumers” to “open-source definers”.
In the past, most innovation happened within Western-led open-source ecosystems such as Linux and Android. This time, for the first time, at the very dawn of a new field with paradigm-shift potential (AI Agents), a core open-source project initiated locally emerged and drew global attention. The giants’ collective “forking” and heavy betting is a fight for the “superstructure” built atop a homegrown open-source foundation.
This is no longer a matter of passively following rules set by others—it is the beginning of actively participating and trying to write the rules together.
What it tests is how China’s tech giants can strike a balanced, respected business model between the universal spirit of “giving back to the global open-source community” and the practical need to “build a sustainable commercial closed loop.” This is not only a technology race, but also a coming-of-age trial of business ethics and strategic wisdom in the open-source era.
The beacon fire lit by OpenClaw will eventually die down. But what it gives rise to will be an AI Agent ecosystem galaxy far more complex than today’s.
There will be no single, winner-takes-all “Android” here, but it is very likely that several thriving planetary systems will form around different gravitational centers (social, content, office, and industry). The spirit of open source ensures the fluidity and sharing of the technological base, while the giants’ battles will be decided on the ecosystem moats each of them builds.
The most profound legacy of this race may not be any monopolistic platform, but rather an important “fact check”:
For the first time, on the eve of the breakout of an entirely new paradigm, China’s AI industry tried—through collective exploration—to practice, in the deep waters of foundational software, how to take control of its own course. This experiment in “open-source sovereignty,” regardless of how each player ultimately ends up, has already been profoundly meaningful in and of itself.
Explore more exclusive insights at nextfin.ai.
