On August 3, a bombshell dropped in venture capital circles: Lingbo Tech, Ant Group's embodied AI subsidiary, has launched its first independent funding round, seeking ¥1.5 billion, with a second round targeted by year-end. Moving from Ant's in-house "physical AI task force" to market scrutiny, Lingbo's independence reflects a key inflection point in embodied AI — the burn rate has outgrown the comfort zone of internal incubation.

Why "Getting Pushed Out"

Sources put it bluntly: the core reason for independent funding is four words — resource squeeze. Inside Ant Group, the LLM business (Doubao) is in a critical phase of the arms race — compute, talent, and capital all tilt toward LLMs. Lingbo, as the embodied AI business, is strategically important but has less near-term commercial clarity, so naturally its resource priority drops.

This isn't an Ant-only problem. Embodied AI has a "double-high" characteristic: high R&D investment + long commercialization cycle. LLMs can monetize quickly through APIs and SaaS — a new model generates revenue the day after launch. But embodied AI is different — hardware needs tooling, mass production, supply chains, after-sales service. Every step is capital-intensive and time-consuming. A humanoid robot can't go from prototype to production in less than 2-3 years.

That's why you're seeing an interesting pattern: internet giants' embodied AI teams are either raising independently or scaling back. Not because they're not optimistic — it's that the burn pace is unsustainable.

What ¥1.5B Means

What level is a ¥1.5B first round? A comparison makes it clear. In early July, Suzhou Lexiang Technology closed a nearly ¥500M Pre-A round led by Ant Group — that's an external startup. Lingbo, as Ant's in-house "darling," is opening with ¥1.5B, almost three times Lexiang's round.

Two meanings behind this number. First, the barrier to entry in embodied AI is rising fast. Two years ago, tens of millions of RMB could fund an embodied AI startup. Now, without ¥1B-level reserves, you can't even qualify for the race. Hardware iteration, data accumulation, team scale — each one is a money pit.

Second, capital market expectations for embodied AI remain high. Although commercialization timelines keep getting pushed back, investor enthusiasm hasn't faded. The logic is simple: AI's next billion-user scenario will likely be in the physical world. The LLM race格局 is mostly clear — embodied AI is still blue ocean.

Independent Funding: Good or Bad?

In the short term, independent funding is good for Lingbo. With independent capital comes independent decision-making power and control over pace — no more fighting with the LLM division for compute and headcount. Moving from the group's "marginal business" to the market's "independent player" also means more flexible team incentives.

But from another angle, independent funding also means weaning off. Previously backed by Ant, brand endorsement, technical resources, and scene access came free. After independence — are those resources still free? How much can you get? Everything needs renegotiation. More importantly, capital market money isn't free — you need a clear commercialization path and timeline.

The startup logic in embodied AI is shifting. 2023-2024 was "story-driven" — a demo was enough to raise money. 2025 was "tech-driven" — you needed real substance. 2026 is shifting to "delivery-driven" — investors are starting to ask: when will you have real orders? When do you break even? Lingbo choosing this moment for independent funding is both being pushed and catching the right rhythm.

See you tomorrow.