A court notice, then a wave
In late August a bankruptcy notice from the Shenzhen intermediate court pushed Yinghe Robotics into public view. The star robotics firm, holding investments from Meituan and ByteDance and having raised more than RMB 600 million, formally entered insolvency proceedings. Within a month, Zhicheng Dynamics was filed for restructuring by creditors, and DeepBlue’s liquidation reached the staff-debt public-notice stage. Three once-celebrated robotics firms fell into a survival crisis in three months. This is not an isolated operating accident. It tears open the real picture of a robotics lane pulled between ice and fire.
The rise, and the hollow core
Yinghe Robotics stood behind the Yingfeng Group, with powerful industrial-capital backing. It took a several-hundred-million-renminbi Series A just ten months after founding and at one point reached a valuation above USD 500 million. Flush with cash, the team spread a vast blueprint, signing a RMB 10 billion industrial base and running five business lines across emergency response, urban governance and warehouse logistics. But beneath the external narrative it never closed a commercial loop. It showed many demonstration prototypes yet never achieved scaled sales in any direction.
In 2025 full-year revenue was only RMB 4.33 million, with a net loss near RMB 67 million. While its own cash flow was not yet stable, the firm ran high-cost operations, burning financing on high salaries and lavish office fit-outs. Much of the hardware it bought back became assets the court disposed of, never converting into customers or revenue.
The incubator model’s hidden risk
Yinghe’s fall reflects the risk inside the parent-incubator model. An industrial group exports brand and resources to help a startup win a high external valuation. If the project succeeds, the parent enjoys the gains. Once operations stall, outside investors bear most of the loss. The firm’s life attached more to the group’s strategic choices than to its own cash generation, and that became the fate of many similar companies.
The other two firms deserve a close look too. Zhicheng Dynamics assembled a team from Microsoft, Xiaomi and DJI, aimed at the overseas pool-cleaning-robot market, once near RMB 100 million in annual revenue with a market-accepted product, yet fell on cash-flow rupture. Orders could be won, but once the collection chain broke, no matter how good the sales figures, they could not hold the firm up. DeepBlue held many disclosed cooperation orders but stuck at delivery, with large orders failing to convert into real revenue, and chronic wage arrears ultimately dragged it into liquidation.
The same kernel beneath different surfaces
The three firms showed different surface problems but a highly consistent core: they all crossed the financing gate and held paper orders, yet failed to complete the full loop from product to payment. IT Juzi’s data shows domestic robotics funding reached RMB 124.51 billion in the first eight months of 2026, already 2.7 times full-year 2025. Capital is still pouring in while a batch of star firms exit. One side a rising tide, the other an accelerating clear-out, and that apparent contradiction is exactly the truest portrait of a shakeout.
Embodied Emergence argues that capital can lengthen a firm’s trial-and-error window but cannot complete the commercial loop for it. Investors now tell industry media the yardstick has shifted: past emphasis on team background, demo effect and story imagination has given way to hard criteria of engineering delivery, real payment and scenario repeat-purchase.
Amid the frenzy some firms confused financing success with business success, treated investment as the final goal, demos as mature products, and framework cooperation as landed orders. When the capital wind tightens and external transfusions stop, these firms quickly expose their fragile underside.
Robotics hardware is a heavy-asset, long-cycle lane. A technology breakthrough is not a commercial victory. A working prototype only proves technical feasibility. Mass manufacturing, cost control, after-sales and willingness to pay are each a huge test, and capital alone rarely clears them. The incubator model needs healthier checks and balances, and external investors must impose hard constraints on cost control and commercial progress rather than chase the halo of shareholder background. The whole industry should quit the showmanship-first habit. Policy has already signalled against blind expansion, and practitioners should return to the commercial essence and answer why a customer will pay before chasing grand industrial stories.
Bankruptcy is not the failure of the whole robotics lane. It is the market removing bubbles detached from reality. The firms that cross the cycle need not raise the most or build the biggest base, but they will complete the loop of technology, product and revenue. When the tide recedes, the market teaches a plain lesson: a robotics firm’s footing was never the capital’s applause, but the customer’s real payment.
Editor’s note: This is an adapted translation of the original OFweek Robotics report. It has been trimmed and restructured for readability for an international business audience.