Rockchip Micro, making a stunning debut!

（Source: Langge Finance）

Behind the “raising lobsters” boom, who’s making quiet money?

In March 2026, the open-source AI agent OpenClaw quickly broke into the mainstream. Its GitHub star count surpassed 280,000, overtaking long-standing leaders like Linux and React in one fell swoop, becoming a phenomenon-level AI project!

Because the product icon is a red Boston lobster, and users must complete intelligent agent personalization through data feeding and instruction training, the whole process looks like “raising livestock,” and this gameplay is also aptly called “raising lobsters.”

OpenClaw’s sudden popularity may help AI officially step into the Agent era!

In the wake of this wave, related industry chain companies have all been enjoying development dividends.

The AI server product from Langchao Information has already been adapted to the OpenClaw runtime environment; cloud vendors like UCloud have even taken the lead in launching dedicated deployment images, gaining a first-mover advantage in the cloud deployment market.

However, they have overlooked an important area: edge AI chips.

Edge AI chips can be understood as the “smart brain” installed in devices like computers, smartphones, and cars. Without sending data to the cloud, AI recognition, judgment, and decision-making can be completed locally.

In fact, edge AI chips have long been a high-growth track. Data shows that in 2024, the domestic edge AI chip market size reached 21 billion yuan, and it is expected to surge to 86 billion yuan by 2030, with a compound annual growth rate of over 20%.

Behind this are accelerating deployment of terminal AI applications, along with the ongoing expansion of edge-side scenarios such as AIPC and smart home appliances, all jointly driving steady growth in demand for edge computing power.

The trend toward localized deployment of OpenClaw is also adding fuel to the demand for edge AI chips!

And in this blue-ocean track, Rockchip has already moved early to secure a foothold!

Hard-core product adaptation to OpenClaw for strong technical groundwork

Rockchip is a leading SoC chip design company in China.

A SoC chip refers to integrating various functions such as CPU, GPU, and NPU onto a single chip, enabling functional unification. This is the core carrier for edge AI chips to realize on-device intelligent computing.

As a typical Fabless (no wafer fab) model company, Rockchip focuses on the core chip R&D stages, outsourcing wafer foundry and packaging/testing.

This model enables the company to run with light assets.

As of the end of Q3 2025, the company’s fixed assets were only 50.06 million yuan, accounting for just 1% of total assets. Light-asset operations bring advantages of low depreciation costs, and also allow the company’s gross margin to stay high for a long time.

From 2021 to 2024, Rockchip’s gross margin remained above 30%. In the first three quarters of 2025, its gross margin was as high as 41.77%, leading chip design companies such as Himax Technologies and Allwinner Technology.

In product R&D, Rockchip has formed a clear, tiered layout.

In 2021, the company launched the first domestic high-performance general-purpose processor RK3588; its next-process flagship version is expected to be上市 in 2026. In 2024, it introduced the mid-to-high-end AIoT processor RK3576, completing the tiered product layout.

Rockchip states that multiple chips such as RK3588 and RK182X are suitable for personal deployments of products like OpenClaw, and there are already solutions based on RK3588 that call cloud-based large models.

Meanwhile, entry-level processors like RK3506 and AI audio-dedicated processors like RK2118 have also been rolled out one after another, enabling Rockchip to build a full-range AIoTSoC chip platform covering “entry-level—mid-tier—mid-to-high-end—high-end.”

With strong product capabilities, Rockchip has already established stable cooperation with well-known terminal customers such as Anker Innovations, BBK Electronics, and Lenovo.

And behind all of this is the company’s years-long, unrelenting R&D investment.

From 2021 to the first three quarters of 2025, Rockchip’s cumulative R&D investment reached 2.65 billion yuan. Its R&D expense ratio has continuously stayed above 14%, and in 2022 it was as high as 26.36%.

As of the end of June 2025, the proportion of R&D personnel was nearly 80%. The company has applied for 1,337 patents, of which 736 have been authorized. Its R&D strength has become a solid backing for steady business growth.

With advantages in products, technology, and more, Rockchip has already firmly established itself in the edge AI chip arena. But what is the core driving force behind its years of sustained high R&D and stable technical barrier building?

Earnings surge by 100%

Strong cash flow

Rockchip’s ongoing technical investment comes from the solid support of its abundant cash flow.

According to its earnings forecast, Rockchip expects to achieve revenue of 4.387 billion yuan to 4.427 billion yuan in 2025, net profit of 1.023 billion yuan to 1.103 billion yuan, with year-over-year net profit growth of over 70%. The growth rate of non-GAAP net profit (after deducting non-recurring items) can be as high as 99.3% at most—truly stunning!

Benefiting from strong performance growth, from 2021 to the first three quarters of 2025, the company’s net operating cash flow totaled 2.575 billion yuan, injecting strong momentum into continuous R&D for SOC chips.

The rapid growth in performance comes from key breakthroughs across multiple core product lines—especially in these two areas.

One is automotive electronics.

As intelligent driving accelerates its penetration, it greatly increases demand for components like LiDAR, cameras, and millimeter-wave radar chips.

With a forward-looking perspective, Rockchip has laid out six major areas including intelligent cockpits and in-car instrument clusters. In the intelligent cockpit segment, the company’s primary offering RK3588M—thanks to its strong performance—has already been successfully adopted in models from GAC Group.

By the end of 2024, this chip had reached mass production in more than ten vehicle models, with another 20-plus designated projects.

The other is humanoid robots.

A humanoid robot’s “brain” and “small brain” determine its capabilities in intelligent interaction and motion control, and a SoC chip is precisely the key carrier for these two major cores.

Rockchip has worked deeply in the robotics field for many years. High-performance general-purpose processors such as RK3588 and RK3576 can realize core functions including human-machine interaction, motion control, and environmental perception.

At present, Rockchip has partnered with robot manufacturers such as Unitree Technology and DeepBlue Technology to launch multiple forms of robot products.

Judging from changes in contract liabilities, the company’s subsequent performance growth is also well assured.

Contract liabilities can be understood as the deposits collected when the company signs contracts with customers, which to a certain extent can measure the company’s on-hand order situation.

From 2021 to the first three quarters of 2025, Rockchip’s contract liabilities surged from 4.65 million yuan to 56.26 million yuan. The rapid growth in on-hand orders provides support for the company’s performance.

Rockchip leverages abundant cash flow to sustain R&D, and its continuously rising on-hand orders further solidify the foundation for development. Driven by multiple tracks, the company is steadily opening up growth space for its chip business.

From the “raising lobsters” craze sparked by OpenClaw, to the industry trend of AI moving toward the Agent era, the explosive demand for computing power has become an irreversible trend. As the core for edge-side deployment, edge AI computing-power chips are also ushering in a new blue-ocean of development.

With continuous deep cultivation, Rockchip occupies an early position in the edge AI computing-power field, and this will also become its core backing for continually capturing AI industry dividends.

A vast amount of information, precise interpretation—everything on the Sina Finance app