Track Hyper | Arm joins the self-developed chip battle group

Author: Zhou Yuan / Wall Street News

On July 30 local time, British semiconductor IP giant Arm announced a forecast for the next fiscal quarter that fell short of market expectations.

Data shows that Arm expects revenue for the second fiscal quarter to be between $1.01 billion and $1.11 billion, in line with the market expectation of $1.06 billion. However, Arm also predicts adjusted earnings per share to be between $0.29 and $0.37, with the median falling below the market average expectation of $0.36.

This news disappointed investors, but what shook the market even more was Arm's plan to invest part of its profits in manufacturing its own chips and other components.

According to public information, Arm CEO René Haas revealed that Arm is investing in the development of its own chips. This significant information marks a major shift in Arm's business model, which originally licensed chip IP design blueprints to other companies.

Traditional Model Achieves Glory

In the chip industry, Arm has long played a unique and important role.

As an intellectual property supplier, Arm does not directly produce chips but licenses chip design solutions to major semiconductor manufacturers worldwide.

Arm's licensing model mainly includes processor licensing, processor optimization packages, physical IP package licensing (POP), and architecture/instruction set licensing.

Through this model, Arm collects royalties from each chip produced using its technology, with upfront licensing fees generally ranging from $1 million to $10 million, payable in a lump sum, depending on the complexity of the licensed technology.

Royalties are Arm's share of the revenue after the chips are shipped, typically 1%-2% of the chip's selling price; however, for some new architectures or overall computing solutions, the royalty rates may be higher: for example, the royalty rate for the Arm v9 architecture is 2%-4.5%, and the royalty rate for the overall computing solution CSS can reach 8%-10%.

In the context of Arm's business, CSS refers to "Compute Subsystem," which is a more integrated chip design solution provided by Arm.

Unlike purely licensing architectures (such as Arm v9), CSS includes a more complete combination of hardware and software, typically integrating key components such as CPU cores, GPUs, interconnect buses, and memory controllers, and may even include basic software stacks and development tools.

This one-stop solution can help customers (such as chip design companies) significantly shorten product development cycles and reduce design complexity. For example, in the first half of the year, a well-known domestic company adopted the CSS solution, resulting in a finished chip with outstanding performance.

Arm's light-asset business model has achieved tremendous success.

From smartphones to IoT devices, from automotive electronics to data centers, Arm architecture is ubiquitous, with 99% of smartphones worldwide using chips designed based on Arm architecture.

With extensive licensing partnerships, Arm has about 1,000 licensing partners and 320 partners, with astonishing processor shipment volumes.

Despite Arm's remarkable achievements under the traditional licensing model, in recent years, growth in its core market—the smartphone sector—has gradually weakened IDC released data on July 14, indicating that during the period from April to June 2025, global smartphone shipments will only grow by 1%, a decrease of 0.5 percentage points compared to the previous quarter.

The saturation of the smartphone market has limited Arm's royalty revenue growth in this field. At the same time, tariff fluctuations and macroeconomic challenges brought about by global trade tensions have further weakened end-market demand, casting a shadow over Arm's market prospects.

In the rapidly growing data center market, although Arm has made some inroads, clients such as Amazon's cloud computing division have begun using Arm technology in this area, it faces fierce competition.

Traditional chip giants like Intel and NVIDIA have a solid foundation in the data center market, with strong technological research and development capabilities and a wide customer base.

If Arm wants to gain a larger market share in this field, relying solely on the licensing model seems inadequate.

By developing its own chips, Arm can create more competitive products, directly participate in market competition, and enhance its voice in the data center market.

In recent years, Arm has been promoting semi-custom IP integration services, expanding revenue through the "Compute Subsystem (CSS)" platform, combining high-margin new products, and increasing royalties per chip to enhance profits.

If chip self-development is achieved, Arm can better integrate its CPU, GPU, and other IP with CSS solutions into a system platform, achieving deep collaboration and innovation in technology.

Finished chips, as the physical embodiment of Compute Subsystem (CSS) products, can more intuitively showcase Arm's technological strength, providing customers with more complete solutions to meet their demand for one-stop services.

Impact: Customer Relationships Face Restructuring

Arm plans to increase investment in developing its own chip products, a move that signifies a significant change in its relationships with many customers.

There have long been signs regarding Arm's intention to develop its own chips.

In December 2024, during the court hearing of the lawsuit between Arm and Qualcomm over technology licensing issues, Qualcomm's legal team presented documents prepared for the board by Arm CEO René Haas, indicating that Arm was considering designing its own chips to supply directly to customers.

This means Arm could become a major competitor to clients like Qualcomm, but at that time, René Haas refuted these claims, stating that Arm does not manufacture chips and has never ventured into this industry.

However, subsequent news "ruthlessly slapped" René Haas in the face: In February of this year, British media reported that Arm was developing its own chips, with the first self-developed chip expected to be launched as early as that summer, to be manufactured by TSMC, and Meta might become one of the first customers.

On July 30, Arm announced an increase in investment in self-developed chips; Rene Haas straightforwardly stated that the company plans to create a complete product line including small chips, physical chips, motherboards, and systems.

This significant shift from IP or semi-custom solution licensing to directly manufacturing chips transforms Arm's relationships with many chip design companies, including Qualcomm, from partners to direct competitors According to media reports, Arm has even recruited talent from its clients; this behavior has further intensified the tension between Arm and its original partners.

In the future, Arm needs to reassess its cooperation model with clients and find a new balance to maintain relationships with existing clients while expanding its customer base.

If Arm succeeds in the self-developed chip field, even if it's just a start, it will have a profound impact on the existing chip industry landscape.

In the data center market, Arm's entry may break the relative monopoly of Intel and NVIDIA, bringing new competitive vitality to the market and prompting other chip manufacturers to accelerate technological innovation and product iteration.

In emerging fields such as the Internet of Things, if Arm can leverage its advantages in low-power chip design to launch competitive self-developed chips, it is expected to further expand its market share in these areas and drive industry restructuring.

Arm's strategic shift from the traditional IP licensing model to self-developed chips is a response strategy and choice in the face of challenges such as market growth bottlenecks and increasing competitive pressure.

This transformation, while full of uncertainty and risk, also brings new development opportunities for Arm. However, whether Arm can succeed in the self-developed chip field, reshape its relationships with clients, and rewrite the industry landscape may not be so smooth, judging by the capital market's reaction on that day.

Of course, the final answer will depend on whether Arm can truly launch powerful and competitive chip products; as for the capital market's perspective, it is not that important