Qualcomm Acquisition of Edge Impulse Adds Huge Value to Dragonwing Announcement

On February 25, Qualcomm announced a new product line, Dragonwing. This new portfolio works alongside Snapdragon and is tailored for industries and consumer use cases (including industrial & embedded Internet of Things (IoT), networking, and cellular infrastructure). The announcement was quickly followed on March 10 by the acquisition of Edge Impulse, a platform that simplifies the process of building, deploying, and scaling Machine Learning (ML) models on edge devices. This acquisition aims to 1) augment Qualcomm AI Hub by integrating Machine Learning Operations (MLOps) capabilities (including model optimization); 2) reduce the barrier between Edge Impulse’s 170,00 active developers and Qualcomm’s new Dragonwing portfolio; and 3) add software monetization alongside Qualcomm’s hardware-only commercial model. ABI Research expects that this acquisition is likely to have a positive impact on Qualcomm’s new product and market entry strategy:

• Build Differentiation Through Software Integration & Innovation: Qualcomm has relied on AI Hub and third-party partnerships to support the developer ecosystem. However, given the market competition with as many as six Tier One IoT vendors actively competing in Original Equipment Manufacturer (OEM) product cycles, building a truly differentiated solution requires leading software capabilities with significant applications/developer support. By tightly integrating Edge Impulse’s platform into its software stack, Qualcomm can offer developers a unique experience by improving hardware-level optimization & lowering deployment barriers. This will accelerate Time to Value (TTV) and alleviate some of the bottlenecks slowing Artificial Intelligence (AI) adoption within the IoT.
• Target Specific Verticals: Horizontal solutions can struggle to scale, because OEMs want highly-optimized solutions that target specific applications/pain points. Edge Impulse’s strategy has, so far, targeted and built strong industry awareness in key industries (including industrial, wearables, health, and appliances). Qualcomm can utilize this vertical focus to support Dragonwing’s go-to-market strategy.
• Build Toward Solution-Orientated Product Portfolio: As a previous ABI Insight recommended, vendors must move toward a solution-orientated approach that combines hardware, software, and applications to provide “turnkey” solutions that solve specific pain points. Deeper integration of MLOps tools into the Qualcomm AI Hub can push Dragonwing further toward a turnkey solution, rather than one that relies on third-party partners.
• Co-Development & Improvement: Although Edge Impulse’s platform will remain agnostic, through tighter integration with Qualcomm AI Hub and pursuing co-development opportunities it can drastically improve hardware optimization, creating greater value for developers. Tighter integration can be achieved by integrating workflows between the AI Hub and Edge Impulse, integrating and optimizing Qualcomm’s Software Development Kit (SDK) into the Edge Impulse platform. Clearly, Qualcomm & Edge Impulse recognize this, as they have emphasized that integration improves inference performance by 4X. In addition, the acquisition can solve one of the major time-to-value challenges in the edge market. Currently, hardware vendors build chips and then expose their capabilities to software partners toward the end of the Proof-of-Concept (PoC) phase. However, bringing Edge Impulse in-house means longer-term visibility and closer Research and Development (R&D) alignment across the lifecycle. This will have a positive effect on developers, as they will be given early access to build/test models & applications on PoC hardware, accelerating time-to-value for new hardware releases.
• Data, Data, Data: Data remains one of the biggest challenges for industrial adoption of AI use cases, as most players lack the necessary data to tune foundation models and build new applications. Edge Impulse’s data solution that enables developers to aggregate datasets for training will support the buildout of new applications mapped to Dragonwing hardware.

Bringing a new hardware solution to market in the saturated and maturing IoT market is going to be challenging. However, by deeply integrating Edge Impulse within its software stack, Qualcomm may be able to drive adoption/migration through improved experience and deeper optimization. On top of this, by remaining agnostic to other hardware partners, Qualcomm will be able to effectively monetize both hardware and software, following an approach similar to that of Syntiant.

Qualcomm’s decision to acquire Edge Impulse looks very interesting on paper and certainly will support the entry of Dragonwing into an already congested marketplace. However, given Edge Impulse’s market position, the major impact of this acquisition may not be on Qualcomm, but on other key players.

Edge Impulse is a critical edge AI software partner in the ecosystem (with many regarding them as leaders). It has developed deep partnerships with key players, including STMicroelectronics, BrainChip, Renesas, Syntiant, Silicon Labs, and Alif Semiconductor. Some of these partners even rely exclusively on Edge Impulse to provide vital MLOps tools within their software ecosystem. For example, STMicroelectronics relies on Edge Impulse to add critical MLOps capabilities (data collection, deep model optimization, model training) to its ST Edge AI Suite for Microcontroller Units (MCU)/Microprocessor Units (MPUs).

Although Edge Impulse has announced that it will continue to honor these existing partnerships by remaining agnostic, there are many potential downsides for Qualcomm competitors as a result of this acquisition:

• Tight integration between Qualcomm & Edge Impulse, alongside the likely co-development, will place Qualcomm’s solution capabilities and experience above other Edge Impulse partners, leading to a migration away from its hardware solutions.
• Although Qualcomm’s vertical focus seems aligned with Edge Impulse for now, this could easily change moving forward as it commercially scales Dragonwing. Qualcomm may pivot and explore new areas that would impact Edge Impulse’s strategic focus, potentially having a negative impact on its partner’s vertical strategy.
• There is, of course, no guarantee that, in the future, Qualcomm will take Edge Impulse fully in-house, which will eliminate their partnerships with other hardware vendors and significantly inhibit their value proposition.

Although this acquisition may not spell immediate danger for Qualcomm’s IoT competitors, it should certainly be seen as a long-term risk, given their huge reliance on Edge Impulse’s MLOps capabilities. Exploring other partnerships or Mergers and Acquisitions (M&A) opportunities with players like Neuton.ai is certainly a good idea to de-risk their positions in the market.

Qualcomm’s acquisition of Edge Impulse is not a shocking revelation, as it mirrors competitor strategies over the last couple of years. Infineon acquired Imagimob in 2023 to expand its software stack and add AutoML capabilities, while this business remains a largely separate entity. Renesas acquired Reality AI in 2022, adding its software environment to support product development and lifecycle management. STMicroelectronics acquired Cartesiam in 2021, adding its specialized AI development tools for Arm-based microcontrollers.

Each of these vendors recognized the importance of building strong software capabilities to move toward “turnkey” end-to-end solutions that drive engagement and accelerate application development. But what can Qualcomm and Edge Impulse learn from these acquisitions:

• It’s a Thin Line Between “Startup” Autonomy and Simply a Slightly Deeper Partnership, so It Is Important to Effectively Balance This to Maximize the Benefits of the Acquisition: Edge Impulse’s platform will still remain agnostic and support other hardware players. Qualcomm needs to effectively manage the integration between the two companies to ensure they maximize the benefits of this acquisition in terms of product co-development and optimization, while still enabling startup autonomy to monetize software and quickly pivot to new opportunities.
• Syntiant Offers a Strong Case Study of a Company That Both “Monetizes” and Utilizes Software as a “Value-Add,” Blending These Models Is Challenging and Requires Effective Differentiation: As Edge Impulse remains agnostic, it is evident Qualcomm wants to monetize its software platform. However, blending this with “value-add” software means that Qualcomm and Edge Impulse must build effective differentiation for Dragonwing to incentivize developers to choose the MLOps platform and the hardware solution. Targeting power efficiency through improved optimization will be the most effective product strategy, while ensuring tight platform integration will improve developer experience and incentivize end-to-end usage. At the end of the day, the best-case scenario for Qualcomm is if developers choose to use Edge Impulse’s platform to develop applications targeting Dragonwing.
• Look to Use Edge Impulse’s Software Talent and Qualcomm’s Deeper Pockets to Build & Expand the Software Environment to Target Key Pain Points: One of the major challenges for AI/IoT developers remains test and evaluation. Qualcomm should provide the capital to support significant investment in this software environment to build out its product capabilities and enhance its overall value proposition. Branding this under Qualcomm would mean that it offers differentiation against Edge Impulse’s platform partners.

The edge AI market is very competitive with vendors reporting that OEM contracts are being contested by more than six mostly identical competitors. Qualcomm’s decision to enter the market alongside a headline-making acquisition was smart and will certainly help build its value proposition. However, given the maturity of other solutions in this space, Qualcomm and Edge Impulse will need to put their heads together to build an effective and highly integrated product and go-to-market strategy to ensure they can showcase their differentiation to OEM customers, while continuing to monetize Edge Impulse’s platform through hardware agnosticism.