Author(s): Martin Garner
Qualcomm's approach to Internet of things (IoT) markets is different from that taken by other players. It combines the connectivity and computing elements into platforms that are tailored for various segments. These platforms can appear as a full reference design, a third-party board or a dedicated chip for each product type, and they're offered together with appropriate software and development kits. This provides manufacturers with the guts of the device, leaving them able to differentiate on form factor, software and services, such as cloud services or machine learning.
Years ago Qualcomm adopted a reference design approach to entry-level and mid-range smartphones. This has developed into a comprehensive suite of platforms covering several tiers of processors, multiple operating systems and operator pre-certification. For IoT markets, Qualcomm has developed a similar, separate program that is building up across a broad range of segments. It now offers more than 25 IoT platforms, spanning devices such as smart watches, drones, IP cameras, smart speakers, automotive systems and connected light bulbs.
As more equipment manufacturers use these platforms, Qualcomm will build up more-detailed knowledge of the nuances in each segment. Then, as those markets move beyond their early experimental phase, and the ecosystems take shape more clearly, it can progressively adjust the hardware and software capabilities to fit the needs and challenges more precisely. For example, drone platforms have almost the same subsystems as smartphones, with many of the same power limitations, but they are different from the platforms needed for smart speakers. Similarly, in connected light bulbs, there's a wide array of features being tried with no combinations yet emerging as clear winners.
This approach means that Qualcomm has, and will continue to have, a slightly different position in each segment, depending largely on the nature of the products. For example, its drone platform is a highly integrated board, needing relatively little further work to create a fully-fledged drone, while the wearables platforms are chipsets that require a higher level of engineering and product design expertise to develop finished products. Its industrial offerings at this stage are more-generic boards commercialized by third-party module manufacturers, focusing mostly on providing cellular connectivity. In healthcare Qualcomm's gone much further, offering a cloud service approved by the US Food and Drug Administration; the service stores data generated by devices and provides monitoring and analytics. The company also offers different support and business models across this sector.
This variety makes good sense: platforms can be offered to fit market conditions — what is right for a consumer product with a short life cycle won't be appropriate for an industrial system that will be installed and used for at least 10 years.
Given its strength in mobile phones, Qualcomm has a wide range of technologies at its disposal. These range from Bluetooth to 4G for connectivity and include CPUs, graphics chips, digital signal processing, multimedia and security for computing elements.
However, at present Qualcomm is focused on widely accepted, standardised, wireless technologies – a philosophical position that stems from its success in mobile phone markets where scale is hugely important. Yet IoT markets currently use a more-fragmented set of connectivity options, especially in the smart home and industrial areas.
There are many connection types that Qualcomm does not support natively on its platforms today – both wireless and wired. These include Z-Wave, specialist narrowband networks such as Sigfox and LoRa and various protocols used in the 800 MHz band, as well as wired connections including Modbus and the assortment of different industrial flavours of Ethernet. While it is possible to add external interface boards, that will add time and cost, so the lack of native support will limit Qualcomm's addressable market.
The problem of scale versus fragmentation isn't unique to Qualcomm —all large IT and telecom providers face the same dilemma in addressing IoT markets. This is especially true in the embedded computing market for industrial sectors, where a high degree of customisation is commonplace and technology fragmentation is extremely high. For customers there will be commercial benefits in dealing with suppliers used to operating on a large scale, but the trade-off is that fewer technology options are supported.
One major benefit of the reference platform strategy is an opportunity to become seen as the "Switzerland" of the market — a trusted neutral player providing platforms that support all major ecosystems in a market by including relevant software and cloud interfaces. This is a stronger position than being tightly aligned to any one ecosystem and it reduces risk for Qualcomm's customers.
But a major hurdle as a neutral player working with all the various ecosystems is how to be seen as a thought leader or agenda setter. This particularly applies to some of the higher layers of the IoT stack, such as analytics and machine learning. In these areas, ecosystem operators such as Amazon, Baidu, Google and Microsoft are setting the agenda, and other chipmakers are starting to establish strong positions, including Nvidia and Intel. Obviously in areas such as connectivity or heterogeneous computing, Qualcomm is already very well-known and better positioned to push its agenda.
In addition, like all suppliers moving into the Internet of things from a successful base in adjacent markets, Qualcomm faces the challenge of how to set up its sales organisation to serve so many new segments effectively. Shipment volumes in IoT markets will be somewhat smaller than in smartphone platforms. Qualcomm's acquisitions of CSR and Atheros will be of some help here and the company has implemented a new organisation to help achieve that goal. We expect the recent acquisition of NXP Semiconductors will also have a big effect (see Instant Insight: Qualcomm Buys NXP Semiconductors).
Overall, Qualcomm has a strong range of assets it can bring to the IoT area. It is aiming at a broad set of markets and is capable of addressing most of the layers in an IoT system. Its strategy in most market areas is to start with technologies from mobile devices and adapt them to a more vertical position as that market develops. This means it has numerous different positions across the various IoT markets it serves. By packaging its technologies as a platform or reference design for each area, Qualcomm is agnostic to ecosystems that are forming and is enabling more players to take part. The strategy is sensible and pragmatic, and it's worked for Qualcomm before.