A clearly more comprehensive approach
From wireless communication specialist to system provider
Nordic Semiconductor is well known for offering highly energy-efficient ICs for wireless communications. But the company is now taking a much broader approach. According to Kjetil Holstad, EVP Strategy and Product Management at Nordic, the company is now able to take on many more roles in systems.
Markt & Technik: Nordic is well known for its Bluetooth LE ICs, but the company has significantly expanded its range in recent years...
Kjetil Holstad: Yes, starting from a very well-established base with Bluetooth, we have also established other connectivity protocols such as cellular IoT in recent years. In addition, by taking over the entire development team for Ensigma WiFi IP from Imagination Technologies, we have established a position for ourselves within WiFi. However, we also see newer protocols such as Matter and Thread on the rise. We are working on covering these new technologies with a clear product offering, combined with the known benefits of Nordic. To put it simply, you could say that we have evolved from a Bluetooth company to a wireless company.
However, we are now also dealing with completely different topics. Because we want to solve our customers' problems, not just in terms of connectivity, but for the entire system.
What does that mean in concrete terms?
For example, we have started to look at the problems from the battery to the antennas. That's why we have developed our own power management ICs to complement our wireless communication ICs. Because if you really want to achieve low power consumption in the system, you have to think from the battery to the antenna.
In addition, technologies such as mobile communications and WiFi naturally have a cloud component too. Accordingly, we also have a cloud-based offering in our portfolio that can be used to perform over-the-air updates and things like that. This means that we don't just supply a chip, but an entire solution.
This expanded approach has come about as a result of requests from our customer base, who have presented us with problems that require far more than just the sale of a chip. It's about selling a solution, an ecosystem around the software, a development platform, and so on.
Power management ICs are offered by many companies, how can Nordic differentiate itself here?
Our power management ICs are important for customers who really need very low power consumption, because in this case power management also plays a big role. Our radio ICs are designed for the lowest possible power consumption, which is what we are known for, but you have to look at the entire path to minimise power consumption.
Of course, our customers can also use power management ICs from other suppliers, such as TI or Analog Devices for our communication ICs, but if they use our power management ICs, the power consumption will be lower.
Does Nordic offer PMICs for all its supported communication standards and can these ICs also be used in combination with communication ICs from competing suppliers?
We have only just started to develop this product segment, which means that the product portfolio is not yet comprehensive. As I said, our aim is to supplement our existing product range of ICs for wireless communication with power management ICs so that we can solve a larger part of the customer's problem, but of course the components can also be combined with Bluetooth ICs from other manufacturers.
But the system approach that Nordic is now pursuing goes even further?
Yes, it does in terms of the need for higher computing power at the edge, i.e. in the sensors or very close to the sensors. There are two reasons for the demand for higher computing power: in one case, it's about developers wanting to do the same thing as yesterday, but much faster, because that also saves power. In the other case, power consumption is not the issue - it remains the same, but the developer wants to do much more with this power consumption. The sensor should therefore become more intelligent. Accordingly, we are working on increasing the computing power on the one hand and reducing the power consumption for these calculations on the other.
Are you talking about the pre-processing of sensor data?
Yes, that too. But I would put it another way: in general, it's about bringing AI to the edge.
There is also another point: the protocol stacks for the various communication standards require more and more computing power. So, it's about all aspects of data processing and we are addressing this trend with our latest product generation, which increases computing power tenfold compared to the previous version, and yet reduces absolute power consumption. This has clear advantages for our customers, as many now use a Nordic chip and a microcontroller alongside it. By increasing the computing power and capabilities of the Nordic ICs, they no longer need this external MCU.
Which processor cores does Nordic use, ARM or RISC-V?
We use ARM cores in most designs. However, some products from the newer generation use a RISC-V as a complementary core, i.e. a combination of ARM and RISC-V. The ARM cores function as general purpose cores, while we use RISC-V for dedicated tasks such as data acquisition or security issues.
Do you use licensed RISC-V cores or have you developed your own?
We have taken the instruction set and developed our own cores, as we only use RISC-V for special tasks.
Considerable for the size of Nordic...
Yes, but that has its justification. We are not interested in replacing ARM. We are more than happy with that, the ecosystem and the innovative strength of ARM are perfect. This means that there is absolutely nothing wrong with ARM's offering and therefore there is absolutely no reason for us to want to replace ARM.
The reason why we are nevertheless investing in RISC V is that RISC V allows us to do things that are important to us and our customers a little differently, i.e. to be more innovative. And we couldn't do that with cores from SiFive or similarly ambitious suppliers of RISC-V cores. Because these companies are trying to replace ARM with RISC-V. That's not what we're about, we want to create innovations and that's why we're convinced that it's worth investing in our own RISC-V cores.
Can you give specific examples of when Nordic uses its own RISC-V cores?
For example, in our latest nRF54H family, which includes multiple RISC-V coprocessors. Thanks to the 22nm process used, we can now do functions that we have typically implemented in hardware with a small processor core, such as the RISC-V. I don't want to go into too much detail, but when it comes to low latency, sensor-related functions or the emulation of a serial interface, i.e. good old bit banging, for example, we take the standard instruction set from RISC-V and optimise it for what we want to implement. This could also be implemented with ARM cores, but with RISC-V we can optimise the core precisely for the purpose, because in these cases an ARM core would be overkill from a power consumption perspective.
Nordic uses a 22nm process for its SoCs?
Yes, for our fifth generation of Bluetooth products, it is now available to selected customers and will be launched later this year. Compared to the 2- or 3-nm processes you read a lot about, the 22-nm structures don't seem "cutting edge" at first. But for communication ICs, this process technology is extremely advanced. We will offer these 22nm products in the form of a high-end series and a low-end series of wireless SoCs. There is another important innovation with this generation: we are sourcing these 22 nm components from two different foundries: in addition to TSMC, GlobalFoundries in Germany is also manufacturing the new components - so we want to ensure that supply is guaranteed in all cases.
Does GlobalFoundries use its 22nm FDSOI process to manufacture these SoCs?
Yes, we use the FDSOI process for the high-end products and TSMC's 22nm process for the low-end series.
Does the FDSOI process bring advantages in terms of power consumption?
Both process technologies, i.e. from TSMC and GlobalFoundries, are very good process technologies in terms of power consumption, but clearly the FDSOI process brings additional advantages.
When did Nordic start working with GlobalFoundries?
The decision to pursue a dual-source strategy was made after the experiences of the coronavirus pandemic. Of course, it will take time for this decision to be implemented, but the first products manufactured by GlobalFoundries will be launched on the market this year.
Nordic Semiconductor is also heavily involved in cellular IoT, what are Nordic's activities here?
Nordic has always been about being "Lead on Connectivity". This means reducing power consumption and increasing RF performance and always being better than competing suppliers. That's why we use very advanced process technology with 22 nm structures for Bluetooth LE, for example. Of course, Bluetooth LE cannot be directly compared with cellular IoT, because Bluetooth LE is much older. There are now products for Bluetooth 5.4, but we are already working on Bluetooth 6 and Bluetooth 7 in the background. Essentially though, the basic idea is the same for mobile IoT technologies, i.e. we are also pursuing the same approaches for cellular IoT as for Bluetooth, i.e. increasing integration and reducing power consumption. Power consumption is particularly important in IoT because we are often talking about battery-operated devices. But this is in our DNA, so we do everything we can to consume as little power as possible, and this aspect is also very essential for the collective IoT.
Nordic wants to score on the wireless side, is it more a question of process technology or design?
Every company has access to the latest process technologies, so that's not a competitive advantage, it's the design that counts.
That's true, but perhaps Nordic is simply faster when it comes to using more modern production processes?
I think what sets Nordic apart is that we develop the entire radio unit in-house. We also develop our own connectivity stacks, which is just as important for our customers. Sure, we have a lot of competitors, but if there's a problem with their ICs, it's often difficult for customers. We, on the other hand, can solve the problems quickly because we have developed everything in-house. This makes us much faster when it comes to solving problems or improving things.
There is another point that is becoming increasingly important: Five years ago, no customer was interested in this topic, but now it is the number one issue for many.
Desirable, a study by Cisco has actually described the opposite, i.e. the topic is still the last priority for most providers.
Yes, but the regulatory authorities are slowly waking up and demanding that devices that are connected to the Internet also have certain security mechanisms. We, at Nordic, have been dealing with this issue for many years, which means we are working at the forefront to ensure that we are ahead of our customers.
But security is a "moving target", does Nordic offer any options for upgrades or updates?
Yes of course, we've been doing this for more than ten years, and since then it has been possible to upgrade Nordic ICs over the air, which is an important part of our security strategy.
Another topic is "multi-protocol", a feature that is demanded by the market and therefore important for Nordic?
Yes definitely, multiple protocols are essential for us because we are convinced that some of these protocols complement each other. For example, when it comes to a smart home system with Matter, Bluetooth is typically also important for provisioning. But even if a system house uses different types of mesh networking, in many cases it has to create a bridge between old and new systems, let's say Zigbee and Matter. This means that the different communication standards complement each other in many applications, i.e. they explicitly do not replace each other.
From our point of view, Bluetooth is the common denominator that works well with pretty much all other technologies. Bluetooth enables easy access to mobile phones so that many tasks can be carried out via the phone. This also applies to mobile IoT and tracking applications. Even in mobile IoT, many users want to use WiFi, for example, to connect to existing WiFi networks and perform tracking. It is quite clear that many of our customers are implementing several protocols at the same time, but clearly not all of them. Some will continue to use only one technology. But we as a company are trying to ensure that we offer multi-protocol capabilities that are interoperable and can co-exist.
Is there a greater need for multi-protocol ICs in various areas?
Smart home is certainly an application that requires frequent use of multi-protocol ICs, because it typically uses multiple communication protocols. Whether it's a combination of Thread and WiFi or Thread, WiFi and Bluetooth, or Thread and Bluetooth, etc., it's typically a combination of communication standards. But even in non-consumer asset tracking, multiple communication technologies are the norm. This starts with fleet management and ends with tracking containers etc. - one of the reasons for using different communication standards is power consumption. Because one thing is clear: using mobile communications for everything is too expensive from a power consumption perspective, which means that if these users can use another technology, they will.
The interview was conducted by Iris Stroh
