Your browser is not supported. Please update it.

19 January 2021

Qualcomm CEO hands his successor a valuable gift with Nuvia purchase

Qualcomm has announced that its president, Cristiano Amon, will take over as CEO on June 30 from Steve Mollenkopf, who is retiring after a seven-year stint. Unlike the change of CEO at Intel (see separate item), this one was predictable and long-expected, though in other ways, the two companies have similar strategies for leadership transition. Like Intel, Qualcomm changes CEO seldom (Amon will be its fourth in a 35-year history) and always chooses its own veterans, and like his predecessor, Amon has prepared for his new role with stints at the helm of the largest business unit, QCT, and as president of the firm.

Just because the leadership changes are undramatic, though, does not mean that Mollenkopf’s tenure has been smooth sailing – nor that Amon’s will be either. Since the end of the period when Qualcomm was run by the Jacobs family (co-founder Irwin and then his son Paul), the company has faced a string of huge challenges, some potentially existential.

These have centered on legal and antitrust actions against its licensing models and the huge power it exerts through its patents; while its core mobile processor business has suffered slowing growth, price wars and the rise of new challengers, mainly from Greater China. And of course, Qualcomm went through a failed takeover bid by Broadcom, as well as failing in its own bid to acquire NXP.

“Steve navigated through unprecedented circumstances during his tenure, facing more in his seven years as CEO than most leaders face in their entire careers,” said Mark McLaughlin, chair of Qualcomm’s board of directors, in a statement.

As at Intel, there have been various strategies to diversify the business, but none has delivered the kind of market dominance Qualcomm has enjoyed in mobile devices. Some have remained minor markets, such as home entertainment and PC processors; some have, so far at least, been damp squibs, like the launch of a server processor range.

Others, however, are showing far greater promise, notably the expansion in connected automotive markets and in AI/ML. But, again like Intel, the sector Amon will most need to conquer is infrastructure. Qualcomm recently re-entered the macrocell equipment segment under Amon’s personal supervision, and is likely to try again at cloud processors, at least to support the move to vRAN, even if it hesitates to launch sometime more generic at this stage.

Meanwhile, Amon has also placed his personal stamp on an important area of success, the expansion of Qualcomm’s RF business, which enables it to colonize a greater share of the smartphone’s bill of materials, while also extending its reach into the growing range of connected devices and equipment that 5G will enable. Qualcomm has always had fierce rivals in the processor market, but in cellular modems it has been, generation after generation, scarcely touchable. If Amon can ensure that it retains that pre-eminence – which stems from engineering capability as well as market and licensing power – and replicates the success in RF too, it will be hard for the processor supremos to secure the level of influence they would desire in any connected wireless segment.

Among takes the helm at a time when Qualcomm is looking buoyant – its share value has more than doubled over the past 10 months as demand for 5G smartphones has risen and the chip provider has successfully hung onto its dominance of the high end of this market. A lull in antitrust actions, and the settling of legal disputes with Apple and Huawei over the past year or so, have provided some breathing space for Qualcomm to focus on new business and potential acquisitions.

Amon said: “In addition to driving the expansion of 5G into mainstream devices and beyond mobile, Qualcomm is set to play a key role in the digital transformation of numerous industries as our technologies become essential to connecting everything to the cloud.”

One of Mollenkopf’s departing actions is the acquisition of a two-year-old start-up called Nuvia for $1.4bn. True to Qualcomm tradition, it is prepared to pay apparently high prices for companies that come with valuable intellectual property, even if their revenues are low (a famous example was Flarion, which Qualcomm bought in 2006 for $600m, forming a key enabler of its diversification of its intellectual property base beyond CDMA and into OFDM technologies).

Nuvia’s immediate contribution will be to strengthen Qualcomm’s core product ranges for mobile devices, cars and laptops, but considering the start-up was originally focused on cloud servers, it is possible that there is also a longer game in play, which will see Qualcomm reviving its efforts to develop low power server chips, as companions for its new range of macrocell wireless infrastructure products. That would give it a 5G portfolio to mount a challenge to major players in the 5G cloud networks space, such as Marvell and Intel.

Nuvia’s main product is a CPU which claims to be able to “significantly outperform the competition both in raw performance and power efficiency”. Already, Qualcomm clearly sees broad application for Nuvia’s technology. It said it would integrate Nuvia CPUs into smartphones and other devices, but also digital cockpits, advanced driver assistance systems (ADAS), extended reality platforms and mobile network infrastructure solutions.

It said this will enable it to build on its own Snapdragon SoC platform, to deliver performance and power efficiency gains in order to address 5G era demands. It will add Nuvia CPUs to Qualcomm’s own graphics processing units, AI engine, digital signal processors and dedicated multimedia accelerator to expand the capabilities of the Snapdragon platform.

That raises many questions about why Qualcomm needs to buy in technology to do this, rather than relying on its own engineering prowess and resources to evolve Snapdragon’s various elements. The most obvious may be that Nuvia has its own architecture, so may provide Qualcomm with an alternative route forward, should Nvidia succeed in acquiring ARM.

This could explain the high price paid for such a young chip start-up – a hedge against Nvidia buying ARM, and Qualcomm failing to make an acceptable relationship with ARM’s new owner, or finding it too difficult to cede control of its architecture to a competitor. In that event, the San Diego firm might embark on developing its own core architecture. That would be a massive undertaking, but less than for some of ARM’s licensees, since Qualcomm is an architectural licensor, which means it develops significant unique core designs on top of the ARM foundations.

Nuvia does use ARM cores too, but has devised a new architecture of its own which could accelerate Qualcomm’s own efforts to become self-sufficient. Combining Qualcomm’s and Nuvia’s custom architectures could ease the transition to an alternative instruction set to ARM’s in future.

The start-up’s SVP of engineering, John Bruno, wrote last year that Nuvia was “creating a server CPU that is built in a new way, with a complete overhaul of the CPU pipeline. Our first generation CPU, codenamed Phoenix, will be a custom core and central to our Orion SoC.”

Qualcomm was staying vague, with Amon saying in a statement: “Together with Nuvia, we are very well positioned to redefine computing and enable our ecosystem of partners to drive innovation and deliver a new class of products and experiences for the 5G era.”

Nuvia was founded in 2019 by a group of former chip executives from Apple and Google, and co-founder Gerard Williams was chief CPU architect at Apple for over a decade.

The company’s SVP of silicon engineering, Manu Gulati, was lead SoC architect for Google’s consumer hardware unit, and before that, was also at Apple. The third member of the founding leadership team was John Bruno, SVP of system engineering, who was also a system architect at Google, and held previous roles at Apple and AMD.

When Nuvia first emerged from the shadows, it had designed its own CPU server core and associated system-on-chip (SoC) to support high performance cloud applications such as vRAN and AI. Its pitch was that it would be able to achieve the low power consumption of ARM, but with no performance trade-offs, taking “some of the lessons that were applied in the mobile SoC world around those thermally-constrained envelopes, but then add a layer of performance to that. That hasn’t been seen before in a data center perspective …to be able to achieve that step function of performance improvement that those hyperscalers need because of all the data they’re processing and transacting, within that constrained power envelope, because we’re not going to be able to get a lot more power into those data centers.”

Since then, however, the focus seems to have shifted away from the data center, perhaps because of the cost and effort to get a chip commercially ready for such a demanding and competitive environment – something that Qualcomm’s deep pockets could change.

“With Nuvia’s efforts in developing server processors, Qualcomm could combine this with its existing IP and engineering group to further advance into communications infrastructure and data center applications in the future,” wrote Jim McGregor, an analyst with semiconductor specialists Tirias Research, in a client note.

Nuvia has raised $293m in venture capital in its short life and delivered an extremely quick exit for its investors, which include Dell, WRVI Capital, and the founders of Marvell Technology Group. Navin Chaddha, managing director of one of the investors, Mayfield Fund, said: “Our under-two-year journey has been fast and furious and sets a record in our recent history at Mayfield Fund for the shortest time between serving as founding investor to creating significant financial value with a unicorn exit.”

McGregor wrote in an interview with Forbes magazine: “Nuvia’s first products aimed at server applications are not expected until at least 2022. While many ARM-based server start-ups have come and gone, Nuvia’s team is highly respected and its products highly anticipated. As one of its key investors, Dell EMC was expected to leverage Nuvia’s products. Qualcomm’s ability to take Nuvia’s cores into smartphone SoCs may be a direct challenge to Apple.”

In Qualcomm’s statement about the deal, 22 of its partners offered support including Microsoft, Google, Samsung, Acer, Asus, Bosch, Continental, General Motors, HMD, Honor, HP, Lenovo, LG Electronics, LG Mobile, OnePlus, OPPO, Panasonic, Renault, Sharp, Sony, Vivo and Xiaomi.