Nokia's 5G chip choice leaves it exposed

When Nokia was busy absorbing Alcatel-Lucent after its €15.6 billion ($17.3 billion, at today's exchange rate) takeover in early 2016, it took a 5G decision that has come back to haunt it. That decision, about the components used in the company's 5G equipment, now threatens Nokia's 5G competitiveness against chief rivals Ericsson and Huawei. As customers start to roll out their 5G networks, it could even result in a loss of market share.

The decision was to use 5G products based heavily on field-programmable gate arrays (FPGAs). To those outside the chip industry, these are essentially circuits that can be configured or reprogrammed after they have been designed. The alternative was to use what is known as an application-specific integrated circuit (ASIC), intended from the outset for a given purpose. FPGAs, Nokia decided, would bring flexibility and time-to-market advantage.

That was deemed important while Nokia was still wrestling with the Alcatel-Lucent integration. Bought mainly because of its expertise in fixed-line and core network technologies, Alcatel-Lucent also came with its portfolio of mobile products. Integrating this with Nokia's range of mobile goodies, to avoid any overlap or conflict, would have been a mammoth task given the size of the two businesses.

Nokia's Share Price in Helsinki Since October 21
(Source: Yahoo Finance)
(Source: Yahoo Finance)

"Nokia was dealing with the migration of the portfolio that nobody else in the industry had to deal with in regard to this Alcatel-Lucent phase-out," said Rajeev Suri, Nokia's CEO, during a call with analysts last week. FPGAs, with their inherent attractions, would help to overcome those constraints.

But there was an important trade-off. While FPGAs bring flexibility, they are also a lot more expensive than ASICs, bloating Nokia's bill for cost of goods sold. Nokia's 5G equipment, in other words, has turned out to be less profitable than gear that relies heavily on ASICs instead of FPGAs. While a detailed breakdown of product lines is unavailable, the gross margin at Nokia's networks business shrank five percentage points in the recent third quarter, to just 29.1%, compared with the year-earlier period. At Ericsson's networks division, the gross margin ticked up from 41.3% to 41.6% over the same period.

Nokia's problems were made worse because one of its suppliers apparently fell short of expectations. Commenting on the FPGA decision, Suri told analysts: "That is not the most efficient from a cost point of view… and, admittedly, one supplier let us down as well."

In a research note issued after the publication of third-quarter results, Michael Genovese, an analyst with MKM Partners, said the supplier in question had failed to deliver a low-cost 10-nanometer product, while Andrew Gardiner, another analyst with Barclays, appeared to identify Intel as the culprit when questioning Suri about ReefShark, Nokia's range of 5G chipsets. "I mean, you had this issue with ReefShark with Intel's 10-nanometer process," he said. Suri did not dispute the remark, and Intel had not responded to a request for comment at the time of publication.

Nokia is now trying to clean up the mess by moving from these externally developed FPGAs to what Suri describes as "equity SoC-based [system-on-a-chip-based] products." That will entail a significant increase in 5G research-and-development costs. The shift began last year and gathered momentum in November when Nokia replaced Marc Rouanne, then head of its mobile networks business, with Tommi Uitto, who previously led mobile network sales. "I expect that we'll be in a much better position at the end of 2020, beginning of 2021, but it will progressively mitigate over the course of the year," said Suri.

The Finnish vendor's share price has tumbled about 28%, compared with its level at the start of last week, and there is little evidence of analyst sympathy. Genovese calls the decision to make FPGA-based 5G hardware "short-sighted" and says the negative impact on gross margins is likely to be felt throughout next year.

Scandi noir
It comes at an especially bad time. Having ringfenced its own research-and-development unit from broader cost-cutting activities, and offloaded various non-mobile assets, a stronger Ericsson is suddenly on the attack. Its mission, it has said, is to reclaim some of the market share it lost in the 4G era when China's Huawei was in the ascendant. The strategy seems to be working. Ericsson says its market share is now growing by around one percentage point annually -- a figure that tallies with research by analyst firm Dell'Oro.

The danger for Nokia is not necessarily that its 5G gear is inferior to Ericsson's but that Nokia's weaker margins give the Swedish firm the upper hand. Ericsson has indicated a willingness to accept less profitable 5G contracts for market share. That could mean replacing another vendor's 4G network as part of a new 5G deal. Nokia's cost predicament gives it far less wiggle room in negotiations. It cannot be as aggressive as Ericsson without risking a loss. But refusing to countenance a loss could mean ceding market share.

Want to know more about 5G? Check out our dedicated 5G content channel here on
Light Reading.

The pressure on the Finnish company is huge. With its new R&D plan, it has slashed its latest cost-saving target to €500 million ($555 million) in annual cuts, from €700 million ($776 million) previously. Nokia has also paused dividends until the net cash balance has risen to €2 billion ($2.2 billion), from €344 million ($382 million) in the recent third quarter. It still expects a strong final quarter, but any setbacks in China, whose operators are due to award 5G contracts, could trigger further upset.

Some analysts continue to look on the bright side. "We think the guidance has now been set low enough that the next revision is more likely to be up," said Genovese in his research note. "We also think it is still extremely early in the promising 5G cycle." Any more disappointment and even the most optimistic observers will find it hard to keep smiling.

Related posts:

— Iain Morris, International Editor, Light Reading

DanPittPaloAlto 10/31/2019 | 7:32:23 PM
Share price y-axis starting at zero? Hi Iain,

Many thanks for this well-written article. But could you produce a figure of the share price in which the y-axis starts at zero? The way it looks now, the stock price appears to have collapsed to a small fraction of its price. But as the y-axis starts at a very high number that is deceiving. I first learned about this method of deceiving the reader (which I do not at all think is your intention) in high school when I read a fantastic little book called "How to Lie with Statistics". It's still available on Amazon.

Dan Pitt
Sign In