Contrasting quarterly fortunes for Intel and Qualcomm
Intel shocks investors with only 5% growth in servers, too little to offset PC decline and mobile failure; but Qualcomm beats forecasts on earnings and outlook
The chip giants have been basing their strategies on a core set of assumptions in the past couple of years, which are common across most of the majors. Basically – cloud and network infrastructure, good; human devices, especially PCs but also handsets, bad; non-human devices, for the Internet of Things (IoT), essential though potentially a bloodbath. But Intel’s and Qualcomm’s latest quarterly results highlight how unpredictable all these chip sectors are, with the former disappointing investors with poor server growth, while Qualcomm bucked the smartphone segment’s stagnant trend with strong performance in its core market.
In reality, then, the quarter-by-quarter results of Intel, Qualcomm, Nvidia, Samsung and the others point to far more fluctuating and unpredictable patterns than the grand strategic statements suggest – and that’s even without the new shock of ARM’s likely change of ownership when it is acquired by Softbank. The figures show that servers are not a guaranteed pot of gold after all; smartphones may have more life in them than expected, if companies can work out new ways to deal with China; and IoT devices are a poor prospect for any company which relies on respectable margins.
The growth of cloud services and mobile internet is certainly creating momentum in low power server chips, and Qualcomm, Cavium and other ARM-based players are heading that way, while IBM is breathing new life into Power, and Oracle into Sparc, for the same purpose. Yet the price pressures are enormous, the biggest cloud platform players are evaluating internally designed options, and other infrastructure builders, including mobile operators, are seeking to reduce capex by shifting their network priorities to software.
That servers are not always the golden goose after all was clear in Intel’s shock second quarter results, which saw a 51% decline in net income and server sales growing by only 5% year-on-year, instead of the double digits the company had forecast. Its data center business unit missed its revenue growth target for the third quarter in a row, and for the second consecutive quarter it reported growth below 10%.
Second quarter net income fell 51% to $1.3bn or 27 cents a share, hit by a $1.4bn charge for Intel’s aggressive job-cutting program, which will cost 12,000 positions. Revenue was up 2.6% to $13.5bn. Within that, data center group revenue was up 5% year-on-year to $4bn; while the Client Computing Group saw its revenue fall 3% to $7.3bn with operating profit of $1.9bn.
“While we remain cautious on the PC market, we’re forecasting growth in 2016 built on strength in data center, the Internet of Things and programmable solutions,” Krzanich said in the earnings statement. He expects the PC market to decline in a “high-single-digit” range this year, but he is looking for double-digit growth in all the other businesses for fiscal 2016.
For the third quarter, Intel expects revenue of $14.9bn, which would be ahead of Wall Street estimates, with gross margin fairly stable at 62% (it was 61.8% in Q2).
Intel certainly needs to deliver on these forecasts. If the first half pattern in server chips were were repeated in the second half of the year, Intel would be well short of its targets.
“Intel has some explaining to do,” Kim Forrest, an equity analyst at Fort Pitt Capital Group, told Bloomberg, reflecting the angry tone of the earnings call, in which analysts bombarded Intel with questions about getting server chip growth back on track. The company said several factors would help produce a better second half of the year in this market, including an improved outlook for large enterprise sales and new design releases.
“It’s not slowing down in the long term,” CEO Officer Brian Krzanich said on the conference call. “It’s going to be driven by the many more devices that are going to connect to the cloud.”
One issue in the infrastructure market is that sales are increasingly dominated by a few large hyperscale cloud firms, which may buy huge numbers of chips one quarter, for a new data center, and then very few the next. This “lumpy” pattern, as Krzanich described it, is far harder to predict and to present to shareholders than the relatively smooth distribution across the quarters, when most sales were spread among large numbers of enterprises.
This is another reason why Intel is suffering from the decline of the PC and the limited presence in other device sectors, since end user equipment has very different sales patterns to servers and can balance the peaks and troughs.
The great fear for Intel is that it will be unable to offset the decline of the core PC chip business with continuing server strength. The other businesses in which it has invested to counter that trend have either failed – smartphone processors – or are too immature to impact the figures significantly yet – Cloud-RAN, SDN, IoT devices. Some of its activities, like the former Infineon modem business and the network infrastructure division, are robust and would be a great asset to a smaller player, but Intel’s scale, and the size of the PC hole it will need to fill, make them look like mere drops in the ocean.
It now has a high mountain to climb in server sales in the second half, and no sizeable mobile or IoT business to fall back on. And Krzanich’s confidence about a server bounceback is not shared by everyone. The company is becoming over-reliant on a few giant customers, and suffering from the trend for enterprises to scale back on inhouse servers and rely on cloud partners. “You need a lot of growth in the back half and they’re awfully confident, but I don’t know,” said Stacy Rasgon, an analyst at Sanford C. Bernstein.
In recent quarters, Qualcomm and Intel have been almost like a weather house – when one is in the shadows, the other comes out into the sun of a strong earnings report. Qualcomm turned in a storming set of results, beating analysts’ forecasts both with its fiscal third quarter earnings and its future outlook.
Its profit was $1.4bn, or 97 cents a share, up from $1.2bn last year, on revenue up 4% to $6bn. Equipment and services revenue rose 1%, and licensing revenue 8.9%.
For the final quarter of its fiscal year, Qualcomm sees revenue coming in between $5.4bn and $6.2bn with adjusted earnings per share of $1.05 to $1.15.
In particular, it seems to have addressed the issue of Chinese companies withholding licensing fees more rapidly and decisively than many had expected – including some Qualcomm executives. Many Chinese companies held back on paying patent fees until the US vendor sorted out its antitrust issues with their country’s government, but even after that settlement was reached in February 2015, Qualcomm warned it would remain hard to collect the payments quickly (though they were reduced by its deal).
Even last November, Qualcomm’s president Derek Aberle was still saying there were many agreements left to negotiate, including “fewer than a handful” with Chinese OEMs which “in a negotiating tactic have stopped reporting sales and paying royalties”. He admitted they did not expect all these firms to agree terms during 2016.
But some breakthrough deals, like one announced at the start of this year with Xiaomi, have opened the floodgates and Qualcomm said on its earnings call this week that it was “overcoming resistance” to paying licensing dues among Chinese phonemakers and chip suppliers, though it has resorted to suing some firms, notably Meizu.
According to SeekingAlpha, Qualcomm now has 195 royalty-bearing single-mode OFDM/OFDMA licensees, up from about 180 in the previous quarter, and the number of CDMA-based licensees has improved from 310+ to 320+. It also benefited from delayed revenue recognition following the settlement of a patent dispute with LG Electronics.
And processor sales outdid expectations too. Qualcomm had predicted mobile chip shipments of 175m to 195m but actually turned in 201m, that was still down 11% from the year-ago quarter.
Of course, Qualcomm has daunting challenges, notably slowing growth in smartphones. Like Intel, it is hugely dependent on one core business and newer activities like server and IoT chips are not yet delivering significant percentages of its revenue. Yet the smartphone growth is forecast to grow by only 3.1% this year, compared to more than 10% in 2015 and 28% in 2014. So it is imperative that Qualcomm increases its market share to outperform that curve – a tough ask when it is already the market leader, and when there are powerful new competitors in Greater China.
CEO Steve Mollenkopf said it is outperforming the market and should gain share because of its improved performance in China, in chip sales as well as licensing. “It was a really strong quarter. It was strong in both businesses, in particular the license business,” Mollenkopf told Bloomberg. “We got some money a little earlier than we would have thought.”