[Cryptography] Did Spectre help torpedo Qualcomm?

Henry Baker hbaker1 at pipeline.com
Mon Sep 24 11:34:00 EDT 2018

FYI --


How Qualcomm Tried and Failed to Steal Intel's Crown Jewel

By Ian King September 20, 2018, 5:00 AM EDT

In early November, Qualcomm Inc. Chairman Paul Jacobs stood on a stage
in the heart of Silicon Valley and vowed to break Intel Corp.'s
stranglehold on the world's most lucrative chip business.

The mobile internet and cloud computing were booming and the data
centers running this digital economy had an insatiable thirst for
computer servers -- and especially the powerful, expensive server
chips that Intel churns out by the million.  Qualcomm had spent five
years and hundreds of millions of dollars designing competing
processors, trying to expand beyond its mobile business.  Jacobs was
leading a coming-out party featuring tech giants like Microsoft
Corp. and Hewlett Packard Enterprise Co., which had committed to try
the new gear.

"That's an industry that's been very slow moving, very complacent,"
Jacobs said on stage.  "We're going to change that."

Less than a year later, this once-promising business is in tatters,
according to people familiar with the situation.  Most of the key
engineers are gone.  Big customers are looking elsewhere or going back
to Intel for the data center chips they need.  Efforts to sell the
operation -- including a proposed management buyout backed by SoftBank
Group Corp. -- have failed, the people said.  Jacobs, chief backer of
the plan and the son of Qualcomm's founder, is out, too.

The demise is a story of debt-fueled dealmaking and executive
cost-cutting pledges in the face of restless investors seeking quick
returns -- exactly the wrong environment for the painstaking and
expensive task of building a new semiconductor business from scratch.
It leaves Qualcomm more reliant on a smartphone market that's
plateaued.  And Intel's server chip boss is happy.

"They were the ones that had enough clout to make a mark," said Alan
Priestley, an analyst at Gartner Inc.  "Qualcomm had the best chance."

The idea of a Qualcomm server chip business started in late 2012 as
the company looked for new markets beyond smartphones.  Server
processors can sell for more than $10,000 each in a market utterly
dominated by Intel.  The idea was to get the energy efficient
properties of mobile chips into a more powerful design that would
appeal to data center owners such as Microsoft and Google that were
looking to save on operating costs and needed a balance against Intel
in price negotiations.

Qualcomm repurposed its design center in Raleigh, North Carolina,
adding engineers from Intel, IBM and Advanced Micro Devices Inc.  It
was home to some of the most accomplished processor engineers in the
industry.  They'd already designed some of the company's best mobile
chips, products that made the San Diego-based company the biggest
provider of silicon for smartphones.

The 1,000-person team was tasked with creating chips capable of
performing some of the toughest jobs in computing, from making complex
calculations that predict the weather to mapping the human genome.
The most common use, though, is crunching the information flowing
through giant data centers owned by the likes Microsoft, Facebook,
Amazon and Google.  When you search for something on the web, check
your social-media app or buy something online, these are the chips
making it happen.  Intel rakes in almost $20 billion a year in revenue
selling the components.

To break into this business, you can't just make a good chip.  What's
really needed is a road map of future processors that will improve
over time dependably.  Then comes software and engineering support to
convince data center operators that all their programs and services
will work with the components.  One server chip can't just be popped
out and replaced with a different one in an afternoon.  The process
takes years and customers are loath to change what's already working
-- even if they're sick of paying Intel so much.

Just designing a server chip costs hundreds of millions of dollars
these days.  The last time Intel was seriously challenged was in 2006,
when AMD got a fifth of the market.  Work began on this AMD server
processor in 2000, when Mark Zuckerberg was still a teenager and the
word "cloud" referred only to white fluffy things in the sky.  AMD's
market share ended up back at almost zero as its new products came
late or didn't live up to their billing.  By 2015, some analysts
worried AMD might go bankrupt.

Qualcomm was one of the few companies capable of challenging Intel
again.  Its $5.5 billion annual research and development budget is
third only to Intel and Samsung in the chip industry.  By the launch
event in November last year, Qualcomm was beginning to attract public
support from companies such as Microsoft and Facebook, leading some
analysts to predict an end to Intel's lock on all the profit.

All was looking good as Qualcomm and Jacobs prepared for the
coming-out party in San Jose, California.  But two days before the
event, rival Broadcom Inc. announced it wanted to buy Qualcomm in the
industry's biggest-ever deal.

The hostile takeover battle that ensued pitted two divergent views of
the industry's future.  Broadcom's Hock Tan saw inevitable
commodification and preached consolidation and cost-cutting to
prepare.  He also planned to borrow as much as $106 billion to pay for
the deal, leaving less cash for R&D.  Jacobs argued that, rather than
pouring cash into debt payments, chip companies needed to keep
investing in emerging technology to find new markets.

The U.S. government blocked the deal in the end.  But before that
happened, many investors sided with Tan, forcing Qualcomm to slash
spending plans by a billion dollars.  The budding server business was
the main victim of those cuts.

By early this year, it was clear Qualcomm's support for the server
effort was evaporating, setting off a frantic round of deal
negotiations to try to save it.  Anand Chandrasekher, who oversaw the
business, persuaded a group of investors including SoftBank and
Singapore's Temasek Holdings Pte. to support a management buyout.
They made an offer, but Qualcomm Chief Financial Officer George Davis
demanded the transaction close by a specific time, scuttling the deal,
according to people familiar with the process.

After Jacobs was ousted from Qualcomm in March, he offered to buy the
server chip business, too.  He even proposed letting the company keep
a minority stake.  Qualcomm could avoid spending money on the project,
but could still profit later if it worked out.  But Qualcomm's board
insisted that Jacobs give up any attempt to return to Qualcomm and
abandon his plan to try take the whole company private, the people
said.  He refused.  Once again, the unit was left hanging.

Former Intel executive Renee James considered buying the unit to
integrate into her data center chip startup, Ampere Computing,
according to the people.  However, Chandrasekher had earlier brokered
a deal with the Chinese province of Guizhou to fund part of Qualcomm's
server chip work.  In return, the local government demanded the
transfer of chip designs and exclusive rights to sell the processors
in China.  Ampere's James balked at the conditions, which would have
locked her out of a giant market, the people said.

In May, Chandrasekher left after leading Qualcomm Datacenter
Technologies for almost half a decade.  Others followed.  By June, the
company had cut about 280 jobs in Raleigh and 43 in California.
Almost half the unit's engineers have gone now, and some of them are
working for James at Ampere.  That's left the Qualcomm unit short of
the expertise needed to continue, even if it had full funding,
according to former employees.

Qualcomm President Cristiano Amon said the company has "right-sized"
the server business for the market opportunities.  The division will
now concentrate on selling chips to a couple of the largest cloud
providers, and Chinese customers such as Alibaba Group Holding
Ltd. and Tencent Holdings Ltd. through the joint venture with Guizhou

Behind the scenes, work has been shelved on a successor to the chip
that rolled out with much fanfare at the November event.  Qualcomm
will consider adapting the existing design for use in smartphone base
stations, but as a direct competitor to Intel, the effort is over.

"Qualcomm became more like Broadcom even though the acquisition didn't
happen," said Shane Rau, a semiconductor analyst at IDC.  "You had a
company with deep pockets and an understanding of what the customer
base wanted.  The market was hungry for such a vendor."

Qualcomm Chief Executive Officer Steve Mollenkopf is now saying future
growth will come from the company's traditional strength in the mobile
phone industry, which will start a transition to so-called fifth
generation, or 5G, technology next year.  His board is backing him and
plans to announce a bigger pay package for the CEO, according to
people familiar with the situation.  The company also said it's on
course to generate $5 billion in sales from other markets in its 2018
fiscal year.

That leaves Intel, which is more than ever the only major supplier of
chips that power the internet.  In its latest quarter, the company
turned $5.5 billion of sales from its server business into $2.7
billion of operating profit.

This story doesn't completely ring true to me.

I wonder if the server chips Qualcomm developed were hopelessly
haunted by Spectre?

Yes, Intel chips were also affected by Spectre, but if the
up&coming Qualcomm chips were affected as much or more, then
Qualcomm may have lost an entire generation of chip sales while
they scrambled to fix Spectre-type problems.  For these customers,
better the devil that you know...

Jus' speculatin' ...

More information about the cryptography mailing list