Intel Further Boosts CapEx to Meet Demand for 14nm Chips
by Anton Shilov on December 6, 2018 1:30 PM EST
Earlier this year Intel said that it had reaffirmed its plan to invest $1 billion in a bid to increase the number of chips made using its 14 nm process technology. Apparently, the processor giant did not stop there and is now further amplifiying its CapEx expenditures by $500 million for 2018.
Intel originally planned to invest $14 billion on capital expenditures (i.e., production equipment, building new facilities, and similar expenses) before this year, and then early on added an extra $1b which was reaffirmed mid year. Because demand for its multi-core server and client processors grew beyond the company’s expectations and Intel could not meet it, the company had to increase its production capacities. Back in September Intel said that it the additional $1 billion allocated on manufacturing tools to meet demand for its 14 nm products was in full swing. Since then, Intel has further increased its CapEx spending and now the expenditure will total $15.5 billion for the year, the company said this week.
“I think one of the consequences of getting $6 billion of upside in revenue is we have been chasing supply and some of our businesses are somewhat constrained in terms of equalizing demand and supply,” said Murthy Renduchintala, chief engineering officer and president of technology, systems architecture and client group at Intel, at the Nasdaq 39th Investor Conference. We have invested about $1.5 billion of extra CapEx since our January guidance, bringing our CapEx investment for 2018 to about $15.5 billion.”
Allocating an extra half of a billion USD on CapEx in several months indicates how severe Intel’s capacity constraints are today and gives an idea regarding Intel’s expectations based on orders coming from its clients.
Since Intel rarely discloses information about its production capacities, it is impossible to say how significantly it increased its 14 nm manufacturing capabilities this year and what targets did it set for Q1 – Q2 (when it is expected to meet demand). Meanwhile, it should be noted that $1.5 billion was mostly spent on switching to newer more advanced production tools that enable it to increase output of existing fabs. Many of these tools are going to be re-used for production of chips using more advanced process technologies, so Intel will take advantage of them for years to come.
Related Reading:
- Intel Investing $1B to Meet 14nm Demand: Prioritizing High-End Core and Xeon
- Intel 10nm Production Update: Systems on Shelves For Holiday 2019
- Intel’s Xeon Scalable Roadmap Leaks: Cooper Lake-SP, Ice Lake-SP Due in 2020
Sources: Intel, SeekingAlpha
Facebook
Twitter
RSS
20 Comments
View All Comments
FreckledTrout - Thursday, December 6, 2018 - link
It's still too much demand regardless of the reason. Sure it's very much due to the fact Intel did not spin up a new 14nm fab as they planned to move to 10+ but there is still very high demand. If they were overcharging then there sales would be down and they have had year on year record sales. Every single company overcharges for its products just as much as demand allows. I am not a fan of Intel either but everything you said is not accurate.HStewart - Thursday, December 6, 2018 - link
AMD would also charge more for there products if they had the demand - but they are mostly limited to desktop sales.sajal - Thursday, December 6, 2018 - link
Not new, already mentioned on their last earnings call!!HStewart - Thursday, December 6, 2018 - link
Sounds like they had plans for this,1. Get rid old and obsolete micron plant
2. Build up 14nm plants to meet demand
3. Put FAB 42 online for 7nm.
I believe most companies do such expenses at end of year for financial reasons. Preparing for 2019.
Bob-o - Thursday, December 6, 2018 - link
> 1. Get rid old and obsolete micron plantSurely you jest? Or just not familiar with the JV, and who had the upper hand?
Long_on_AMD - Thursday, December 6, 2018 - link
There are three reasons for Intel's 14 nm capacity shortfall; only two of which are typically mentioned. The first is AMD pressuring Intel to up their average CPU core count. Intel first launched a quad-core processor in January of 2012. Up until late last year, that was still the highest core count of their client CPUs. In the absence of real competition, Intel choose to slow-walk innovation and milk their customers. AMD's launch of Ryzen in the spring of last year finally forced Intel to slowly up the core count of their client CPUs, much as EPYC did the same on the server side. More cores = more area = more 14 nm demand.Then there are chipsets. Historically, these were always on the last node. That was the plan this time around, but Intel's 10 nm process was an abject failure, so now both CPUs and chipsets are competing for fab capacity at 14 nm.
But there is a third driver of 14 nm demand, which gets a lot less press. Security flaws, of which many have been revealed this year, have taken a dramatic toll on the performance of vast numbers of Intel processors in cloud and other server deployments. Hits of 30% or more are widely cited. Those server customers were hit hard and fast by these issues, and did not have the luxury of time to consider switching to AMD. It is a panicked "give me more of exactly what I bought ASAP to get my throughput back to what it was" reaction, which will goose Intel for a quarter or two, but at the prices they are charging, will lead to defections in the long run.
Dotans - Thursday, December 6, 2018 - link
Your first sound reasonable, the second is a really known issue. The third one is maybe one side of the large demand for servers CPU's.Another known 14nm issue is moving modem capacity inside Intel to 14nm and have to keep Apple happy with their desire quantities.
The last one is really an uptick in the CPU global demand. The PC is back and no one predicted it.
peevee - Monday, December 10, 2018 - link
" Intel first launched a quad-core processor in January of 2012."You must be kidding. Core 2 Quad (Kentsfield XE) is from 2006.
jjj - Friday, December 7, 2018 - link
Actually he does not state that they have increased the CAPEX further to boost 14nm supply and that means it's not that at all. He implies but does not state so the intention is to mislead.Opencg - Friday, December 7, 2018 - link
Intels long term plan to limit supply backfires when amd actually has competitive chips. Management and engineers communication and ability to work together cohesively fails miserably at egocentircly dominated company.