Intel Takes the Crown of World's Largest Semiconductor Supplier in 2019

[AleksandarK]

Intel is set to become the world's largest semiconductor supplier of 2019, according to the research from IC Insights. Intel held a crown for the largest semiconductor supplier since 1992, until 2018 when Samsung overtook it because of the booming DRAM business driven by high demand and not enough supply. Being Samsung's main business, any DRAM price/demand fluctuation was having a massive impact on its business. Due to high demand and high pricing, Samsung saw a massive revenue jump which resulted in a new king of the world's largest semiconductor supplier.

However, having seen predictions for a fall of 34% for this year, the decrease in demand will result in lower revenue for all DRAM market suppliers. SK Hynix, Micron and Samsung are expecting their revenues to decline around 29% on a year-over-year basis given the situation. This is resulting in lower revenue for Samsung than Intel has, and makes Intel the king of semiconductors once more. Intel's revenue is expected to reach around 70 billion USD, which is similar to last year's numbers.


View at TechPowerUp Main Site

 

[champsilva]

People: Intel is hurt, Intel shortage is going to have massive loss in profit.
Intel: Record after record,become (again) the largest semiconductor.

 

[cucker tarlson]

Yet

 

[kapone32]

And yet they can't keep up with demand. Anecdotal data claims that AMD is seriously clawing back market share. The 10NM debacle is never-ending. Yet another vulnerability. It really scares me to think what Intel would have made if none of these challenges existed.

 

[londiste]

It really scares me to think what Intel would have made if none of these challenges existed.

Why does that scare you? Intel would have made Cannonlake on 10nm in mid-2016.

 

[kapone32]

Why does that scare you? Intel would have made Cannonlake on 10nm in mid-2016.

Well they had already said they expect to have revenue of $70 billion for 2019.....without those mitigations I mentioned. That is an insane amount of money for any company to make.

 

[Steevo]

Well they had already said they expect to have revenue of $70 billion for 2019.....without those mitigations I mentioned. That is an insane amount of money for any company to make.

And to invest back into R&D, and their fab plants. There is no Scrodge McDuck vault of gold they go swimming in, that money is used to pay other people, buy everything from coffee cups, to private planes. Which also provide jobs for people Making those items.

Profit is not evil, evil is what you can do with profit and has nothing to do with the money.

 

[DeathtoGnomes]

And to invest back into R&D, and their fab plants. There is no Scrodge McDuck vault of gold they go swimming in, that money is used to pay other people, buy everything from coffee cups, to private planes. Which also provide jobs for people Making those items.

Profit is not evil, evil is what you can do with profit and has nothing to do with the money.

Evil is ignoring the corpses of people abused to get that profit.

 

[fancucker]

14nm parts that maintain a ST advantage over 7nm competitors and two consecutive uarchs (ice/tiger) with massive ipc gains. I wouldn't worry.

Even backported ICE-L would ward off AMD hence their scrambling towards 5nm and Zen 4. Zen 3 will deliver modest gains at most. AMD also better pray 10nm issues arent fully resolved.

 

[ZoneDymo]

in other news, grass is green.

 

[yeeeeman]

Why are people surprised? It is much better from a business point of view to use an older process which is cheap, has amazing yields and is done in house. I bet Intel spends peanuts on 14nm cpus, compared to AMD fabbing stuff on 7nm TSMC. That is how Intel is still making profit, no problem.
Adding to this, 9th gen parts are still very good, even if they are overtaken by Zen 2 in several metrics. The fact that a 5 year old uArch can have so much life in it is nothing short of amazing and Intel is doing the rights things in my opinion. 14nm process is not bad either, getting to 5Ghz at crazy low voltages, efficiency is very good in low power parts. I really believe that Intel found, by accident, a great business model. Just keep reiterating an old process, improve silicon, add more frequency, add more cores, while your process is getting more and more efficient in yields, costs, frequency...
But they need to move to a better process in the end.

 

[dicktracy]

14nm parts that maintain a ST advantage over 7nm competitors and two consecutive uarchs (ice/tiger) with massive ipc gains. I wouldn't worry.

Even backported ICE-L would ward off AMD hence their scrambling towards 5nm and Zen 4. Zen 3 will deliver modest gains at most. AMD also better pray 10nm issues arent fully resolved.

Golden Cove is the one to look forward to. This is the one that will highly likely to resume their monopoly.

 

[R0H1T]

The fact that a 5 year old uArch can have so much life in it is nothing short of amazing

Easy there, the uarch is not 5 years old. Not unless you want to add an * to that statement, like the ones we see in presentation slides

 

[Bones]

Evil is ignoring the corpses of people abused to get that profit.

Which goes right back to Steevo's statement.
Profit itself didn't do anything, if there are any corpses it was people that made it so and their lust for profit was the reason for it.

People do things, profit just..... Sits and does nothing on it's own.

Samsung and others in that part of the market have done some shady stuff related to plant fires, shortages to drive pricing.... We've seen it and know it's true.

 

[KithKhan]

By 'largest', they mean transistor size, right?

 

[TheGuruStud]

By 'largest', they mean transistor size, right?

Largest dies wasted.

Golden Cove is the one to look forward to. This is the one that will highly likely to resume their monopoly.

In 2023? Lol. Look at the roadmap without drunk goggles for the true timeframe haha