Toshiba Pares Losses on Report of Apple Investing in Chips | Hacker News Reader

Toshiba Pares Losses on Report of Apple Investing in Chips

bloomberg.com

106 pointsrayuela53 comments

Reuters' story mentions that Apple would take a 20% stake while Foxconn would own 30%. With Toshiba keeping a stake as well [1].

It's certainly not in Apple's interest to let this industry consolidate further [2]. Memory prices have already been on the rise and Apple is using more of it (that actually could be the reason why it's rising).

[1] http://www.reuters.com/article/us-toshiba-accounting-apple-i...

[2] https://www.statista.com/statistics/275886/market-share-held...

Over the past decade, the memory market has been flooded with cheap chips, causing most players in this market to lose value while breaking even or making a slim profit on memory sales.

Hopefully, Apple is planning to single source or majority source their memory through Toshiba, and give them funds to develop more advanced memory on newer processes, as that is what this market needs to develop further and faster. Apple may extend their engineering talent into building their own memory controller perhaps, similar to how they have done a ton of in house design for the ARM SoCs used in the iPhone/iPad lineup, instead of using off the shelf ARM designs.

Even if Apple invests in Toshiba's unit there's no way they're going to use them as a single source. They might give them a guaranteed percent of their NAND orders (especially if they get favorable pricing). It would still be in Apple's interest to shop around.

Also they already design memory controllers. They purchased the Israeli company Anobit in 2012 to do just that [1]. That's why Apple's storage benchmarks on recent products are really good.

[1] https://en.wikipedia.org/wiki/Anobit

Apple put their storage on a pcie bus, not too surprising it performs so well..

I am surprised to hear the negative sentiment on the memory market as for most users who build computers don't always look for the high-end memory. In fact, I am not sure how much performance gain one get from a high-end RAM on a gaming system.

IIRC it can give a 30% to 40% performance boost on APU based systems, as the increased bandwidth to the GPU allows for much faster processing. Alternatively, Intel decided to put a few hundred megs of cache on their iGPU, which has made their graphics performance much better, and lessened the gains from increased Ram speed.

https://semiaccurate.com/2014/01/15/exploring-effect-memory-...

Gamer machines don't use APU, they have a dedicated graphic card.

Ugh, I've been trying to tell my friend that, but he won't even invest in the most minimal of GPUs.

Ryzen is directly affected by memory speed due to the fabric ticking at the mem frequency

About 1% last I've seen benchmarks about that. It's so little, it's rather difficult to measure and it doesn't justify the investment.

Vertical integration (e.g. designing and fabricating their own hardware components, sourcing their own raw materials) is probably the best investment of Apple's cash reserves.

Why? Don't they benefit tremendously from building (mostly) on an ultra-competitive, near-commodity components market?

When they want something strategic (Retina screens, custom ARM) they make a big investment, but I can't see how getting directly into the fab business gives them any kind of long-run advantage.

Apple has huge offshore cash reserves. Investing (or acquiring) an offshore company makes sense for them. There are some speculations[1] about Apple buying Disney. I do not think that will be possible without a major tax break.

[1] http://www.businessinsider.com/rbc-analysts-speculate-apple-...

The Apple-Disney fantasy has been peddled around for years. It crops back up every couple of years, when some analyst or business writer dusts it off for attention.

For example, 2011:

http://www.marketwatch.com/story/why-apple-should-buy-disney...

The first time I heard this was back in 1999 and the reason they were merging was because the bite out of the apple and the "3" in the G3 logo looked like a sideways Mickey head: http://i.imgur.com/YaUWcXM.jpg

This idea made more sense when Jobs was leading Apple. He had significant shares in Disney due to his Pixar holdings, and at the time there was cross-board membership between the two companies.

Of course maybe that was a good reason for them never to merge while still coordinating strategically.

FYI, Bob Iger (Chairman and CEO of The Walt Disney Company) sits on the Apple board.

https://www.apple.com/pr/bios/

There was one about Apple buying Twitter that I enjoyed reading. I wish they would either do that or a similar acquisition, it would put their foot into the cloud (outside of iCloud) in a way that might be interesting.

Apple-Nintendo seems more likely in the context of offshore cash reserves.

Personally I hope they acquire Tesla to acquihire Elon Musk and get innovation going again. Apple is the new Microsoft, boring by profitable.

I honestly believe that if Apple and Tesla were to be come one, it wouldn't bring Apple up, it would take Tesla down.

With no visionary at the helm, Apples reputation for readily usable tech seems to be decreasing while Tesla has been advancing their own technology at a steady pace.

I hope this never happens.

If you have another view on how things would improve if they did merge, I'd love to hear it.

I could understand Apple going into CPU or GPU business. After all when you have 250M annual devices shipment, roughly the size of the whole PC market segment, these high margin cost could be eliminated, not to mention they aren't getting the best or what they need.

But Memory? But NAND and DRAM are commodity, in fact for most of their life time Memory companies tends to lose money. Not to mention China are pouring in tens of billions into memory Fabs alongside with billions of other incentives. In pretty much every business Chinese steps in is a race to bottom. I just dont see any value of getting a holding on to a memory fab / capacity.

Tell that to Samsung. Their 3d vnand gave them a crushing lead for the last several years in the nand market.

Apple is already designing their own chips, including both CPU + GPU:

https://en.wikipedia.org/wiki/Apple_A10

https://en.wikipedia.org/wiki/Apple_mobile_application_proce...

I guess running a chip foundry isn't something they would benefit a lot from (they can contract for process improvements just as well as they could build them out).

This could change when/if computational RAM goes mainstream.

Well, I can imagine that if you can build a GPU, then you can also build a memory with processing elements attached to it. Therefore, going after CPU/GPU companies seems more sensible.

Memory/processing integration is not as simple as it sounds: https://electronics.stackexchange.com/questions/134585/preci...

For Apple, you should not think in terms of profit. They probably care more about power consumption and having the best of the best.

Owning that much cash, why they don't just buy the companies that they need?

Apple as an ARM cofounder could have bought ARM in the past but they didn't want to own it completely if these could be useful for other industry players. Foxconn would be benefits from Toshiba.

No way a foreign company will be allowed to buy Toshiba due to national security concerns. One of domestic players will be to chip in and be partial owner.

Toshiba has been having a lot of trouble lately thanks to Westinghouse, so the Abe's government might consider foreign control to be a lesser evil than massive layoffs or a bailout.

https://www.bloomberg.com/news/articles/2017-04-11/toshiba-r...

Why national security..? If there's a world war and all of a sudden Japan doesn't have computers? (genuinely curious)

>If there's a world war and all of a sudden Japan doesn't have computers?

Pretty much. If you're at war you need that materials to feed the war machine, which includes the metal to build the plane as well as the computers that go in the cockpit (and let's not forget the knowledge to build it all).

Got it.

Kind of makes sense.

I don't think the issue is actually national security as much as preserving Japan's relatively closed economic system of interlocking corporate ownership / keiretsu.

Why am I being down-voted? That's the standard bargain Japan has with the US. It goes back to the Cold War. I'm not passing judgment; it's just a known matter of fact that their economy is rather closed to outside investment, especially controlling interests.

They just don't want the Chinese to get DRAM fab technology and then flood the market with cheap parts.

Perhaps they meant economic security.

If you operate in too many fields it may get difficult to take care of all of them. Parts of the organization can become inefficient. Also it means that top management needs to split their focus on more topics.

Apple buying the memory business seems uninteresting. That's a commodity business with low margins. Toshiba does make ARM CPUs, and Apple might want to get into the CPU business, but that's not the business being offered for sale.

> Apple might want to get into the CPU business

Apple is already in the CPU business?

Every iPhone and iPad sold in the last 5 years has used an Apple designed CPU core. [0]

Apple has chosen to keep the CPU exclusive to their mobile products as they obviously feel it gives them a competitive advantage. They could sell the SoC to others if they wanted to.

[0] https://en.wikipedia.org/wiki/Apple_A6X

Amazing, the A10 is a quad-core 2.34 GHz CPU. My Late-2016 MacBook Pro has a 2 GHz Intel Core i5.

I don't know a lot about CPUs, but isn't the A10 more powerful? Why don't they use them for their laptops, already..?

You can't compare CPUs from different architectures (ARM vs Intel) based purely on clock speed. There are a lot of other differences which mean things like the A10 generally aren't faster than similarly-clocked laptop chips.

ARM-based chips (like the A10) tend to have a much smaller thermal envelope and much less battery draw. They were built ground-up for mobile use, while x86 was built for speed & power.

I can't think of many fanless x86 laptops -- the only one that comes to mind is the current 12" Macbook, but that makes a ton of compromises and is running a CPU that's basically 5-6 years old. On the other hand, the idea of a cell phone with a big heatsink and fan assembly is laughable.

That's true.

Yes, I don't know a lot about CPUs.

It's true, a recent MBP will destroy the iPhone or iPad in a benchmark. You can't compare Apples to oranges with CPU speeds alone. However, you can run a benchmark. And on a popular one, GeekBench, the iPhone generally beats the non-Pro MacBook.

And:

"The iPhone 7 scores better on both single- and multi-core than any MacBook Air ever made, and performs comparably to a 2013 MacBook Pro."

https://browser.primatelabs.com/v4/cpu/2385400

https://browser.primatelabs.com/geekbench3/6170528

https://9to5mac.com/2016/09/15/iphone-7-geekbench-benchmark-...

where can I find an iPhone with single-core A processor that outperforms a MBP?

The benchmark runs either single core or multi core. Is not a single core iPhone.

The A10 isn't a "true" quad-core - it's a 2x2 design with 2 "heavy" (processing efficient) and 2 "light" (power efficient) cores. But, importantly, only one type can be active at a time - ie. you effectively only have 2 cores.

https://en.wikipedia.org/wiki/Apple_A10 has more information.

It would make sense to get Intel since lot of development tools and businesses relied on Intel technology. We won't know until LLVM, Clang, Rust and Swift could replace C/C++ in the future.

I see.

Perhaps they could do a virtual machine type of thing like they did when they transitioned to Intel (was it called Rosetta?).

> Why don't they use them for their laptops, already..?

They will, it's just a matter of when.

Reply on news.ycombinator.com