我沒有說Intel(英特爾)太便宜，NVIDIA太貴喲!

sec2100

https://seekingalpha.com/article ... fter-intels-numbers


(以上全文連結)

sec2100

Soraemon
Yesterday, 2:34 PM

Comments (1)
|
@igli1234 not sure about nvdia beating intc in self driving, mobileye is miles ahead in products and partnerships and is actually launching tangible service soon such as robotaxi in germany, and partnerships with nio and toyota

sec2100

https://www.intel.com/content/www/us/en/research/overview.html

intel labs landing page

sec2100

The issue is when a traditional analyst tackles tech. They look at EVERYTHING about a company except what matters most.
The KEY element in tech in talent. Your greatest asset walks out of the building every day. Young talent, up on latest technologies is much more valuable than most older people, especially if they have been at a company for over 5 years. Brutal, I realize, but in my 45 years of tech, that is generally true.
That said, where would a newly minted PhD or a 30 something engineer at the peak of their skills want to work, hot NVidia or old, stodgy Intel.
While numbers matter, better to check GlassDoor or any of the list of forums on places to work. When a company in chips or software is considered a 'slave ship' by employees, that means more than the slick PowerPoint investors deck.

Maybe the new CEO can turn around Intel culture and/or match NVidia compensation (pay + options), but there is limited time.

sec2100

"where would a newly minted PhD or a 30 something engineer at the peak of their skills want to work"
This is a key point. NVDA has Selene, the 5th fastest supercomputer on earth, sitting in their data center in Santa Clara. I can't fathom the number of freshly minted AI/ML PhD researchers who want to go to work with that kind of resource in the tool box.
And then the idea NVDA are using Selene to design and build products and platforms has me pondering how ANY chip company keeps up without their own Selene.

sec2100

Fabless designers are a dime a dozen these days. Intel is going to start making graphics cards and could take market share at the low and mid range (along with AMD). High end cards are mostly used for crypto but that could change with the Ethereum 2.0 protocol changes or AI ASICs.
I have an Nvidia GTX 2070 which is a 3 year old card and works just fine for 99% of games on an UltraHD monitor.
Nvidia is a great company but a valuation north of half a trillion on relatively meager net income?
What will be interesting is if Nvidia can buy ARM holdings. I'm not bearish I just see the wisdom in owning a cash cow like Intel that is possibly on its way back up with Fab problems getting resolved.

sec2100

I will add to my NVDA position under $200. Probably around $180-$185.
Been buying Intel weekly for last 16 months.
New products coming (have to be on time as scheduled), New CEO, new product roadmap, new foundry expansion with massive amounts of increased capacity, huge amounts of yearly revenue and profits, dividend, and many catalysts over the next 5-10 years to capitalize on. MobileEye may be one of the smartest acquisitions of any company recently. Facts are that Intel is a great company, but under new mgmt, the market has Intel in a show me/prove it to me trading range.
If you believe in Pat, which I do, Intel is a strong buy here. Remember, before Pat, Intel was thinking of outsourcing most of its new tech and eventually becoming less and less a foundry. Pat immediately said heck no, that is not what Intel is at its core and he immediately doubled down. Took balls that early on. Also, the fact that many senior engineers followed Pat back to Intel. This may be one of the most bullish indicators because it shows he is trusted and respected.

Price target is $80 by the spring, but any positive product roadmap news, geo-political issues in Taiwan, and this thing could move higher quickly.
Should be a fun 5-10 years to watch it all play out

sec2100

Agreed that eventually valuations do matter. However, I don't believe the comparison of INTC to NVDA is as straightforward as you assume: i.e. they are both semiconductor companies and need to be valued by the same metrics. NVDA and INTC are very different companies even though they both manufacture semiconductors. NVDA has made good decisions over the years to focus more on high growth sectors of the semi market like AI and gaming. INTC has floundered with its tech roadmap and execution and has a much higher portion of its business coming from slower growing segments of the semi market. Confidence in growth also plays an important role in valuation of a stock and investors now have more confidence in NVDA's growth potential. INTC is in a position where it now has to prove it can grow at a higher rate and re-capture lost market share. Can it do it? The jury is out on this and it will take time (years) to prove this. Sure, by the numbers, INTC appears "cheap" but there is uncertainty now about its ability to execute. All this plays into the valuation. My suggestion is own some of both NVDA and INTC. They are different animals and you own them for different reasons: NVDA for growth, INTC as a turnaround/restructuring play. Maybe you hold your nose and own a smaller position in NVDA along side your larger position in INTC. I think it is a mistake not own some NVDA. I see your logic on INTC, but it may take a long time before you see a payoff on that bet.

sec2100

jayn
01 Oct. 2021, 10:57 AM

Comments (1.14K)
|
Intel's new EUV fab in Oregon is opening in 1H 2022. Their pilot line for the Meteor Lake chip began ramping in q3. Production to begin in mid 2022.
Intel has additional fab and packaging expansions planned in Ireland, Israel, Arizona and New Mexico, already announced. They stated they will announce, by the end of the year, additional large fab expansion plans in the U.S. and EU, tied to their IFS plans.
Their story, under the new CEO, is that they are going stop buybacks and sink their money into IFS. Aside from that, they were already targeting new developing markets in autonomous vehicles, 5G infrastructure, silicon photonics, Optane memory, ai ASICs and are putting their stake in the ground for discrete GPUs beginning next year.

sec2100

If Intel can build high volume on TSM N6, they can gain Xe-HPG GPU market share in 2022 on availability alone.


Don't forget Xe-HP and Xe-HPC, which are designed for 42TF FP32 and 42TF FP64, respectively.