AI's Investment Implications


#101

https://www.investors.com/research/industry-snapshot/artificial-intelligence-investing-it-doesnt-take-a-rocket-scientist/

It doesn’t take a rocket scientist to know that the cutting-edge technologies known as artificial intelligence will transform many U.S. industries.

Goldman Sachs consider it possible that the next Apple (AAPL) or Facebook (FB) could be an AI startup.

Nvidia aside, there’s a slew of tech companies that Wall Street analysts tout as possible AI stock plays. They include enterprise software providers like Salesforce.com (CRM), internet giant Facebook, tech consultancy Accenture (ACN), Alphabet’s (GOOGL) Google, Microsoft and lesser known names like Splunk (SPLK).

Need to load up more SPLK :roll_eyes: Have thought is a big data play :thinking:

Computer security specialist Palo Alto Networks (PANW) uses AI machine learning, which involves computers sifting through and learning from data, to detect hacker threats.

Whoa, PANW also AI stock. Bought only 20… time to double or triple my stake.

AI as a service: GOOG, AMZN, IBM and MSFT
AI enabler: NVDA
AI as a weapon: AAPL, CRM, NFLX, PYPL, PANW, FB and ZEN. Articles mentioned these but almost all companies would be here.

Which one of these could be next FAANG?
Graphcore, Cerebras, AIMotive, Sentient Technologies, Ayasdi, Digital Reasoning Systems, Vicarious Systems, DataRobot, StackPath, Flatiron and Zoox.

CV1 chip of AMBA is an AI chip, but can it compete with NVDA and INTC?


#102

Buy these seven shares to profit from driverless cars and artificial intelligence

NVDA - Obvious
XLNX - manch’s favorite
CGNX - This stock soars like an eagle. Think we are Johnny come late.
2 are listed in UK or Germany, and 2 are for driverless cars.


#103

XLNX is only one of my bets, not sure whether it’s my favorite. I am more certain about Micron.

CGNX looks interesting. Need to learn more about it…

The training war is done. Nvidia won. We are too late for that. Although Xilinx and Intel may have some design wins there but those are small and won’t move the needle. The industry has settled on Nvidia.

The far bigger war is inference. You only need to train once in the data center, but apply many, many times on difference devices in the field, or “at the edge” as cool kids would call it. Nvidia doesn’t have an advantage there. Pretty much everybody in the semiconductor industry is angling for a piece of the action.


#104

Actually we may not be too late. Nvidia could potentially double from the current level.


#105

Added two names to my AI portfolio. Both listed in Shenzhen.

000333.SZ Midea Group. Chinese robotics company. It even makes robot vacuum cleaners…
002415.SZ. HIKVision. Chinese surveillance company. Has lots of surveillance camera data and leader in facial recognition.

You won’t find much coverage in the English media. Google with their Chinese names if you want to know more.


#106

WSJ: AMD and Intel planning joint PC chip to take on Nvidia


#107

The semiconductor industry today reminds me of the “Seven Warrior States” period in ancient China. Companies are seeking alliance with one another while at the same time busy fighting each other.

Must be a lot of fun being CEO at a semi company now. Wartime!!


#108

Which one is the Qin?

Why 7 warrior states and not romance of the 3 kingdoms?

NVDA
INTC
AMD

Only 3.

In the 7, Qin won.
In the 3, none won.


#109

Amd, intel, and nvda. Who are the other 4?


#110

Broadcom is swallowing Qualcomm. TSMC and Samsung are big players too.

In fact we have more than 7. :smile:


#111

#112

Assuming manch is referring to semi that would benefit from the rise of AI and listed on US stock exchanges:

NVDA
INTC
AMD
MU
QCOM
XLNX
TSM

Still not 7 :slight_smile:
Edit: Added TSM = TSMC, SSNLF is not listed on US stock exchange

Non-semi companies like GOOG, smallish companies such as AMBA and startups are excluded from above list.

IoT would go AI too, which semi is best suited for IoT? ARM is an UK company and bought by SoftBank already. My take is IoT needs network security, hence I bought PANW. How about service providers like Comcast, AT&T and VZ?


#113

All chips are fabbed by either TSMC or Samsung. In the IoT world TSMC will do very well. Its stock price reflects that.

Also don’t forget Western Digital. We will need plenty of NAND.


#114

Totally forgot that WDC acquired SNDK. Need so many memory guys? MU and WDC?
I need only 1 AI semi :smile: Guess just long some calls instead of buying underlying.


#115

Where Major Chip Companies Are Investing In AI, AR/VR, And IoT


#116

We can’t invest in those.


#117

UBS: AI is growing like crazy — and these chip makers are set to dominate the industry (NVDA, AMD, INTC, QCOM)

Artificial intelligence is going to change every aspect of your life — here’s how to invest in it (NVDA, GOOGLE, INTC, QCOM)

Dropping AMBA from the prosperous 8, don’t think the CV1 chip can compete with the big boys nor as game-changing as the startups. Now has to decide which big boys to buy, NVDA or INTC or QCOM? The next AAPL/FB/GOOG AI startup is not listed yet, so best alternatives are these big boys.


#118

AMBA CV1 is not even an AI chip. It is computer vision and traditional computer vision has nothing to do with AI. AAPL is also pretty weak on AI. I haven’t seen any significant contribution from it. GOOG is the leader in AI, with full stack support from hardware to applications. Among the most notable AI scientists, Hinton works for GOOG. LeCun works for FB. I don’t know who Bengio works for.


#119

Presumed you didn’t know the context because you didn’t read the entire thread.


#120

Thanks for the nice summary. :slight_smile:
I agree with you that big companies stand a much better chance on the AI due to the huge amount of data needed for training purpose. In that front, I don’t think we’ll see any surprise.

However, there are some new algorithms which will give better chances for new startups. New algorithms such as GAN network generates its own training data on problems that can be evaluated automatically, such as the win/loss of a GO game. For these problems, the large dataset is no longer a prerequisite.