I picked Equinix in 2019 because of their expansion into emerging markets. They don’t seem to be impacted by anything other than business as usual which is great to me. I wouldn’t try to make an entry now but Digital Realty is the only other stock I choose to anchor my already existing AI picks. They’ve performed well. So well I trimmed all of them to buy more of the stable and undervalued plays I picked. All in all this video was very informative. I will go take another look at this sector because of it.
I've invested in PAVE. Global X U.S. Infrastructure Development ETF. Not just datacenter oriented but the CHIPS act, Infrastructure bill and data centers plus buildout for more energy supply, all should keep things going for a while. Not the HOT ticket around AI but just good longer term business.
Global X seems to produce some decent ETFs. PAVE has an interesting collection of stocks and surprisingly $7.4 billion in AUM. Thank you for the comment.
I wonder if it makes sense to build data centers in naturally cold environments with low humidity. Perhaps, they can take advantage of outside air and low underground temperature for cooling.
Yes, as long as there is access to cheap electricity. The problem is that those places are far from where most people live, so network latency is a problem in many cases.
Good point. Microsoft was working on underwater data centers a while back (not sure what ever happened to that plan) and the Chinese are said to use this technique.
@@Nanalyze One of Google's fascinating failures were these floating data center barges that used the below water decks as a natural coolant. I remember liking the idea but I also liked the Google Glass so don't look to me for concept analysis.
@@fuhishva You've never been to Winnipeg Manitoba Canada, have you? Winter gets to -40 with low humidity. Winter is usually 7 months of the year. Temps 0celcius to -40c with a 1million population. So yes there are places like this. Also the 2nd lowest Electricity prices in the country as we sell our Hydro generated electricity to the USA.
Outstanding video, loved the deep dive on the ETF and how scammy these sales products are and why it’s so important to dig into the underlying businesses within.
Celestial AI is working to reduce power consumption by 90%. The core of Celestial AI's strategy lies in its chiplets, interposers, and optical interconnect technology. By combining DDR5 and HBM memory, the company aims to significantly reduce power consumption while maintaining high performance levels. The chiplets can be used for additional memory capacity or as interconnects between chips, offering speeds comparable to NVLink or Infinity Fabric.
When it comes to AI/Technology/Datacenter, or whatever they're calling themselves funds, I like VGT (similarly VITAX). The other play that never gets love is MSCI, what I consider the pick and shovel of the index fund world.
Joe P. here. You cannot beat Vanguard ETFs. VGT is really heavily weighted for top-ten (MSFT, APPL, and NVDA = 45%) and it's a sector index. Not a bad thing, just notable. Goes to show how big some companies are these days. Also, I worked at MSCI for over a decade so I know the business well. Maybe we'll do a piece on MSCI stock one of these days.
Above all else: NVDA. Then VRT for cooling (cream of crop, more efficient way to cool needs to be put into all data centers liquid immersion cooling). VST for energy.
Under your video is Vrt ceo on CNBC talking about AI 😂😂😂 right before you said it I went on robinhood and seen that the stock went from $8 to $77 that yells AI hype bubble, fast forward I unpause your video and you say the same thing 😉 first day subscribed and I’m learning
The best thing to do is search our website www.nanalyze.com for articles. For example, we last wrote about Xometry last year: www.nanalyze.com/2023/02/xometry-stock-short-report/
@@Nanalyze Wikipedia "List of small modular reactor designs" are almost all just studies. NuScale is a newcomer with stock symbol SMR ("small modular reactor") getting close to approval.
What for me the most interesting aspect regarding this AI topic is, is the shift from maximum 1 year ago regular investors thoughts "oh look, now everyone is putting the hype buzzword AI on their agenda" to "AI is here to stay"-mentality.
If you are referring to Marvell ,the short answer is yes on MRVL being a beneficiary of the building of new data centers and the revamp of existing data centers . Marvell made a strategic acquisition of a company by the name of IPHI in 2021 ,which allowed MRVL ,access to the optical processor market vs being limited to copper based processing within the data center servers . Going by the last earnings report from MRVL ,this investment in acquiring IPHI is yielding increases in data center revenues . By disclosure I am a long term holder in MRVL and was a holder of IPHI stock at the time of acquisition by MRVL .
@@RRGtherailroadguy The question is just how much exposure a MRVL investor gets to data center hardware. This should be broken out into a revenue segment so that investors can track it over time.
@RRGtherailroadguy yes, I watched their presentation recently . Was interested in the silicon photonics part. Doesn't look like marvel is an industry leader there . But my thoughts after listening were that marvels version could be a future leader.
When we did the math in my last interview, AI did the work 70-130x cheaper than a human would. When the C-suite figures out AI can replace workers, the demand for this will shoot to the moon and a lot of the savings can be taken directly by companies providing the compute, because it's going to become incredibly scarce for anything but productive work. The law of economics however will drive this profitability to fall off a cliff with infinite resources becoming available for data centers. It's just not plausible this transformation would really allow for capitalism as it currently is to survive, because once a position is automated, it's automated for everyone in that position everywhere in the world. So my advice is, invest like the world would survive, but understand you are betting on this technology to destroy jobs and livelyhood's on an absolutely eye watering scale to give you the profits you seek.
Gartner's Hype Cycle means that at some point we will overestimate the usefulness of every technology. If all companies will benefit from this, then simply investing in the broader market should show good results down the road.
@@Nanalyze Yes, but we've never seen anything this powerful advance this fast. The world will be shocked to discover how many jobs exist purely because they're ever so slightly too complex to hard code out of existence - but well within the reasoning ability of what ever is the next OpenAI model. All you really have to do is go look at Palantir's customer stories about what they've done with AIP to see the future: automated decision making. Today it's the millions of 5$ decisions, tomorrow who knows. Another interesting thing to ask yourself when watching those AIP video's is how much more can this do when by next year the underlying LLM's are 1000% better? If anything, AI is _underhyped_
@@MrVohveli Yes, AI is great and will do great things. Making money off that as a retail investor is a different story. Marketing videos from a company that sweeps underperforming financial metrics under the rug are to be taken with a grain of salt. The proof is always in the revenue growth.
The McKinsey research cited in this piece can be found here: www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/investing-in-the-rising-data-center-economy
Not too many good options for retail investors to get exposure to nuclear energy. We've covered a few in the past such as BWXT: www.nanalyze.com/2019/02/invest-nuclear-energy-stock/
VRT: high cost, low-margin industrial indecently voguing as part of the AI clown show. 7B sales & a negative PTB for 35B, Giordano? You better really like rack-mounted power supplies. Should be offshored post haste, starting with executive team.
AI issues are power and the entire model of chip manufacturing process. Edge Compute and a complete distributed process model is where it will move, but current telcos and cable cos are not financially prepared to accommodate. They are not prepared to integrate or navigate a regulatory environment which will challenge them and call for innovative thinking beyond the inadequate I infrastructure and inadequate power to support AI training data centers since “green energy is grid destabilizing and inadequate. Only nuclear build out can meet the goals, but must be accelerated.
Your comment on nuclear mimics what a lot of people are saying about AI's energy needs. Seems like the regulatory constraints need to be loosened and some government incentives put in place for it to grow.
@@Nanalyze True, but the issue remains that we’re 5-7 years behind being able to meet data center power needs as it now stands without upsetting consumer energy costs that would impact utilization of said data centers and other industrial electrical needs. We been working on new processor materials science tech for some years reducing power consumption by very large percentages and algorithms to complement that process. These needs have been ignored for decades with malinvestment and network structural build failings in last mile and large infrastructure
Nanalyze is interesting take, but I believe that this misses the data center serious issues: 1. POWER as the GREENIES have been all happy talk and no real planning like development of large scale nuclear. 2. New chip technologies that move more services to the edge and leverage better services utilization of infrastructure since the last mile and network long haul is weak, dated and unable to support demand. 3. New physics/chemistry that impacts chips, new materials science and completely new hybrid optical-radio structure in compute that replaces from data center to hand held device to satellite. It hardens devices against radio interference attack and other issues. AI as described thus far by Microsoft and others is going to be a legal quagmire and not a money maker. Curated specialized AI Services make more sense.
Good comment here. To be fair, we did mention edge computing as a trend to watch (per the McKinsey research). Nuclear is one way to solve the energy needs and it's quite green (arguably). Materials science may be spurred on by generative AI inventing new stuff. Your comment on AI being a legal quagmire is quite interesting. Thank you for taking the time to share your thoughts!
This seems to fall under the "standard networking equipment will see a demand as data centers grow" category. Since networking equipment is a mature category, many of the large companies that provide it will see some boost but it's likely to be muted unless they are pure plays.
Subscribe to our channel for more content on investing in disruptive technology:
ruclips.net/user/nanalyze ❤
I picked Equinix in 2019 because of their expansion into emerging markets. They don’t seem to be impacted by anything other than business as usual which is great to me. I wouldn’t try to make an entry now but Digital Realty is the only other stock I choose to anchor my already existing AI picks. They’ve performed well. So well I trimmed all of them to buy more of the stable and undervalued plays I picked. All in all this video was very informative. I will go take another look at this sector because of it.
Really glad to hear you found the video informative, thank you for the comment!
I've invested in PAVE. Global X U.S. Infrastructure Development ETF. Not just datacenter oriented but the CHIPS act, Infrastructure bill and data centers plus buildout for more energy supply, all should keep things going for a while. Not the HOT ticket around AI but just good longer term business.
Global X seems to produce some decent ETFs. PAVE has an interesting collection of stocks and surprisingly $7.4 billion in AUM. Thank you for the comment.
I wonder if it makes sense to build data centers in naturally cold environments with low humidity. Perhaps, they can take advantage of outside air and low underground temperature for cooling.
Yes, as long as there is access to cheap electricity. The problem is that those places are far from where most people live, so network latency is a problem in many cases.
Good point. Microsoft was working on underwater data centers a while back (not sure what ever happened to that plan) and the Chinese are said to use this technique.
@@Nanalyze One of Google's fascinating failures were these floating data center barges that used the below water decks as a natural coolant. I remember liking the idea but I also liked the Google Glass so don't look to me for concept analysis.
@@fuhishva You've never been to Winnipeg Manitoba Canada, have you? Winter gets to -40 with low humidity. Winter is usually 7 months of the year. Temps 0celcius to -40c with a 1million population. So yes there are places like this. Also the 2nd lowest Electricity prices in the country as we sell our Hydro generated electricity to the USA.
@@dbalcewich I never claimed cold places with cheap electricity do not exist. Only that they have small populations, which 1 million is.
Outstanding video, loved the deep dive on the ETF and how scammy these sales products are and why it’s so important to dig into the underlying businesses within.
Thank you for the positive feedback!
Celestial AI is working to reduce power consumption by 90%. The core of Celestial AI's strategy lies in its chiplets, interposers, and optical interconnect technology. By combining DDR5 and HBM memory, the company aims to significantly reduce power consumption while maintaining high performance levels. The chiplets can be used for additional memory capacity or as interconnects between chips, offering speeds comparable to NVLink or Infinity Fabric.
Cool tech and getting lots of funding. Still private so we don't know much. We'll see how it goes. Thank you for the heads up!
This was awesome. Exactly the info I was wanting
Great to hear! We love getting it right.
When it comes to AI/Technology/Datacenter, or whatever they're calling themselves funds, I like VGT (similarly VITAX). The other play that never gets love is MSCI, what I consider the pick and shovel of the index fund world.
Joe P. here. You cannot beat Vanguard ETFs. VGT is really heavily weighted for top-ten (MSFT, APPL, and NVDA = 45%) and it's a sector index. Not a bad thing, just notable. Goes to show how big some companies are these days. Also, I worked at MSCI for over a decade so I know the business well. Maybe we'll do a piece on MSCI stock one of these days.
@@Nanalyze Thanks Joe! Would love to hear a trip down memory lane on MSCI one day :)
Above all else: NVDA. Then VRT for cooling (cream of crop, more efficient way to cool needs to be put into all data centers liquid immersion cooling). VST for energy.
We covered VRT here: ruclips.net/video/2z69hVzMy3E/видео.html
@@Nanalyzethx. Exciting times
Vertiv VRT has done me proud.
Vertiv appears hyped. We have another piece on data center stocks coming out soon.
infrastructure and energy stocks that needed for AI as well
We touched on the "AI driving electricity demand" thesis here: ruclips.net/video/nT4kbdJRYc8/видео.html
Data is the new GOLD!!!
AI algos are only as good as the data you feed them
Under your video is Vrt ceo on CNBC talking about AI 😂😂😂 right before you said it I went on robinhood and seen that the stock went from $8 to $77 that yells AI hype bubble, fast forward I unpause your video and you say the same thing 😉 first day subscribed and I’m learning
That's great to hear you're learning from our videos. That's all we could ask for.
Just came across your XMTR vs PRLB article from 3 years ago. Would love to see you revisit that 😊
The best thing to do is search our website www.nanalyze.com for articles. For example, we last wrote about Xometry last year: www.nanalyze.com/2023/02/xometry-stock-short-report/
AI data centres will need their own small nuclear power, eg Small Modular Reactors (SMRs).
That's what they say! We'll keep an eye out for ways retail investors can play that.
@@Nanalyze Wikipedia "List of small modular reactor designs" are almost all just studies. NuScale is a newcomer with stock symbol SMR ("small modular reactor") getting close to approval.
@@formica. We covered them ,here: www.nanalyze.com/2022/01/nuscale-power-stock-nuclear/
What for me the most interesting aspect regarding this AI topic is, is the shift from maximum 1 year ago regular investors thoughts "oh look, now everyone is putting the hype buzzword AI on their agenda" to "AI is here to stay"-mentality.
Both can be true, and are true frankly.
@@Nanalyze True.
Is marvel a beneficiary or data centers?
Good question. We haven't researched Marvel at all so wouldn't be able to offer up an intelligent answer.
If you are referring to Marvell ,the short answer is yes on MRVL being a beneficiary of the building of new data centers and the revamp of existing data centers . Marvell made a strategic acquisition of a company by the name of IPHI in 2021 ,which allowed MRVL ,access to the optical processor market vs being limited to copper based processing within the data center servers . Going by the last earnings report from MRVL ,this investment in acquiring IPHI is yielding increases in data center revenues . By disclosure I am a long term holder in MRVL and was a holder of IPHI stock at the time of acquisition by MRVL .
@@RRGtherailroadguy The question is just how much exposure a MRVL investor gets to data center hardware. This should be broken out into a revenue segment so that investors can track it over time.
@RRGtherailroadguy yes, I watched their presentation recently . Was interested in the silicon photonics part. Doesn't look like marvel is an industry leader there . But my thoughts after listening were that marvels version could be a future leader.
When we did the math in my last interview, AI did the work 70-130x cheaper than a human would. When the C-suite figures out AI can replace workers, the demand for this will shoot to the moon and a lot of the savings can be taken directly by companies providing the compute, because it's going to become incredibly scarce for anything but productive work. The law of economics however will drive this profitability to fall off a cliff with infinite resources becoming available for data centers.
It's just not plausible this transformation would really allow for capitalism as it currently is to survive, because once a position is automated, it's automated for everyone in that position everywhere in the world. So my advice is, invest like the world would survive, but understand you are betting on this technology to destroy jobs and livelyhood's on an absolutely eye watering scale to give you the profits you seek.
Gartner's Hype Cycle means that at some point we will overestimate the usefulness of every technology. If all companies will benefit from this, then simply investing in the broader market should show good results down the road.
@@Nanalyze Yes, but we've never seen anything this powerful advance this fast. The world will be shocked to discover how many jobs exist purely because they're ever so slightly too complex to hard code out of existence - but well within the reasoning ability of what ever is the next OpenAI model.
All you really have to do is go look at Palantir's customer stories about what they've done with AIP to see the future: automated decision making. Today it's the millions of 5$ decisions, tomorrow who knows.
Another interesting thing to ask yourself when watching those AIP video's is how much more can this do when by next year the underlying LLM's are 1000% better?
If anything, AI is _underhyped_
@@MrVohveli Yes, AI is great and will do great things. Making money off that as a retail investor is a different story.
Marketing videos from a company that sweeps underperforming financial metrics under the rug are to be taken with a grain of salt. The proof is always in the revenue growth.
Wonder if it would be feasible to use helium for cooling. NASA and rocket companies use it for cooling. Just a thought.
Interesting idea
It’s expensive and limited. MRI uses helium to cool
@@joeschnur4632 Good to know, thank you.
Cant wait till its more adopted im looking to leverage in helium 👍
Where can we find the McKinsey research?
The McKinsey research cited in this piece can be found here: www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/investing-in-the-rising-data-center-economy
Nuclear energy
Not too many good options for retail investors to get exposure to nuclear energy. We've covered a few in the past such as BWXT: www.nanalyze.com/2019/02/invest-nuclear-energy-stock/
My wife's the leading producer of Artificial Intelligence.
*ba-dum tss* 🥁
(If you see this babe, love you and your many degrees. 🙏)
Underrated comment this
VRT: high cost, low-margin industrial indecently voguing as part of the AI clown show. 7B sales & a negative PTB for 35B, Giordano? You better really like rack-mounted power supplies. Should be offshored post haste, starting with executive team.
Definitely hyped
AI issues are power and the entire model of chip manufacturing process. Edge Compute and a complete distributed process model is where it will move, but current telcos and cable cos are not financially prepared to accommodate. They are not prepared to integrate or navigate a regulatory environment which will challenge them and call for innovative thinking beyond the inadequate I infrastructure and inadequate power to support AI training data centers since “green energy is grid destabilizing and inadequate. Only nuclear build out can meet the goals, but must be accelerated.
Your comment on nuclear mimics what a lot of people are saying about AI's energy needs. Seems like the regulatory constraints need to be loosened and some government incentives put in place for it to grow.
@@Nanalyze True, but the issue remains that we’re 5-7 years behind being able to meet data center power needs as it now stands without upsetting consumer energy costs that would impact utilization of said data centers and other industrial electrical needs. We been working on new processor materials science tech for some years reducing power consumption by very large percentages and algorithms to complement that process. These needs have been ignored for decades with malinvestment and network structural build failings in last mile and large infrastructure
@@PlanetFrosty Good points
Nanalyze is interesting take, but I believe that this misses the data center serious issues: 1. POWER as the GREENIES have been all happy talk and no real planning like development of large scale nuclear. 2. New chip technologies that move more services to the edge and leverage better services utilization of infrastructure since the last mile and network long haul is weak, dated and unable to support demand. 3. New physics/chemistry that impacts chips, new materials science and completely new hybrid optical-radio structure in compute that replaces from data center to hand held device to satellite. It hardens devices against radio interference attack and other issues.
AI as described thus far by Microsoft and others is going to be a legal quagmire and not a money maker. Curated specialized AI Services make more sense.
Good comment here. To be fair, we did mention edge computing as a trend to watch (per the McKinsey research). Nuclear is one way to solve the energy needs and it's quite green (arguably). Materials science may be spurred on by generative AI inventing new stuff. Your comment on AI being a legal quagmire is quite interesting. Thank you for taking the time to share your thoughts!
Nividia and vrt together
No cheerleading please. You can't just repeat Fintwat mantras and expect that wealth will follow. Try to add some value to the conversation.
In a gold rush, sell shovels
Good point
When are you roasting another portfolio la?
We would never do such a thing
neither etf would work for this convo
Precisely
Holding a 10% lost on Snowflake, let see how it works out
Wasn't it Buffett who once said don't hold a stock unless you're comfortable seeing the share price halve? Always invest in companies, not stocks.
@@Nanalyze funny Buffert hold this stock as well. And Im hold this stock because I'm the data engineer and my companies use Snowflake lol
Good point! Note that Buffett's position is very very small relative to his broader portfolio.
Its crazy how much data you have to cover and analyze just to understand a company before investing/buying a company stock...
That's why the three ETF portfolio approach is so popular. It's easy and it works!
fiber to ethernet hardware
This seems to fall under the "standard networking equipment will see a demand as data centers grow" category. Since networking equipment is a mature category, many of the large companies that provide it will see some boost but it's likely to be muted unless they are pure plays.
@@Nanalyze Thanks Nan! Gave you a shout out on the stocktwits HCP board last week for calling that buy out.
@@fafillionaire Really appreciate you helping to spread our brand. Thank you!
cloud a.i ain’t the future it’s edge a.i on device so investin in data centers is dumb
This is the sort of quality banter our readers have come to expect.
You talk too fast and show screens at the speed of light. Sorry, I am not Superman
(Makes mental note to give up meth for Lent.)
Turn down the speed control. (The wheel at the bottom right.)
ALCC seems like the next meme. As of now, just an altman backed architecture company (*spac)😂
Yes that was just approved and will begin trading this week. Perhaps we can do a video on it.