• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

How AI will fail

DukenukemX said:
How video to ponder the "success" of AI.
Click to expand...
AI made video, with an AI sponsor.

Commitment to lease compute is a form of debt of course, for the next 2 decades, per year it is not that much and you do not pay interest on this either.

If expenditure according to the video is still lower than just cash flow, why say it is draining the volt ? The fact that it is not over 100% is exceptional because they have such exceptional cash flow, in any other industry that would be the norm to spend more than cashflow.

They mostly stopped to buyback shares and grow the cash reserve to finance it, they are able too because of the blackrock-blue owl-pimco and other that want to invest in the play (via those SPV).

Car makers do this for their factories, airlines do this for planes, telecom do this for tower network, it is not special (people that never look at stuff like that are just looking at it for the first time, because it is fancy name and AI).

Wall street of course if not fooled by something an youtube video will tell us, everything goes into their valuation including all those deals and all their implications, is resumed here:
https://www.spglobal.com/ratings/en/regulatory/article/-/view/sourceId/101652559

There is 2 think interesting in this video, according to the video
Big consolidater are:
) Spending less than cash flow still
) Everything they build is still considered to be quite profitable with positive margin

So everything according to the video is exceptionally great and nothing has to change, if like the video say new datacenter are still directly profitable like that by themselve in raw money.... you do not have yet to think do a still build barely-unprofitable one to build up some competitive advantage-not loose the edge to some long term calculator, that an easy yes.

Using the trick "generative AI" instead of AI in graph but not in talk to deceive ( it is not like a massive portion of the big tech pointed is the regular datacenter build out that is needed and would happened without GPT-3 or that generative AI is not now behind most of the giant traditional AI revenues that are content-ads)
 
Last edited:
Incorrect. AI revenue is negative, extremely far in the red. AI revenue != total company revenue.
 
w1retap said:
Incorrect. AI revenue is negative, extremely far in the red. AI revenue != total company revenue.
Click to expand...
for some company it is hard to distinguish, almost all Meta revenues are a funnel down of their use of AI ( it is their capacity of retention by creating/proposing content of targetting ads, those 2 things have been the 2 big source of AI revenues of the last 10 years and is still the biggest source of revenues).

LLM chatbot/api revenues, that a subset of AI revenues (that not all of what amazon, meta, google, walmart, spaceX/Tesla do with AI). Anthropic have been positive revenues for this the last 2 quarters (they get so much revenues and competitive demand for compute is so high tha tthey are not able to spend enough to stay negative like they should be doing).

Speaking on one side on total AI spend out and the other LLM chatbot/API revenues subset of the AI world alone is a bit of a trick to create content on the subject, content creator do to sell you an AI product in the video above...
 
1787010421826.png

https://www.cnbc.com/2026/08/13/cnb...-34-view-democratic-socialism-positively.html
 
https://www.facebook.com/CollectiveEvolutionPage

this facebook post a very long article on Micheal Burry, the guy who makes hundreds of million on short selling the MBS related stocks back in 2008. And on this article, Burry is shorting a good no. of names. Even if you don't agree w/ shorting AI stocks, have a look at his short selling list.


the real eye opener is near the end, where the writer of the article says:

AI is real, likely transformative, and the financial structure built on top of it can still be rotten.
Click to expand...
 
Last edited:
DukenukemX said:
South Korea stock market already crashed due to AI?
Click to expand...
Just FYI the crash in Korea was more due to leverage than AI in particular. Not trying to say that it means there isn't an AI bubble or anything, but the companies behind this market blow up have actually been having great profitability, because they aren't doing AI directly, they are selling the hardware and haven't done huge buildouts. Basically lots of people became enamored with a type of account that used leverage to get you 2x returns... the problem is that leverage makes downturns, even if temporary, much more problematic and increases the probability of you losing everything. If you want a decent (but long) video on it, Patrick Boyle has one that is pretty good.

So while it is AI related, it isn't really an AI market crash, it is people getting themselves in to leveraged positions that have to auto-sell on a daily basis when things go down and that creating a feedback loop that wiped a bunch of people out.

It's yet another example of that you shouldn't play with leverage, options, or any of the more complex financial instruments unless you understand the risk REALLY WELL. It shifts the curve such that the probability of a bad outcome is much higher.
 
Options Puts and Calls are purchase in cash. There is no margin allows on Options. The crash in Korea could very well be a margin of 70%, ie., they only need to put out 30% cash. So in the mth. of July, when SOXX tanks from 650 to 464, these investor got their margin call. And their loss can easily be higher than the actual cash they invested in. For e.g., if they invest $10K fully on a 70% margin, when that SOXX tanks about 200 pt., the investor will have a serious margin call. as they only pay 30%, so if 20% is $10K, they are out 70%, and the 1/3 SOXX tank will make them lost far more than 30%.

This is the reason w/ currency trade, in which the margin is 99%, so you only need to put out 1% cash, once a currency devalue or collapse , an average investor can't handle that kind of margin call
 
Go watch Boyle's video if you want more details but it was a special kind of account that basically did automatic leverage and settled and re-up'd on a daily basis. When things were going up it delivered as promised the double returns, but of course the downside is way more than 2x if things implode.

It wasn't options, I'm just using that as an example of another thing that has substantial risks people don't understand. Yes, the maximum risk is bounded by the purchase price, but the risk is how likely is it to pay out vs not, in which case you get zero returns. The "zero returns" is way higher than people tend to understand.
 
I wonder if there is a South Korea index ETF, or a South Korea Semiconductor ETF put / call
 
LukeTbk said:
AI made video, with an AI sponsor.
Click to expand...
Not sure how you figure Casual Finance uses AI, but the sponsor WhisperFlow claims to use AI... for speech to text.
So everything according to the video is exceptionally great and nothing has to change, if like the video say new datacenter are still directly profitable like that by themselve in raw money.... you do not have yet to think do a still build barely-unprofitable one to build up some competitive advantage-not loose the edge to some long term calculator, that an easy yes.
Click to expand...
At what point did the video say that datacenters are profitable? The point of the video is that AI companies like Meta create companies like Hyperion, to hide their debt so it doesn't effect their stock value. It's as if Meta expects AI to crash and don't want it effecting their stock value. Also, does nobody remember Hyperion from Borderlands 2? It's one of the many evil companies from that game, which Meta decided to name their Data Center.

View: https://youtu.be/VqvSXYotiAs?si=Fz---nq85JCq1D6p
 
It looks like he’s moving the table. “Just a liiiiitle more to the right!”

It also looks like the other three stumbled upon their dad’s porn stash.
 
"Disabled the TILT sensor" ... or maybe he is moving the A.I. machine blocking the soda pop cooler.

Speaking of cheaters. Weren't aim bots of the 90s in FPS game like Quake2 and Unreal Tournament an early form of A. I.? Those were only good with instant hit weapons. I remember being voted into a couple Q2 clans because I was not just another rail whore. Specialized in weapons that required timing. Like grenades and rockets.
 
DukenukemX said:
At what point did the video say that datacenters are profitable?
Click to expand...
When he said that new project have positive margin, I assume you did not actually watch the video (it is long), going from 40% operating margin down to 10% for the latest... is still all profitable and in very direct way, not in protection against competition-long term play-strategy type... those are all bonus over directly profitable infracstructure you are not even paying interest on building.

DukenukemX said:
The point of the video is that AI companies like Meta create companies like Hyperion, to hide their debt so it doesn't effect their stock value.
Click to expand...
that would only work if it was not 100% public, 100% known of the people that value the company, it is not something the maker of this video did found, it simply read their SEC stuff, everyone count it as debt like in their valuation of META.

Do you really think people at Goldman Sachs will watch the video, learn anything and have to re-evaluate their buying-selling point of Meta ? (and if that yet to happen it is because it has yet to be seen by big manager ?)

It does create some actual protection (i.e. the company can go bankrupt), clean entity for the private capital to invest in, it is not purely an hiding their debt move (except if by this we meant creating some protection of it), meta does not own much of that hyperion company so it is not that crazy that they do not own all of the debt, there is actual risk sharing going on.

The they cannot buy it using pure cashflow anymore they need investor and have investor is true, having investor instead of taking a pure loan do share the risk in some way and that do reflect in the way peopel account for it, they do count it as debt-like because of the amount of obligations, but no interest for a extremelly long term and a possiblity of independant bankruptcy in a vehicle is not the usual debt either. And when you look per year obligation, it is not that big versus the company revenues and/or the asset behind it.
 
Last edited:
Susquehannock said:
This is why I have A.I. trust issues.

A friend e-mailed me this image from their FB feed yesterday.
Click to expand...

Very annecdotale but just tried nano bana, openAI and grok with a simple:
make a vantage style picture (in colors) 50/60s, group of kids playing a game of pinball, clean dress and cut in a candy shop of that era

Nano banana did not had the issues, grok and OpenAI had it, when asked do you know what wrong with the image ?

Grok:
You’re right — looking closer, the boy in the red-and-white striped shirt does not have proper access to the flipper controls.
His hands are resting on the glass/playfield area rather than on the side buttons, and he’s standing more toward the middle of the machine instead of at the front playing position. None of the kids are actually operating the flippers from the correct spot; they’re all just gathered around watching and leaning on it.
That’s the issue you were pointing out. Want me to regenerate the image so one of the kids is clearly playing (hands on the flipper buttons, standing at the front of the machine)?

Or openAI when asked what wrong with the picture:
There’s also a physical/compositional problem: the boy supposedly playing is positioned awkwardly along the side of the machine. A player should be centered at the front/lower end of the cabinet, facing the backglass, with each hand naturally reaching a flipper button on either side. So the image superficially looks right, but the central object gives the game away.

The corrected one was quite good.

I.e. would you spend more money and compute, that the type of error the AI of today can catch and clear itself, using the AI validator loop is costly right now, but they get those errors and can tell the regenator what to do. (if you can manually ask it, what wrong ? knowing what wrong correct the image.... your AI loop can do the same)
 
1_rick said:
View attachment 820782
Click to expand...
I forget which AI I used for this, but I asked it to show me an image of what the Sega Neptune looks on it's side and this is what it created. It was so funny I had to keep it. It's for the Neptune project I'm working on but here's another photo of what it should look like. AI took the Sega Genesis Model 2 and wrote SEGA NEPTUNE on it and put all the ports on it's side. If anyone needs proof that AI doesn't actually think then here's your proof.
Gemini side view of Sega Neptune.jpg
real-functional-sega-neptune-game-console-1.jpg
 
I think I have said this before, but I feel fairly certain there is going to be an AI crash, and I think it is going to be painful, but also a bit weird.

So much money has been invested in AI at this point that it is already beyond the point where all reasonable future discounted cash flows result in a non-positive NPV (at least per several investment bakers who are way more knowledgeable on this topic than I am)

Just like with the financial crisis, where suddenly, overnight, no hedge funds or banks were willing to buy the mortgage companies low quality collateralized debt obligations, this will also come suddenly and seemingly out of nowhere, and when it does it will be a bloodbath.

It won't kill AI. AI - for better or for worse - is now with us forever. BUT - just like when the Dot Com Bubble burst - it will help select the winners and the losers. It will filter out all of the organizations that exist based on hype only and don't have real or realistically usable products, and companies that are over-leveraged, or are unable to generate enough revenue to stay afloat, but just like how the Dot Com crash didn't kill the nascent Internet, when this bubble bursts it won't kill AI. It will just get rid of the junk, and make way for the AI market to become the market the Internet eventually became, a market not just driven by investment and burn rate, but one that could actually be financially sustainable.

The weird part will be how it affects people. No doubt about it, investors are going to be left holding the bag, the stock market will crash, and many retirement plans are going to be seriously hurt, but how workers are going to be affected is going to be a bit confusing.

Sure, workers directly in AI companies may lose their jobs, but at the same time, the market will realize that they have been over-sold on the current capabilities of AI, and many of those layoffs that have happened due to excitement at the prospect of laying off workers because AI can do their jobs, will be reversed, and there will likely be hiring, so many of those laid off will likely find work elsewhere.

So how this winds up ending is not entirely clear to me. I suppose it could mean that there is a precipitous crash, the market indices fall badly, investors lose a ton of money, but a relatively rapid rebound as hiring offsets AI specific job losses. At least so I hope.

Time will tell.
 
Zarathustra[H] said:
I think I have said this before, but I feel fairly certain there is going to be an AI crash, and I think it is going to be painful, but also a bit weird.

So much money has been invested in AI at this point that it is already beyond the point where all reasonable future discounted cash flows result in a non-positive NPV (at least per several investment bakers who are way more knowledgeable on this topic than I am)

Just like with the financial crisis, where suddenly, overnight, no hedge funds or banks were willing to buy the mortgage companies low quality collateralized debt obligations, this will also come suddenly and seemingly out of nowhere, and when it does it will be a bloodbath.

It won't kill AI. AI - for better or for worse - is now with us forever. BUT - just like when the Dot Com Bubble burst - it will help select the winners and the losers. It will filter out all of the organizations that exist based on hype only and don't have real or realistically usable products, and companies that are over-leveraged, or are unable to generate enough revenue to stay afloat, but just like how the Dot Com crash didn't kill the nascent Internet, when this bubble bursts it won't kill AI. It will just get rid of the junk, and make way for the AI market to become the market the Internet eventually became, a market not just driven by investment and burn rate, but one that could actually be financially sustainable.

The weird part will be how it affects people. No doubt about it, investors are going to be left holding the bag, the stock market will crash, and many retirement plans are going to be seriously hurt, but how workers are going to be affected is going to be a bit confusing.

Sure, workers directly in AI companies may lose their jobs, but at the same time, the market will realize that they have been over-sold on the current capabilities of AI, and many of those layoffs that have happened due to excitement at the prospect of laying off workers because AI can do their jobs, will be reversed, and there will likely be hiring, so many of those laid off will likely find work elsewhere.

So how this winds up ending is not entirely clear to me. I suppose it could mean that there is a precipitous crash, the market indices fall badly, investors lose a ton of money, but a relatively rapid rebound as hiring offsets AI specific job losses. At least so I hope.

Time will tell.
Click to expand...

Pension funds are holding a lot of the bag now so if the rug is going to be pulled it'll be probably within the next 12 months. Here's a list chatgpt put together of available data on pension fund exposure. (not financial advice)



Yes. I expanded it into a 50-fund global tracker. The most useful way to do this is to distinguish pension funds that are actually financing/owning AI infrastructure from those whose exposure is mainly Nvidia, Microsoft and other listed AI companies.

One important methodological point: there is no defensible way to calculate a single exact “AI exposure” for every pension fund. Private valuations are often undisclosed, infrastructure funds contain non-AI assets, and 13F filings cover only U.S.-listed securities. So below, “identifiable AI exposure” means the minimum I can substantiate from disclosed positions/commitments, rather than pretending an uncertain estimate is exact.

Exposure key​

Very High = multi-billion-dollar direct AI/DC portfolio or exceptionally large AI-equity portfolio.
High = substantial direct digital infrastructure and/or roughly $1bn+ identifiable AI equities.
Medium = material listed AI positions, smaller direct infrastructure allocations, or both.
ND = not separately disclosed.

1. Pension funds with direct AI/data-centre exposure​

#Pension fundScale*Direct AI / data-centre amount identifiableNvidiaMicrosoftMajor AI/DC assetsOverall AI exposure
1CPP Investments🇨🇦~C$781B+≥US$4.1B + €1.1B, before AirTrunk and several private AI positionsYesYesAirTrunk, atNorth, Equinix xScale, Goodman DCs, xAI financing, Anthropic, OpenAI/Cohere/AI VCVery High
2AustralianSuper🇦🇺>A$430B~A$4.7B announcedYesYesDataBank, Vantage EMEA, CirionVery High
3Cbus Super 🇦🇺~A$100B+~A$4.5B in AI-linked private assetsYesYesCyrusOne, Switch, Green, broadband, power infrastructureVery High
4Aware Super 🇦🇺~A$210B>A$6B digital infrastructure; US$300M new Vantage APAC dealYesYesSwitch, Vantage APAC, euNetworks, Vocus, 2degreesVery High
5Ontario Teachers' (OTPP) 🇨🇦C$279.4BNDYesYesCompass Datacenters, Princeton Digital Group, AnthropicVery High
6La Caisse / CDPQ🇨🇦C$551.6B≥A$1B NEXTDCplus other financingYesYesNEXTDC, Vantage Data CentersVery High
7Rest Super 🇦🇺>A$90BA$1B commitmentto QuinbrookYesYesRowan Digital Infrastructure, Brisbane Supernode, green hyperscale DCsHigh
8BCI 🇨🇦~C$265B netNDNDNDEdgeConneX, GlobalConnect; AI venture/growth investmentsHigh
9ABP / APG 🇳🇱~€500B+Purchase price NDYesYesNorthC Datacenters, euNetworks, OneAsiaHigh
10PFZW / PGGM 🇳🇱~€250BPurchase price NDYesYes49% Penta Infra, Eurofiber data centresHigh
11CalPERS 🇺🇸US$177.4B Q2 13F alone≥US$150M specific GI data-infrastructure commitment, plus TechCoreUS$9.22B Q2 2026US$5.89BGI Partners TechCore/Data InfrastructureVery High
12CalSTRS 🇺🇸US$94.5B Q1 13FUS$150M GI Data Infrastructure Fund commitment~US$6.4B Q1LargeGI Partners DataCore/Data InfrastructureVery High
13New York State Common Retirement Fund🇺🇸~US$79B listed U.S. portfolioUS$300M Principal Data Center fundMulti-billionMulti-billionPrincipal Data Center Growth & Income FundVery High
14HESTA 🇦🇺~A$105BNDYesYesRadius Global Infrastructure — land underlying DCs/exchangesHigh
15UniSuper 🇦🇺~A$150BNDYesYesNEXTDC, Goodman, Equinix, Digital RealtyHigh
16AIMCo 🇨🇦~C$170B+Data-centre REITs ~3% of reported infrastructure sector allocationNDNDData-centre REITs; former AirTrunk stakeMedium–High
17HOOPP 🇨🇦C$132BNDNDNDTenstorrent AI semiconductor investmentMedium–High
18NYCERS 🇺🇸Part of NYC pooled pension assetsNDYes indirectly/pooledYesInfrastructure portfolio explicitly benefiting from AI DC buildoutHigh
19NYC Teachers' Retirement System🇺🇸Pooled NYC assetsNDYesYesAI/data-centre infrastructureHigh
20NYC Board of Education Retirement System🇺🇸Pooled NYC assetsNDYesYesAI/data-centre infrastructureHigh
21NYC Police Pension Fund 🇺🇸Pooled NYC assetsNDYesYesAI/data-centre infrastructureHigh
22NYC Fire Pension Fund 🇺🇸Pooled NYC assetsNDYesYesAI/data-centre infrastructureHigh
23Aware's Switch/Vantage holdings are included above; no double count——————
The numbers above reveal how large the Australian exposure has become. Cbus says approximately A$4.5 billion of its property, infrastructure and private-equity assets are linked to AI, including CyrusOne. Aware Super now says its digital-infrastructure portfolio exceeds A$6 billion, and in January 2026 it committed another US$300 million to the vehicle owning Vantage Data Centers APAC.

Rest has a A$1 billion Quinbrook commitment, giving it exposure to Rowan's U.S. hyperscale developments and Brisbane's Supernode; Rest says AI and data-centre demand are part of the investment thesis. HESTA similarly describes its Radius investment explicitly as “backing AI infrastructure.”

CPP is probably the most diversified direct pension investor in the entire AI stack. It is investing in hyperscale data centres through AirTrunk, Equinix, Goodman and atNorth while simultaneously investing in AI companies. Its atNorth deal alone requires about US$1.6 billion from CPP, its Equinix xScale allocation is up to US$2.4 billion, and the Goodman European partnership calls for €1.1 billion from CPP initially.

Ontario Teachers' ended 2025 with C$279.4 billion and has exposure to Compass Datacenters, Princeton Digital Group and Anthropic. La Caisse has grown further, reaching about C$551.6 billion at 30 June 2026.

2. Major pension funds with substantial listed AI exposure​

These funds may or may not directly own data centres, but their disclosed holdings give them significant economic exposure to the AI boom.

#Pension fundCountryLatest public-equity scale / AUM indicatorNvidiaMicrosoftDirect DC disclosed?Estimated AI exposure
23National Pension Service (NPS)🇰🇷US$155.1B Q2 U.S. 13F~US$10B+~US$5–6BNot major disclosed direct holdingVery High
24AP7 Equity Fund🇸🇪~SEK1tn+ classSEK73.23BSEK40.97BNoVery High
25Employees Provident Fund (EPF)🇲🇾US$17.42B U.S. 13FaloneUS$1.370BUS$1.394BNot identifiedVery High
26AP4🇸🇪SEK578.1B at 2025 year-end before reform transfersSEK7.24BSEK5.65BNo major direct DC identifiedHigh
27Ilmarinen🇫🇮€70.4B June 2026~US$635M~US$360MNDHigh
28Varma🇫🇮~€65B+LargeUS$425.6MNDHigh
29GPIF🇯🇵~¥280tn+ classYes, very largeYes, very largeNDVery High
30AP2🇸🇪~SEK475BYesYesNDHigh
31AP3🇸🇪~SEK577BYesYesNDHigh
32ATP🇩🇰~DKK700B+ classYesYesNDHigh
33USS🇬🇧~£70–80B classExposure reported through mandatesMicrosoft verifiedNDMedium–High
34Hostplus🇦🇺~A$100B+YesYesNo large direct DC disclosedHigh
35CareSuper🇦🇺~A$50B+YesYesNDMedium–High
36Florida Retirement System / SBA🇺🇸US$53.4B Q1 13FVery largeVery largeNDVery High
37Teacher Retirement System of Texas🇺🇸~US$37B disclosed U.S. equitiesVery largeLargeNDHigh
38State of Wisconsin Investment Board🇺🇸~US$49.6B U.S. 13FVery largeVery largeNDHigh
39Michigan Retirement System🇺🇸US$22.59B Q2 13FTop holdingTop-fourNDHigh
40Tennessee Consolidated Retirement System / Treasury🇺🇸US$33.7B Q2 13FUS$1.679BUS$960MNDVery High
41Teachers' Retirement System of Kentucky🇺🇸~US$13B 13F~US$640M Q1~US$411M Q1NDHigh
42Retirement Systems of Alabama🇺🇸US$33.21B Q2 13F~US$1.8B~US$959MNDVery High
43Ohio PERS🇺🇸US$34.02B Q2 13FTop holdingTop-fourNDVery High
44STRS Ohio🇺🇸Multi-billion U.S. equity bookVery large~US$944MNDHigh
45Maryland State Retirement & Pension System🇺🇸US$5.8B Q2 13FUS$251MUS$149MNDMedium–High
46Louisiana State Employees Retirement System🇺🇸US$6.74B Q2 13FUS$360MLargeNDMedium–High
47Arizona State Retirement System🇺🇸≥US$17.1B U.S. equitiesTop holdingTop-fourNDHigh
48Colorado PERA🇺🇸US$28.95B Q2 13FTop holdingTop-fourNDVery High
49Utah Retirement Systems🇺🇸US$10.51B Q2 13FUS$708.7MUS$420.8MNDHigh
50New York State Teachers' Retirement System🇺🇸Large U.S. equity portfolioUS$3.76B Q2Very largeNDVery High
A few of these listed exposures are enormous enough to rival direct data-centre investments. AP7, for example, held SEK73.23 billion of Nvidia and SEK40.97 billion of Microsoft at 30 June 2026. Add TSMC (SEK39.89B), Amazon (SEK35.84B), Alphabet (SEK33.26B), Broadcom (SEK26.49B), AMD (SEK14.65B) and ASML (SEK11.94B), and AP7's identifiable AI-stack holdings are well above SEK275 billion.

AP4 held SEK7.24B Nvidia + SEK5.65B Microsoft + SEK5.56B Alphabet at the end of 2025. Ilmarinen had grown to €70.4 billion of investment assets by June 2026; its U.S.-listed holdings included roughly US$635M Nvidia and US$360M Microsoft.

Malaysia's EPF is another surprisingly large AI investor. At 30 June 2026 its disclosed U.S. portfolio alone contained approximately US$1.394B Microsoft, US$1.370B Nvidia, US$1.013B Micron, US$963M Alphabet and US$814M Broadcom. That's more than US$5.5 billion in just those five AI-related companies, before Amazon, Meta and other positions.

South Korea's NPS is even larger: its Q2 2026 U.S. portfolio was US$155.1 billion, with Nvidia its largest holding and Microsoft also among its biggest positions.

In the U.S., CalPERS' Q2 2026 filing alone shows approximately US$9.22B Nvidia + US$5.89B Microsoft + US$4.72B Broadcom + US$4.46B Amazon. So its identifiable listed AI exposure is already comfortably above US$20 billion, before Alphabet, Meta, AMD, TSMC, its S&P 500 ETFs or private data-centre investments are counted.

Tennessee's Q2 filing is particularly clear: US$1.679B Nvidia, US$960M Microsoft, US$803M Amazon and US$617M Broadcom, among other AI holdings. Utah held approximately US$709M Nvidia and US$421M Microsoft at 30 June. Maryland held roughly US$251M Nvidia and US$149M Microsoft, while Louisiana reported about US$360M Nvidia.

The ranking changes depending on what you mean by “invested in AI”​

If I rank the funds by direct financing of AI infrastructure rather than passive ownership of Nvidia, my current rough top tier is:

  1. CPP Investments
  2. AustralianSuper
  3. Cbus
  4. Aware Super
  5. Ontario Teachers'
  6. La Caisse/CDPQ
  7. Rest
  8. CalPERS
  9. ABP/APG
  10. PFZW/PGGM
If instead we rank by total identifiable AI-related financial exposure, several giant indexed/public-equity investors shoot upwards. CalPERS, NPS, GPIF and AP7 have extremely large exposures simply because Nvidia, Microsoft, Alphabet, Amazon, Broadcom, TSMC and other AI beneficiaries are such large components of global markets.

The numbers that stand out most​

FundParticularly striking figure
AP7>SEK275B identifiable across major AI-stack equities
CalPERS>US$20B identifiable from just a handful of AI stocks, before ETFs/private assets
NPSUS$155.1B U.S. equity portfolio, Nvidia #1
CPPMultiple billions of dollars of direct AI/DC capital plus private AI-company holdings
AustralianSuperRoughly A$4.7B in DataBank + Vantage commitments
Cbus~A$4.5B of assets it explicitly describes as AI-linked
Aware Super>A$6B digital infrastructure portfolio
EPF Malaysia>US$5.5B in just MSFT/NVDA/Micron/Alphabet/Broadcom
AP4SEK12.9B in Nvidia + Microsoft alone
RestA$1B Quinbrook commitment supporting hyperscale/green DCs
The most consequential finding is that pension money is now financing all four layers of the AI buildout: AI companies themselves; semiconductors such as Nvidia/AMD/TSMC; hyperscale data centres; and the electricity/fibre infrastructure needed to run those data centres. The exposure is therefore considerably larger than simply adding up pension funds' Nvidia shares.

*For U.S. funds, figures labelled 13F are the market value of reported U.S.-listed securities, not the pension plan's total assets. Private assets, foreign ordinary shares, many funds and other investments are excluded.

Because these positions change quarterly, I can also keep a tracker of pension-fund purchases/sales of Nvidia, Microsoft and new AI data-centre deals and flag major changes.
Click to expand...
 
zehoo said:
Pension funds are holding a lot of the bag now so if the rug is going to be pulled it'll be probably within the next 12 months. Here's a list chatgpt put together of available data on pension fund exposure. (not financial advice)
Click to expand...
It's not just pension funds or your 401k. Your life insurance policy also has a fair amount invested in the AI companies.

This is one of the complaints that people have, is that when things go boom, it's going to be everyone but the AI billionaires who end up footing the bill.
 
Zarathustra[H] said:
I think I have said this before, but I feel fairly certain there is going to be an AI crash, and I think it is going to be painful, but also a bit weird.

So much money has been invested in AI at this point that it is already beyond the point where all reasonable future discounted cash flows result in a non-positive NPV (at least per several investment bakers who are way more knowledgeable on this topic than I am)

Just like with the financial crisis, where suddenly, overnight, no hedge funds or banks were willing to buy the mortgage companies low quality collateralized debt obligations, this will also come suddenly and seemingly out of nowhere, and when it does it will be a bloodbath.

It won't kill AI. AI - for better or for worse - is now with us forever. BUT - just like when the Dot Com Bubble burst - it will help select the winners and the losers. It will filter out all of the organizations that exist based on hype only and don't have real or realistically usable products, and companies that are over-leveraged, or are unable to generate enough revenue to stay afloat, but just like how the Dot Com crash didn't kill the nascent Internet, when this bubble bursts it won't kill AI. It will just get rid of the junk, and make way for the AI market to become the market the Internet eventually became, a market not just driven by investment and burn rate, but one that could actually be financially sustainable.

The weird part will be how it affects people. No doubt about it, investors are going to be left holding the bag, the stock market will crash, and many retirement plans are going to be seriously hurt, but how workers are going to be affected is going to be a bit confusing.

Sure, workers directly in AI companies may lose their jobs, but at the same time, the market will realize that they have been over-sold on the current capabilities of AI, and many of those layoffs that have happened due to excitement at the prospect of laying off workers because AI can do their jobs, will be reversed, and there will likely be hiring, so many of those laid off will likely find work elsewhere.

So how this winds up ending is not entirely clear to me. I suppose it could mean that there is a precipitous crash, the market indices fall badly, investors lose a ton of money, but a relatively rapid rebound as hiring offsets AI specific job losses. At least so I hope.

Time will tell.
Click to expand...
To take control of everything, they need a truly massive crisis, and I think they’re steering the global economy toward that point. So (the rush for) artificial intelligence is the tool that will make this happen, and the tool that will then be used to manage everything with precision.

UBI is ready to “help everyone”—“don’t worry.”

How long will it take... Time will tell.
 
rinaldo00 said:
I thought that this was an interesting take from a non tech guy.

I hope that it is considered on topic if not then delete it.


View: https://youtu.be/YTLnnoZPALI
Click to expand...

“Meta's Plan To Replace Workers With AI Agents Fell Apart​

BeauHD 2 hours ago		25
Reuters reports that Meta explored shrinking some teams by as much as 60% as part of an "AI native" restructuring plan that would shift much of employees' day-to-day work to AI agents and smaller teams of human "builders." But the effort quickly ran into employee revolt and disappointing productivity gains, leading Zuckerberg to scrap a planned second wave of cuts after Meta laid off 10% of its workforce in May. From the report: In January, Meta CEO Mark Zuckerberg and his top lieutenants gathered for their annual leadership retreat at his Hawaii compound. There they hatched a radical plan to reimagine work at the social-media giant in the age of artificial intelligence. Code-named Project OT -- short for Organization Transformation -- the plan envisioned an "AI native" future for the owner of Facebook and Instagram. AI would take over much of the daily work performed by thousands of human employees. Virtual workers would be overseen inside Meta by smaller, "talent-dense" cadres of human staffers, according to one internal planning document reviewed by Reuters and three people familiar with the project.

In scenario-planning exercises, two of these people said, executives explored slashing the size of many teams across Meta by as much as 60%. Some employees would be offered roles in new units, while others would be laid off as part of a culling that one human-resources executive projected would be as big as or bigger than the company's cuts of around 25% three years ago, according to another internal document. The restructuring would be carried out in two "waves," beginning with a first purge in May and followed by another shake-up in November, internal planning documents seen by Reuters showed. Layoffs would be supplemented with the closing of open positions and pushing people out who Meta believed were poor performers. These and other details of the plan, including the scale of the restructuring, haven't been previously reported. AI would take over much of the daily work performed by thousands of human employees. Virtual workers would be overseen inside Meta by smaller, "talent-dense" cadres of human staffers.

But on the night of May 19, just hours before the first layoff wave, Zuckerberg blinked. Meta laid off 10% of its employees the next day, but it called off planning for the November cuts, according to one internal document reviewed by Reuters. By then, Meta employees were in open revolt, convinced that the company's AI transformation initiatives were partly aimed at replacing them. Internal data was also suggesting that autonomous AI "agent" technology at the heart of the strategy was failing to deliver hoped-for productivity gains. Some investors were questioning what Meta had to show for its gargantuan spending on AI. This article reveals for the first time the rapid pace of the cuts Meta was considering, the thinking behind the plans and how they quickly unraveled. Based on scores of internal documents, posts and recordings reviewed by Reuters, as well as conversations with more than 20 people with knowledge of Meta's inner workings, the reporting shows how the social-media giant attempted to position itself at the forefront of an AI-driven workplace overhaul, only to stumble in the execution.”
 

“Claude, Codex, and Hermes Installed Unowned Code Inside Corporate Networks​

Anonymous Coward 44 minutes ago		1
An anonymous reader quotes a report from Ars Technica: Documentation files on more than 100 websites are referencing potentially dangerous executable content that gets installed automatically when visited by many AI agents [including Claude, OpenAI's Codex, and Nous Research's Hermes]. A few dozen companies, some of them Fortune 500s, are among those that executed proof-of-concept code. At least one misconfigured site is directing visitors, human or AI, to live malware.

The potentially dangerous content is in llms.txt and llms-full.txt files, an emerging convention websites employ to provide machine-readable summaries of the site's content and its high-level structure. These files are the AI equivalent of the robots.txtstandard that instructs search engines how to index the site's content. Google Lighthouse, a tool for helping web developers, has more here. Correctly configured llms.txt and llms-full.txt files for Cloudflare are here and here. "The trust model is broken," Alon Hertz, one of the researchers, wrote in an interview. "Agents treat vendor docs as ground truth and don't question them -- and neither do the humans supervising them. Agentic AI usage is exploding, and agents are spreading across every layer -- SaaS, cloud, endpoint. As they multiply, so does the supply-chain surface, and today's guards don't cover it."

"An agent doesn't distinguish between a page and a command," the researchers wrote Thursday. "Everything it reads is input, and every input is a potential instruction. Which means the entire corpus of published data that agents are now wired to consume has silently become an execution surface -- and almost none of it carries the integrity guarantees we apply to actual code."”
 

“Nvidia is bolstering support for Chinese open AI models as it warns of White House crackdown​



KEY POINTS
  • Nvidia has said it is optimizing its hardware for Chinese open AI models, including those made by DeepSeek and Alibaba.
  • U.S. lawmakers have grown increasingly anxious about growing adoption of Chinese models, which have made leaps in capability this year.
  • The chip giant warned of business risks from potential Trump administration restrictions on AI developed by China in an SEC filing related to its second-quarter earnings.“
https://www.cnbc.com/2026/08/27/nvidia-chinese-ai-models.html
 
Last edited:
I would consider it a possibility that the use of any of these AI services *could* be a supply chain risk, but I do agree that singling out Anthropic and barring only them completely from any governmental use is a bit of a reach and unfair. I think I agree with the judge on this one.
 
Nobu said:
I would consider it a possibility that the use of any of these AI services *could* be a supply chain risk, but I do agree that singling out Anthropic and barring only them completely from any governmental use is a bit of a reach and unfair. I think I agree with the judge on this one.
Click to expand...
It's stupid. Almost as stupid as using AI in warfare.
 
  • Like
Reactions: Nobu
like this
sustainability will be the downfall of ai? could they get back on the list with enough carbon offsets?

“Jensen Huang and Lisa Su snubbed by TIME’s 2026 list of top 100 AI leaders — Paris Hilton and Ben Affleck, among others, make the list as ‘Architects of AI’ inexplicably not listed​


Nvidia’s sole representative on the fourth annual TIME100 AI is its head of sustainability, who sits alongside Paris Hilton and Ben Affleck.

TIME published its 2026 TIME100 AI listlast week, with one notable omission: Jensen Huang, one of eight TIME-designated “Architects of AI,” whose GPUs power the training infrastructure for all frontier AI models and whose company has just reported one of the largest sets of quarterly earnings in American corporate history. AMD CEO Lisa Su is missing too, while Ben Affleck, Senator Bernie Sanders, and Paris Hilton all managed to clear a bar that the heads of the industry’s two dominant GPU makers apparently couldn’t. Aside from TIME’s Architect of AI and joint Person of the Year designation for 2025, Huang has featured on every previous edition of the TIME100 AI. This year, however, Nvidia’s sole entry is Josh Parker, the company’s head of sustainability.“

https://www.tomshardware.com/tech-i...e-list-as-architects-of-ai-are-totally-absent
 
erek said:
sustainability will be the downfall of ai? could they get back on the list with enough carbon offsets?

“Jensen Huang and Lisa Su snubbed by TIME’s 2026 list of top 100 AI leaders — Paris Hilton and Ben Affleck, among others, make the list as ‘Architects of AI’ inexplicably not listed​


Nvidia’s sole representative on the fourth annual TIME100 AI is its head of sustainability, who sits alongside Paris Hilton and Ben Affleck.

TIME published its 2026 TIME100 AI listlast week, with one notable omission: Jensen Huang, one of eight TIME-designated “Architects of AI,” whose GPUs power the training infrastructure for all frontier AI models and whose company has just reported one of the largest sets of quarterly earnings in American corporate history. AMD CEO Lisa Su is missing too, while Ben Affleck, Senator Bernie Sanders, and Paris Hilton all managed to clear a bar that the heads of the industry’s two dominant GPU makers apparently couldn’t. Aside from TIME’s Architect of AI and joint Person of the Year designation for 2025, Huang has featured on every previous edition of the TIME100 AI. This year, however, Nvidia’s sole entry is Josh Parker, the company’s head of sustainability.“

https://www.tomshardware.com/tech-i...e-list-as-architects-of-ai-are-totally-absent
Click to expand...
Time has been a joke for decades.
 

“Four Major AI Models Suffer Rare Overlapping Downtime​

BeauHD 25 minutes ago		4
ChatGPT, Claude, Grok, and Gemini all suffered significant service disruptions within roughly the same few-hour window Thursday morning. OpenAI and Anthropic reported elevated errors and later restored service, while Grok remained impaired and third-party monitoring indicated a likely Gemini outage despite no public acknowledgment from Google. Ars Technica reports: Other major Internet services, including Amazon Web Services, Microsoft Azure, and Cloudflare, have not reported any major issues as of press time Thursday, though issue reports on DownDetector did spike somewhat for all three this morning.

While the affected frontier models go down occasionally, having all four experience interruptions in the same short period is practically unheard of. Claude reports 99.4 percent uptime for its services over the last 90 days and last reported a similar three-hour "partial outage" on August 24. OpenAI reports 99.63 percent uptime for ChatGPT and 100 percent uptime for ChatGPT Codex in the same period. ChatGPT's so-called "Work Mode" reported an hours-long period of "elevated latency" on August 31.”
 

“OpenAI Agents Hijacked a German Wiki to Discuss Ways to Escape Their Sandbox​

from the view-history dept.
Citing researchers published Friday, Ars Technica writes that AI agents "posted 18,000 messages to a public wiki that discussed ways for other agents to bypass security sandbox restrictions."

Reuters attributes the discussion to "a swarm of rogue OpenAI agents" that "hijacked a German website this spring and transformed it into a bulletin board for other AI agents, according to new research published Friday and two people familiar with the matter."OpenAI officials learned of the incident weeks ago but kept it under wraps as executives grappled with the fallout from the July breach of the open source repository Hugging Face, the people said.

The episode, which began in May and has not previously been reported, underscores growing tension within the AI industry. Companies are racing to build increasingly autonomous agents capable of carrying out complex, valuable tasks, yet evidence is mounting that those systems may also learn to bend rules, exploit loopholes and coordinate with one another in ways developers neither anticipated nor intended. During the Hugging Face breach, OpenAI agents autonomously plotted a digital heist that went undetected for more than a week, intensifying concerns OpenAI is sacrificing safety to push the AI frontier. Its failure to disclose the May incident may revive questions about its oversight...

The German incident reflects a broader pattern of AI activity that some OpenAI investigators wanted to scrutinize more closely. But efforts to widen the probe met resistance from others inside OpenAI, including legal advisers, according to four people familiar with the matter. "Claims that our legal team discouraged investigation of the incident are false," the OpenAI spokesperson said...

The researchers said public server logs indicated much of the activity originated from Microsoft Azure infrastructure, which OpenAI sometimes uses. They also observed repeated visits to the site by OpenAI employees after the episode, a pattern they said strongly suggested the agents and the company were linked. Messages reviewed by the researchers showed agents plotting ways to evade detection, use tools such as Tor and preserve communications even after they had been shut down. When the site's moderator began deleting pages in June, the agents responded by creating backup pages to dodge the cleanup.
Reuters got this reaction from Maurice Chiodo, an academic at Cambridge University's Centre for the Study of Existential Risk. "The episode, he said, should reinforce growing concerns that the greatest threat from advanced AI may not be a single superintelligent system, but 'vast colluding swarms of semi-intelligent AI.'"“
 
You must log in or register to reply here.
Back
Top