WTF: I was *NOT* Expecting THIS AI Problem.

요약 보기

대만 9월 수출이 전월비 -5.1%로 컨센서스를 크게 하회, AI 관련 수출 -18.6%, 대미 수출 -23.8%(2015년 이후 최대 감소). 증가율 4.5%로 급락. 비농업 고용은 정부 고용 78.5만 명 증가에 의존하며 민간·자영업 기준으론 약 25.8만 명 감소로 제시. 이를 근거로 경기 침체 가능성과 AI 자본지출 의존 경제에 대한 경계를 언급. 엔비디아는 PEG 1.9로 상대적으로 싸다고 평가. 다만 AI 기업 99%가 파산할 수 있다고 우려.

다듬은 글

※ 자동 자막 원문(다듬기 전)

01. hey everyone me Kevin here as we go into earnings season it's worth knowing that there can sometimes be leading data and this time we're getting leading data from artificial intelligence and well exports that are starting to miss thanks to the United States now this has happened before we have volatility in some of these numbers but the magnitude of this Miss is something we haven't seen since 2015 and it's somewhat odd because really there's there's this large narrative that right now a lot of our GDP is propped up by well government workers 785,000 of them if you read the last government uh BLS report uh and then of course artificial intelligence which uh was likely one of the reasons why we're quite frankly not in a recession yet uh who knows maybe we already are but at least not definitionally so let's take a look at this report out from Goldman Sachs which actually comes from the uh Taiwanese trade Ministry uh if you uh dive into the numbers they're definitely a lot more than meets the eye see if you look at a Taiwanese news site like Focus Taiwan you'll actually see a headline Taiwan exports hit new high in September driven by strong AI demand okay cool this is a Taiwanese paper saying hey everything's fine this is great and if you just read the headline or FR frankly you only read this story this wouldn't seem like it'd be a big issue but it's not until you actually dive a little deeper like the this Goldman Sachs research here that you find that the Taiwanese export numbers were actually way worse than expected these are actually that that sorry that these are actually pretty bad numbers now I'm not looking for things to be bearish I actually really just want to understand what's going on in the macro environment and I want to understand the trajectory if we get continued good jobs data here uh you know I want to make sure I'm positioned correctly when I'm buying stocks and this was very interesting to me so I wanted to look at this with a skeptical Outlook but look at this this is not Taiwan semi this is Taiwan the country though obviously taiwan's semiconductor producing 90% of the advanced chips in the world is basically your exporter coming out of Taiwan right so Taiwan negative surpris surprises in September exports led by drop in AI related exports specifically to the United States now that's interesting to me because it it does make me think of uh a few things that I go into this with the first thing I go into this with is I know that uh Elon is uh done with his uh Nashville Data Center and it's possible that in part because of that you have less shipments coming to Elon via you know his expanded data centers for Tesla or otherwise or even Axe and Gro uh and potentially that's slowing shipments but you know if these are in such high demand we should continue to see more shipments so you know maybe who know who knows maybe there was a supply chain issue or something but I do also go into this knowing that Amazon reduced their orders for uh additional h100 chips now their argument was that they were just waiting for Blackwell and that's sort of what's being advertised but I always like to say look if if the AI demand and the AI profit is so large you know even if the margins are slightly less using the h100 I'd rather fulfill the demand now if the demand starts slow and I actually don't need any more chips right now and you know if I want a future proof I'll just get the new Chips when they're out it's it is a sign somewhat of maybe we have a little too much AI server demand mind you I'm also going into this with the impression that 99% of AI companies will probably go bankrupt I think they're really high risk mostly because I mean I hate to say it but I just I did capitulate uh for those of you who want to know I did capitulate I did end up getting the 16 one of the reasons I didn't want the 16 is because it was going to take me forever to transfer all the stuff I have both phones here right now and it's taking me forever to transfer all these apps because I do it all manually otherwise it just sets up slowly but but sometimes by doing that I also go through and I see some of these other apps that exist and I'm looking at this right now I'm like well I have co-pilot AI that I can use uh the acrobat AI uh you know PDF reader is actually pretty decent on the phone Notions now incorporating AI though I think they use the GPT engine then of course you've got GPT then you've got Gro but then beyond that uh I I know you've got Bing as well then I found you know this other one which is I think anthropics which is cloud or that's not Cloud it's claw whatever uh and uh and to me it's like my gosh how many different AI chat Bots could I possibly need you you know what I mean like there's a so I do Wonder like how many of these companies are training chat Bots when the market doesn't even need that many chat Bots and when we go down to like an olop of of you know AI uh servers basically uh are we cutting out the rest let me clarify what I mean I like and then I want to get into this data because this AI drop is pretty intense but but consider the data for a moment here if uh or or another example of an oligopoly like what what would a good example of an olop I can't even say the word oligopoly be uh I think personally cell towers you know right now if you want different cell service youve basically got Verizon in T-Mobile and AT&T and then you might say no that's not true I use mmobile or whatever but then you realize they just use the T-Mobile bands or you know whatever uh I think the Google five ones I think they use Verizon I'm not sure but anyway you really just have three providers here and so what happens is if you're like oh I want to get into you know making cell towers it'd be crazy for you to spend the capex to get into cell towers like you're better off just doing a Ryan Reynolds and you know making a mintmobile creating your own branding to attract customers and then licensing T-Mobile bands uh and so T-Mobile gets more business and you're picking up people who aren't willing to pay T-Mobile fees I think at T-Mobile you pay maybe 35 40 bucks a month you almost get the same thing at mint for 25 or 15 a month this sounds like an ad for mint I use mint I also use T-Mobile I use both I think they're great I use Verizon I use AT&T I use all of them I hate Verizon Verizon sucks although they have better service than AT&T in my area anyway my point is of course you're going to have an oligopoly in in AI as well so I just wonder like is some of this data that we're getting here uh indicative of this beginning to happen I I I don't know you like I'm not short any AI you know I'm not trying to like pump a book here or whatever I think by now many of many of you hopefully my my my biggest fans or subscribers I appreciate you so much not necessarily the loudest people complaining at me in the comments but I just want to thank you and appreciate you know that like I want you to always know that I'm just going to give you these these facts and the data that I see even if it's unpop talking down AI right now is extremely unpopular and I'm going to get a lot of hate for it I just always like to bring a realistic Outlook I do the same thing in real estate and my other businesses you know house haack's kicking butt uh we'll have uh I think we've got a big house Haack update coming next Wednesday so stay tuned mark your calendar for that it's going to be huge but anyway uh Taiwan exports fell 5.1% month over Monon seasonally adjusted pulling back from a sharp rebound of 88.2% in the prior month so you had a really big August so it's possible this is some normalization but the outcome was even worse uh than uh was weaker than our below consensus forecast so in other words Goldman had a forecast which was below consensus and the outcome was even worse than what Goldman Sachs was predicting exports fell sharply across most major products especially AI related products this is a big red flag here this is not good AI related products fell 18.6 % out of Taiwan uh so now this is important to to note because when they show you the charts they don't just specifically show you air related so AI related products down 18.6% month-over-month seasonally adjusted I'm just going to write that down as sort of one of the little takeaways here by destination uh us uh fell 23.8% month over month seasonally adjusted largest drop since 2015 so those are two really important data point takeaways right there so far so the US getting really hit in reflection of the sharp declines in AI related exports the losses were partially out offset by gains in Japan Emerging Markets South Korea was mentioned in here but look at these numbers here uh the Bloomberg consensus for September exports was 10.9% Goldman was at 99.9% and the previous year-over-year was 16.8 well you actually got a read of just 4.5% so the numberers still positive and when you go over to the Tyran you know news source over here you you'll see they still accurately mention the positive news here on uh on on you know growth it's true it's another month of growth but the problem is not that there's growth or contraction the problem is actually that that growth rate has just collapsed well further than the collapse that was already expected you know Bloomberg was expecting roughly 11% right here down from the 16.8% last year growth rate well now it's just down to 42% this is a shockingly small number here this is not great keep in mind that as part of the exports in information communication AV products AI is lumped into that so sometimes you'll see charts on exports of it Comm and and Audi visual equipment stuff AI is only at one like one4 of that so it's it's really difficult to sort of look at charts on this kind of stuff but basically it's bad and uh it reverses about uh 3 months of modest gains in the uh AI exports and sequential momentum which is quarter to quarter was also weak for other exports after gains in the previous months with machinery and commodity related exports declining 7.8 and 7.1% respectively by destination exports to the US plunged 23.8% sharpest drop since 2015 in a reflection of the sharp declines in AI related exports exports to Europe also declined 12.4% after a strong Rebound in the prior month again maybe this is you know a way of saying hey you know things are normalizing we can't always uh go to the Moon uh and you know in alluded to that as well they had a report on this and they mentioned here that hey you know perhaps this is just a normalization after this really big beat in August and that's entirely possible but they say that you know they we're expecting some of the acceleration of demand in August to translate to September and so far it didn't look like it had so I find this interesting and to me it it it is somewhat you know it's something I want to pay attention to like I don't want to go as far as calling this concerning yet but I do think that a lot of our economy is propped up on artificial intelligence capex spend and one of the downsides of that is when you have a jobs report that tells you your household job increased by 785,000 jobs and it was all government workers uh and absent the government workers the unemployment rate would have skyrocketed uh by 4% which could have potentially triggered us going into recession you know on a definitional basis since it's possible we already are well you know then all of a sudden I get a little nervous uh because it you know I start I I do I really want to be bullish I I want everybody to know that I really I think it it should be crystal clear I want to be that person that every single day is just like by the dip let's go I love this but In fairness to those of you who watch my channel and those of you who you know come to me for information maybe you're not hearing from anywhere else I I just I have to I have to have to have to keep being as uh transparent as possible with my feelings and and I just don't feel good about where valuations are right now now don't get me wrong Nvidia is decent right now unless obviously grow collapses Nvidia is only sitting at a 1.9 Peg it's way cheaper than like a Tesla or paler which I've definitely been running uh but really if you have not looked at it yet please please please take the time look at table alpha 8 on the last BLS labor report and I really want you to pay attention to this seasonally adjusted number of government workers here all of a sudden we're up supposedly 785,000 government workers on a seasonally adjusted household number which is what the unemployment rate is based on this seems cooky dooky to me uh that uh that this this could be true private payrolls uh increased by what do we increased by on private payrolls we did have an increase in seasonally adjusted household private payrolls which is good it was about 133,000 which sounds more appropriate uh but you did also then have a decline of over 200,000 in the self-employed workers so if you're counting as employed as self-employed and let's say maybe you did get a job so you count it over here you actually had a loss of private payrolls there uh so really the way you could see that is you go non a right here and you could go okay we're at uh 159 635 minus out the 785 government workers you're at 158 850 jobs compare that to the prior month of 159 108 you actually lost 258,000 jobs in the uh you know in the combined self-employed and private sector you're minus 258,000 if it wasn't for that big $785,000 or 500 785,000 boost that we got from government workers just seasonally adjusted it already accounts for the you know like removing teachers that are just going back to work wasn't for that be nasty you know add to that the increase we're seeing in the number of people uh you know working multiple jobs or the number of unemployed 27 weeks or over which is a number that generally only Rises and rose again in the last payrolls report uh the September report was up again I mean you were uh almost 100,000 97,000 more people unemployed 27 weeks or more which means now theyve fallen off the ability to apply for unemployment insurance and they're just not considered like unemployed anymore which is crazy as that number goes up you're often usually in recession right so all these things coming together I don't think they necessarily say everything is going to collapse tomorrow but without a doubt they're a warning sign and they're a sign that I want to pay attention to anyway uh big video coming next Wednesday I hope you mark your calendar for it uh it's it's going to be a huge uh investment video uh that I'm making so a huge investment uh coming uh next Wednesday as well well as uh updates on other aspects of uh of businesses house act like a lot of basically news I've got posting on uh next Wednesday so I'm excited to talk about that anyway um thank you so much for watching if you want to get 75 free fractional shares by signing up for Weeble which is that platform that I always like to use for trading make sure you go to metkevin.com Weeble you can then draw lines just like I do uh and you'll get 75 free fractional shares just for signing up go to metkevin.com Weeble uh that is a paid promotion along with life insurance you can get slife metkevin.com slife you can Apple pay or Android pay for it