Why the Market is actually *Crashing* on DeepSeek AI

[@briancunningham5011]

Kevin is always on point.. didn't realize how major the deepseek news was last night. As the night went into morning, i realized that kev had the drop on the market

[@Bigredbulldog401]

They used the news as a perfect time to take some gains 🥱 look at spy chart from 2021 so u get a better idea of the presidential cycle 2 more months to go I already sold all my large caps couple months ago and been getting in small and mid caps and value stocks that are beaten down 😉

[@blueberrrry]

instead, last night it was widely known already and everybody put out content... model was already out last week and I already had conversation back then without any finance hype, so it's more like late on Kevin's side...

[@Bigredbulldog401]

@@blueberrrry exactly look at spy 2021 chart still got another month or 2 of pullbacks for this presidential cycle. everything is lining up including a big short covering event coming and market continues back up after 🔮

[@ericv5337]

Lol

[@JT__Media]

By “on point” he means that Kevin has Twitter 🤣

[@onlypencil]

reminds me of the iron man scene where jebediah screams 'Tony Stark was able to build this in a cave! With a bunch of scraps!'

[@popedonking]

With scraps!😂

[@markfrantz4612]

Yeah actually.. but Nvidia shouldn’t be going down lol

[@DaJrTrades]

@@markfrantz4612they literally reap the most profits, from this AI bubble.

[@vashkarpalma8165]

But the scraps are from nvda

[@bingzhao9314]

Already run it in huawei 910b

[@AndyNick211]

Bro woke up feeling like a waiter

[@jzargovalyrian8679]

He serves us juicy news

[@brianc7453]

He was waitering at a Davos after-party

[@optionmillionaire]

@@brianc7453😂

[@cclcanadiancountrylife3628]

its happening because the Buffalo bills lost last night. it was Good Vs Evil and Evil won last night.

[@SilverChaser22]

Butt hurt?

[@Slaytheday388]

Don’t worry my eagles 🦅 will smash 💥

[@liamboggs3035]

The NFL is Rigged.

[@FamilyofFour30]

Yes, I’m butt hurt big time. The Bills mean the world to us up here in upstate NY (I don’t consider the city part of NY). Couldn’t be more proud of them.

[@dankashiochosinko1080]

Buffalo 66 Moonchild

[@小霸王-h1v]

The CEO of DeepSeek graduated from Zhejiang University, and the annual tuition for his major is approximately 6,000 RMB (about 800 USD). In comparison, the annual tuition for the same major at Stanford University is 65,000 USD.

[@Jojoswhack]

Sounds like the USD is going to lose its reserve currency status in the near future

[@ricnyc2759]

OUCH!

[@iitsBrandonM]

​​@@JojoswhackBRICS

[@Joe-ti7qd]

​@@Jojoswhacknah. Not that time yet.

[@DiFeng-g2o]

In 2021, he bet big on AI-acquiring 10,000 H800 chips and bringing over his top talent from High-Flyer, experts at maximizing Nvidia GPU performance.

[@Theknight615]

Elon did reply to a tweet mentioning that Deepseek potentially has 50,000 GPUs that they are not talking about. And he said "obviously".

[@richardhanes7370]

And everyone is panicking 😂 I'm buying the whole market now

[@douglasgray3206]

But it is open source!

[@mikeramos91]

@@richardhanes7370 but is it wise to buy on the first day of a market dump?

[@vashkarpalma8165]

China lying at usual

[@garydownes6769]

How anyone know these numbers are correct? 😊

[@ARKenMan]

Nvidia is going to be fine for a long, long time. I've been in the computer industry as a professional for 35 years and I can tell you that chip demand for faster chips has never gone down and never will. Hard drive storage/speed, memory capacity/speed is all extremely important and just because the software gets more efficient or becomes open source doesn't mean anything either other than more and more applications will come online that demand more and faster hardware. This is extremely overblown and in 5 years when you look at the stock chart it will literally be a tiny blip on the radar. My 2 cents.

[@Utoko]

I agree but how about we get the premium down to a bit more sane levels?

[@Aut0KAD]

yeah, if training becomes more efficient, they aren't going to do less time training, they are just going to train and do even more.

[@MeetKevin]

This makes me nervous: "I can tell you that chip demand for faster chips has never gone down and never will"

[@jonathanczarnojan127]

@@MeetKevin isn’t that the norm for technological advancement?

[@DeanIsMean]

@@MeetKevin It's true though, at least for a long time ahead. What he didn't specify was that Nvidia are selling their chips at a massive premium due to shortage in supply while demand has been sky high. If it's true it only cost $ 5.6m to train deepseek and at 1/10th the inference cost would mean that Nvidia chips lose their premium, the demand won't be as high.

[@Oneandoneeee]

This I why competition is good for business

[@rabs1440]

Competition great for consumers

[@nicogaminglive1337]

You are completely missing the point, China is misleading on purpose about the true cost of what Deepseek cost to make.

[@pantherdddjvdgx]

China is buying the dip

[@Italianmarcus10]

Trying to undermine trumps stargate

[@i.p.9559]

@@nicogaminglive1337 i agree. Are we certain that they have the tools ?

[@timrobertson2222]

💩 DeepFake! The DeepState uses DeepSeek to attack Trump by crashing the Stockmarket based on fake news. The timing tells all!

[@hkspoon123]

Is open AI, anyone can look if you don't believe it and it get constantly improve and being benchmark tested.

[@morgansofia]

AI Stocks are pretty unstable at the moment, but if you do the right math, you should be just fine. Bloomberg and other finance media have been recording cases of folks gaining over 250k just in a matter of weeks/couple months, so I think there are a lot of wealth transfer in this downtime if you know where to look.

[@JacobsErick-u8r]

you’re right! The current market might give opportunities to maximize profit within a short term, but in order to execute such strategy , you must be a skilled practitioner

[@ThomasLuke-u3v]

I've been in touch with a financial advisor ever since I started my business. Knowing today's culture The challenge is knowing when to purchase or sell when investing in trending stocks, which is pretty simple. On my portfolio, which has grown over $900k in a little over a year, my adviser chooses entry and exit orders.

[@LUCIASMITH-d1z]

Mind if I ask you recommend this particular professional you use their service? i have quite a lot of marketing problems

[@ThomasLuke-u3v]

Melissa really seem to know her stuff. I found her website, read through her resume, educational background, qualifications and it was really impressive. her is a fiduciary who will act in my best interest. So, I booked a session with her.

[@winifred-k9e]

Thank you so much for your helpful tip! I was able to verify her on google and book a call session with her. She seems very proficient and I'm really grateful for your guidance

[@pdaniel8]

This is a tech crash, not a market crash.

[@recycledmotivation]

More like AI crasg

[@quinnh4313]

Definitely it is for now. I think the fed meeting could decide if the whole market crashes though

[@samuelanimaddo7611]

Tech is the market.

[@pdaniel8]

@@samuelanimaddo7611 Most of it, but not all, -1.5% on SP500 is not a crash.

[@Gwiner28]

AI bubble floating around some spikes

[@gurpartapsidhu3353]

The Money was spent on Sam Altman’s Keonisegg not the actual chatgpt

[@Yummr9]

@@gurpartapsidhu3353 true

[@Happycat8385]

And trips to Cambodia read he liked them young

[@FriskyDingo1983]

This too will pass. BTMFD!

[@WeBeatMedicare6969]

I figured out the acronym 👍

[@joshkarriker1057]

buy the motherfucking dip hell yeah!

[@wwt17]

Yes this is silly nothing burger. Yawns picks up codfee

[@ArlyneGoodman]

Predicting inflation starting to rise again this quarter while leading indicators showing economy slowing (not to mention governm*nt figures pumped up for the election). Global economy very weak which affects US. Fed dropping rates 0.50 shows they're VERY worried about financial downturn/crisis. interest rates coming down are also an indication banks are LESS willing to loan money into existence. The question here is where is the inflation going to come from in the near term? Consumers are mostly tapped out which is 70% of US economy (consumption). Yes inflation very likely to return but not before it continues to come down... Inflation can be a concern, but remember, certain assets WA like stocks and Crypto’s acts as a hedge. Long & short-term trading is generally safer, allowing investors to weather market volatility. I have managed to grow a nest egg of around 3.2 B'tc to a decent 27B'tc in the space of a few months... I'm especially grateful to Luna Gilman, whose deep expertise and traditional trading acumen have been invaluable in this challenging, ever-evolving financial landscape.

[@ArlyneGoodman]

She frequently interacts via Telegrams, with the user-name @LunaGilman

[@ArlyneGoodman]

She mainly interacts on Telegrams, using the user-name;

[@ArlyneGoodman]

@LunaGilman

[@KingChan-ud1ho]

One thing I know for certain is crypto is here to stay, the only thing that leaves is the people who don't manage their risk. Manage that, or the market will manage it for you. With the right strategies you will survive.

[@lburg99]

Nice info, i appreciate your concern this will help a lot especially to the young bitcoin investors who have no or lesser knowledge on how bitcoin market works.

[@innorthtex]

The OSRS reference to start the video is wild 😂

[@xSTstSTx]

@@innorthtex Why, because you still play it?

[@innorthtex]

@xSTstSTx no, do you?

[@OG_Tybern]

LOL

[@FINAL-B0SS]

Kevin I think you are missing what DeepSeek’s “break through” means. It means cheaper and faster AI gains. This is still a race. Big tech can’t stop or they will get passed up. The amount of compute you have still equates to your capabilities and dominance. We’ve been running DeepSeek since it dropped, it’s a minor evolution and it’s a lot cheaper but make no mistake some of this is propaganda out of china

[@SmartestDumbGuy]

How is it propaganda? It's almost equal and cheaper. That's what you just said.

[@Joe-ti7qd]

​@@SmartestDumbGuyfair point.

[@FINAL-B0SS]

@ saying they did this with a tiny team, on a shoe string budget and for only $5 million is the propaganda. Reality is it took a $4-$5 billion data center and required OpenAI to be where it is to train on their data. The real story is cost reduction but that is only going to mean more AI chips needed. I’m excited about DeepSeek but it isn’t revolutionary.

[@WhiteWizzard]

@@SmartestDumbGuy it’s free!

[@Happycat8385]

4-5 ,billion is still massively cheaper than what are companies have been spending. Havnt they pissed through hundreds of billions??

[@geraldsmith7240]

Again Kevin, A Flawless, Meaningful Explanation 🌹

[@limemason]

"He invested, go ahead and tank the stock" Every. Single. Time.

[@BeefNEggs057]

Sure feels that way sometimes. I recommend considering just hanging on if you have time.

[@Frankjacob-q5s]

Hit 200k today. I'm really grateful for all the knowledge and nuggets you had thrown my way over the last months. Started with 14k in June 2024

[@WesRoth]

5:56 Elon made a few replies suggesting he supports the idea that Chinese companies have a lot of NVIDIA chips we don't know about. 100% agree that his XAI is also on the cross hairs and we will see some statements by him soon (maybe Tesla earnings call coming up?)

[@TiagoRamosVideos]

Great video 👌 thanks 🙏

[@ClaymorePT]

US Tech sector is no longer efficient at developing advanced technology. And I'm not talking about paychecks. I'm talking about being efficient at their jobs. Too many managers, not enough engineers. Too many CEOs and CTOs and $hitOOs...

[@saquibhussain6000]

This is where the truth is. US tech is all about wall street, money, and too many people with V and C in front of their titles. Not to mention overpaid software engineers

[@iMixMaSteR1]

And China are nothing but plagiarists with zero creativity

[@TonyVear-z8y]

I used to think it was just about buying dip in market, but I didn’t realize there are strategies for managing risk and actually making a profit. Now I feel kinda stuck since I’m not seeing any gains in my portfolio. Is there any recommendations on what I should consider? I’d really appreciate it!

[@Jakepattison42]

It's not only about having funds to invest in the market; it's also about knowing the right strategies to make a profit and manage risk. That's why working with an investment advisor is important-they have insights that aren't available to everyone.

[@Scottydale5]

I totally agree with you. I started out investing on my own too and lost quite a bit. After the 2020 crash, I managed to pull out about $160k. I then invested that money with an analyst, and in just seven months, I made almost $580,000. It's amazing how having the right guidance can turn things around!

[@Speak2susan]

I'm intrigued by this. I've searched for financial advisors online but it's kind of hard to get in touch with one. Okay if I ask you for a recommendation?

[@Scottydale5]

Annette Christine Conte is the advisor I use and i'm just putting this out here because you asked. You can Just search the name. You’d find necessary details to work with to set up an appointment.

[@Speak2susan]

I looked up her full name online and found her page. I emailed and made an appointment to talk with her; hopefully, she gets back to me

[@LarryLaRue2022]

EVERYONE screaming about TikTok being owned by the Chinese but we're gonna trust Chinese AI? WTAF?

[@Theknight615]

The H100 "secret" GPUs story is not just about the quality of chips that Deepseek has, it is also about the quantity. The company could be using as much as 50 thousand GPUs while they are claiming only 2000.

[@墨视天下]

SB

[@AbsurdImprobability]

Market is overreacting. It's an open source model which means all the AI companies will learn and improve their models from it. They also have more compute and more compute at inference time gives better results. This should be good news, so I'd say it's a buying opportunity.

[@SigFigNewton]

Spoken as tho all companies benefiting from ai boom are the same? The fact that ai model makers will incorporate these efficiencies if they’re real is *why* equipment companies are selling off by so much. Yes, agreed, this wouldn’t be bad for the Microsoft!etaAlphabet buyers of the equipment. But it certainly would for the equipment sellers. The distinction is already being made by the market

[@dialy1]

Not good news for NVIDIA. If true…

[@zaevma]

Great news for NVIDIA. This does not effect NVIDIA in a negative way. More efficiency in AI models will mean more demand. AI models are not getting smaller...they are getting larger and larger, there has to be efficiency changes in software (as expected) for AI models to scale out and not solely rely on hardware. Put simply, Jevons Paradox. More efficiency = increased demand. We are not even close to hitting the cap for compute for AI models.

[@madoksback]

One problem with your thesis on AI is that the latest generation of AI models take advantage of models availability to think and the more time it takes to think the more compute is required so the chip demand is more likely to increase So the shift has moved from training to runtime compute

[@VperVendetta1992]

Exactly, and Nvidia will swiftly shift towards inference optimized cards. AMD already started, and other startups like Grok (not X's Grok) are doing the same.

[@gregking267]

Based on my understanding, the code that specifically runs Deepseek-R1 is NOT publicly available. (If you find it, please link it because I would like to read it) We have only some insight into the RL training method they used on their base model (Deepseek-V3 which does have the code publicly available) from the academic paper they provided with the model. Speaking as someone who has some practical experience in this field, I would also say that it is possible that they are getting major efficiency gains from training the model from other sources, creating fine-tuned data sets off of the existing models from other companies, putting more human capital into the equation via supervised learning, etc. There are a TON of other ways for this performance to appear better, without actually having more utility. Given the nature of how LLMs work, the "benchmarks" we are collectively using to compare these models are suspect at best because the real-world performance depends on the task provided to it. This is NOT to discount the fact that the process that Deepseek-R1 used to train is really interesting, because it is. It is mostly to state, that we DON'T actually have all the facts, and the media has been way too quick to jump to conclusions on this in my opinion.

[@CaliLifeTCG]

Buying the dip buying the dip lol song 🎧

[@rioriggs3568]

Chinese consumers just realize why should we buy Tesla when we lead AI and our cars are half the price!

[@MokeAnit]

​​@@rioriggs3568 Because they don't lead in AI. Not even close

[@Realworlddummy]

they dont lead ai, its the cost. consumer mindset is too ingrained in us we are in denial

[@imnotanalien7839]

California is asking homeowners to rent out rooms to fire victims. However, due to squatters rights people are very reluctant. Since you are a Ca. resident and in real estate, this might be a helpful subject to discuss. California needs to change the laws pertaining to squatters rights and start protecting homeowners. And these laws need to be changed immediately, people need help on both sides of this issue. I think you would be the perfect person to lead the charge on this.

[@wenjuntan1717]

@@imnotanalien7839 can’t agree with you more

[@Joe-ti7qd]

Bro...in CA, i let a girl spend the night and she went crazy. Cops wouldn't remove her. Said I had to go through eviction. These ladies see a nice house with a bachelor and just want to take it over. Scary. I'm done dating too.

[@BeefNEggs057]

F em. Make them change the anti renter biased laws first.

[@rayr6278]

After his gubernatorial run agianst Newsome, I'm sure Cali leadership will do everything in their power to suppress or sideline Kev in anything pertaining to politics or legislation.

[@shldngyl8724]

DeepSeek DEPENDS HEAVILY on Nvidia Ecosystem 🐳👾💻

[@theSatanicNaturalists.]

Who told you that?

[@shldngyl8724]

@The DeepSeek app

[@vashkarpalma8165]

​@@theSatanicNaturalists.There is no other gpu

[@daithomas9859]

This just uncovers the greed of major American companies.

[@BeefNEggs057]

But not you - right….

[@Happycat8385]

​tard

[@XunMoon]

Thank you Kevin for the informative video!😊

[@quinnh4313]

Elon Musk actually did comment on Deepseek agreeing with a post that said there was no way Deepseek only used the GPUs they were claiming and that they said they used only those chips because of the export rules

[@Happycat8385]

Cope

[@OG_Tybern]

I feel like this is an overreaction because AI has basically unlimited applications other than Large Language Models (LLMs) and even if it takes less chips for a certain AI application that does not take away from the demand for chips for all the other applications of AI.

[@bmcga1]

But it's opensource. Whatever secrect sauce they have is going to be incorporated into the US models shortly. If DeepSeek really is that much more efficient, it'll run *that* much faster on nvidia hardware...

[@SigFigNewton]

The recent immense spend from meta,Microsoft,alphabet etc wouldn’t even have to go down. If their spend on it simply plateaus due to the models being far easier to train, that alone would be baaaaad news for the equipment sellers.

[@SigFigNewton]

Given valuations priced for rapid growth. So I see your point as being bullish equipment buyers bearish equipment sellers

[@Happycat8385]

It's free, though, right? Doesn't that mess up these companies' plans to maximize profits????

[@bmcga1]

@ Well yeah but I expect they'll reverse engineer it.

[@susans1320]

Love your Vibe today- Cuff links ROCK!!

[@skylerNope]

Theoretically...doesn't this just mean the big companies like open AI could just utilize these methods and accelerate progress with their mega clusters?

[@SigFigNewton]

The reason it may be bad for equipment sellers is that yes, the other ai builders could copy the technique and we don’t know whether they would still buy as much equipment. The reason it may be bad for ai builders is that the cost reduction could let others into the game. If the company no longer has to be megacap to afford to train the model, increased competition decreases what they can charge for use of their models. Which takes us back to the first point. 20x cheaper ai model might not mean same amount of equipment spend but 20x more compute into the training process because of reduced expectations for value of completed model.

[@skylerNope]

@SigFigNewton fair...but this is still a kind of arms race correct? All gas no brakes?..but yes bad for the big business side possibly...but good in terms of the future of this tech as a whole...I think.

[@SigFigNewton]

Maybe. I am unwilling to bet that brakes won’t be pumped.

[@SigFigNewton]

But what do I know I missed out on nvda completely

[@skylerNope]

@SigFigNewton I understand..and I'm not saying nobody will hit brakes..from an investor perspective it's rough...but personally I'd put some $ in tesla because this theoretically pushes grok and optimus bot to another level..I think at least.which is exactly what he'd want to reach mars asap

[@vulian]

Very informative. Thank you.

[@malungisamndzebele2770]

Kevin, Apple is only down 1% so far today because it is already 11% down this year.

[@Aterallus]

Awesome video, very insightful. Thank you Kevin.

[@taz_trader]

Kevin is that a warglaive in the background, and is that a RuneScape mug 😂 I’m new here

[@kraz007]

There's no "code" in an ai model. There's weights, so looking at it won't reveal anything. Open source means its free to use in other products and you can adjust weights or add guard rails.

[@SigFigNewton]

If there’s transparency in their process and they did do it with far less compute, other ai builders suddenly plan to buy less chip than previously anticipated

[@VperVendetta1992]

The code is the training strategy and the training dataset, not just the weights. The training strategy is the most valuable thing here, and other companies will copy what the Chinese did to make their models better and more efficient.

[@raibalama]

This is a clear strategy to buy the dip from scared investors

[@mtownlonewolf7007]

Kevin your wrong, like when they made the steam more efficient, they thought coal would get cheaper. We'll more people came in and bought engines and coal went up.

[@SigFigNewton]

Do they have enough data to bother with a 20x increase in ability to put more data into the ai

[@coleshores]

@@SigFigNewton Its not about training, its about test time compute. Allowing the AI to really think through a question as long as it deems necessary to provide the correct solution. This is how o3 crushed the ARC-AGI test. OpenAI spent $1.7 for inference time compute to solve one problem. If we are to have PHD level AI assistants, we ware going to need to reduce the costs to provide the necessary inference at scale.

[@SigFigNewton]

Cool so it reduces energy requirements too

[@Wibb14]

You’re putting a lot of faith in the Chinese with your assumptions.

[@brejaimecastillo8851]

CLEARANCE ON ISLE AI 25%-35% DISCOUNT

[@justSTUMBLEDupon]

2:34 The only thing about China is that they are mostly allowed to take I.P. trademarks and copyright so it might be something they took, but since this is open source, it’s not as likely it’s stolen. Keep in mind China is very good at training its computers to know what the people are doing (like social engineering). So Ai falls right into their lane of training.

[@brendonthomas2555]

Nice piece, Kevin. Enjoyed your musings. Looking extremely sharp in that dinner jacket as well.

[@Socrates526]

Kevin, why are you looking so dapper this morning?

[@SuMMeRFLi5]

To melt our hearts❤‍🔥

[@josho1778]

He's meeting your mother for lunch 😅

[@briancunningham5011]

The question is, will nvda climb back to the top, or did deepseek hit them with a fade..?

[@duboisalbert1]

Nvidia will expand upward tomorrow. Minimum 2-5%. Just a gut feeling not to be taken as investment advice 😂

[@jordanfilms2334]

Question is, why would anyone ever take this at face value?

[@AnuragShrivastav-7058]

I started using deepseek a few days ago. Unless open ai and others come up with a nuclear reactor powered quantum computer AGI i am using deepseek. not going back anytime soon

[@PepeFuego369]

Meet Kevin king of courses

[@SigFigNewton]

Course King Kev.

[@FINAL-B0SS]

Read the white paper. They used a multi-billion dollar infrastructure to get DeepSeek up and running and it was subsidized because it trained in the back of OpenAI. It’s just the next hot thing. There will be a new hot thing next month. I like to lower inference cost but this will just drive more usage which means more demand for Nvidia.

[@benpielstick]

Demand is limitless. Efficiency only leads to acceleration in a terminal race condition where the value of achieving Artificial General Intelligence is absolute.

[@darin7369]

This will benefit American companies more than China. The reason why I say that is China has a limited stockpile of H100 GPUs and are very limited in their ability to scale. American companies will tear down Deepseek and improve their own technology and then scale it.

[@thestonksmarket1897]

“Limited stockpile” but the point is China is able to make an LLM to rival America’s while spending 5-10% what America does… they only need a fraction of the chips, and you bet the Chinese government and certain companies can get their hands on them.

[@Blest501]

That coffee sip after so much preamble of the coffee sip... I _felt_ that

[@westynils]

I felt the turtlehead poking out

[@cathychats]

DeepSeek feels like the TikTok of AI. Great on the outside, corrupt on the inside.

[@ClaymorePT]

Keep thinking that.

[@cathychats]

@@ClaymorePT I will LOL

[@TRUEjuannonly]

Nah … your statement is the results of American media persuasiveness

[@healer1000040]

yeah console yourselves

[@katum-t5t]

they really collect metadata, its a fact

[@thestonksmarket1897]

LFG! For us diamond handed investors, a multi-pronged macro crash is a GIFT :)

[@wutan234]

DEEPSEEK was released a few weeks ago. Everyone was still pumping on AI and chips, especially with the Stargate announcement. Good old pumping and dump. Let's see if NVDA gets close to 100. 🎉

[@MVPTC]

AI BUBBLE BURST!!!

[@Noobdude534]

Big Kev is the real life Dr.Who. Love the bow tie. ❤

[@Rockall57]

007...... 🔫

[@melissaweyrick5311]

Add to it, farm workers staying home and food rotting in the fields. Now, what do you think Will happen to the economy.

[@Onthe545]

MSFT is in trouble here too. They raised prices on office products by over 20% with promise of AI features via copilot. Companies will not want to pay that premium now. Also, MSFT already paid for all the high end NVDA chips and data centers.

[@Beetalfly]

Very good!

[@violator2695]

They have leveraged the existing models available and optimised them to an extent. If they were building from scratch the sunk cost would be greater. I think the optimisation, and what they have done is a useful addition to progress AI sector as a whole. Evidently, the cost illustrated is not apples with apples, but it might well highlight that those spending billions on capex may not have the original moat they once thought they did have, and how do they protect their IP. I think the pull back in NVDA was much needed, and healthy. Big day with some earnings being reported, and will companies be penalised for spending on AI.

[@jaym9846]

This is a temporary crash, right?

[@wllmshrwn]

I think the ROI from artificial intelligence will never materialize for various reasons one being valuations. There was never justification for big tech valuations because A.I. is never going to be anything more then automated software. People have to want A.I. like we needed Microsoft to operate personal computers. It boils down to A.I. works great for companies to downsize operations but it's not PlayStation or Grand Theft Auto 6.

[@NewWaveSports229]

The video crashed when watching the video because an add came on. This is supreme American technology😂

[@Thedaytradingtraveler]

Love the runescape reference right out the gate!

[@Kms7171]

Remember when you used to be a good and kind KZbinr? We miss those days

[@eugenes9751]

Tesla's FSD is EXACTLY the Deepseek situation. Nvidia will release a virtual training simulator that'll be equivalent to real world data, and their moat is gone overnight.

[@willskaggs15]

The problem with American spending on AI is not technology or infrastructure, it is the human cost. C-Suite earning way too much without proven efficacy in their products and solutions. America has a massive greed problem. Running from AI to the basics, like the cost of food.

[@jordanshurtleff1735]

Loved then intro❤

[@AstroSquid]

If things become more efficient people will just want more, they always want more faster.

[@b3ndotch]

Optimus will have its DeepSeek moment as well soon.

[@SigFigNewton]

If so that’s bad news for nvda

[@saquibhussain6000]

Great thoughtful rationalization! And also let's not forget, newer models can (and should) focus to be even more efficient models. Then we prioritize on how to apply them to wide scale adoption and the model yhemlevs being essentially free and trying to be more efficient. This is very good for society and overall for consumers, we need these efficiency. I know it sucks to be an AI engineer smoking pipes and wanting million dollar / yr, but China showed us there is a better way to do it that helps us all. This is a great day and something we needed.

[@LeighB420]

Oh no, China made the same thing as us but cheaper, faster and better? How the hell did we let that..........oh wait

[@ragnarlothbrok6240]

American tech companies have been overly focused on keeping AI politically correct rather than useful.

[@colddogs]

thanks trump

[@miken5261]

More Fearmonger form Kevin

[@RealEltonChang]

1. Infinite liquidity doesn’t fix a solvency issue. 2. Inflationary monetary policy can’t work against technological deflation.

[@TerryMack-co7lt]

Kevin why would China open source their AI tool if they are doing so much better than everyone else? When companies like Open AI are going closed source and China is a competitor and political rival why would they give out this information and not keep it private. What is the motivation?

[@ScottClevelandmi]

We get punished for using less recourses. This world is so backwards.

[@seapr6]

Kevin cleans up nice.

[@tinates7354]

Kevin,you look great with Suits . Keep it up Kevin.

[@tryintomine4363]

It took people here long enough to see it. It’s been out for a while.

[@jacebabcock2526]

Nice input thanks

[@Fadi-hl8ox]

Happy birthday Kevin

[@TristanHurlebaus]

I agree with everything you said that company are going to make the most efficient models possible. But once they are efficient they will go back to increasing compute power to increase the intelligence of AI. The linear slope of compute power to overall knowledge will still hold. I could be wrong though so take it with a grain of salt

[@Richardkowalski60]

I would be very concerned about the possibility that China has a much more advanced AI. There is no way in hell that China would allow Deepseek to be given out for free if they didn’t have something superior.

[@etfeasy]

XRMI ETF buys protective puts to mitigate the risks of a major market selloff.

[@Tay_Singer]

Love the Runescape reference 😂

[@lalalashundra]

I was waiting on this video

[@DiFeng-g2o]

Why is the market responding now when Deepseek was released three months ago?

[@markschuette2615]

because of the trump anouncement i would guess. was wondering same thing

[@SigFigNewton]

Is the news that it used less equipment to train the model new?

[@SigFigNewton]

Cuz that’s all that matters

[@SmartestDumbGuy]

Nobody has heard of deepsink until just now.

[@DiFeng-g2o]

@@SigFigNewton I think the reason is AGI is a long way to go, let us make sure fund is more fairly distributed.

[@MarkAble8]

Zuckerberg scrambling to find a response.

[@jaymata1218]

They released it on Christmas day. Released the model on Jan 20th.. afterwards the paper came out on 22nd... they have a visual model, 3d model. It's insane seeing people thinking the Chinese even care about what's going on here. If this is the best their private sector has... imagine state sponsored.

[@HollywoodCO_Main]

2025 Drinking Game: Take a shot for everytime you hear Trump say TARIFF 😂😂😂

[@marc31415]

I think the reason why NVIDIA Crashed has to do more with recent Government sanctions, restricting AI Sales in countries that pose a threat to national security (like China)