Until last weekend, i.e., pre-DeepSeek, the AI story was one built on three pieces – that the AI product/service market would be huge (trillions of dollars), that it would require a huge entry fee (in investments in super computing and at a) and that this entry fee would create an end game, where relatively few companies with pricing power would generate super-normal profits. That story has had enough force behind it to push up the market capitalizations of the companies that are actors in this story (from AI architecture companies like Nvidia, Vistra and Constellation Energy to firms in AI product and service space like Palantir) trillions of dollars.
The big story with DeepSeek is that it alters the AI story in a fundamental way, by opening up a pathway into the business without the massive investments in AI chips and data. That, in turn, has the potential of commoditizing the end market for AI products and services, with lots of competition and much less in profits. In my view, speaking as an AI novice, DeepSeek will create a bifurcated market for AI products and services, with a low-grade AI segment (composed of products that don’t need super computers and data) and a high-intensity AI segment, with very different characteristics in terms of competition and profitability.
Slides: https://pages.stern.nyu.edu/~adamodar/pdfiles/blog/DeepSeek.pdf
Blog post: https://aswathdamodaran.blogspot.com/2025/01/deepseek-crashes-ai-party-story-break.html
Valuations:
1. Nvidia in September 2024: https://pages.stern.nyu.edu/~adamodar/pc/blog/Nvidia2024.xlsx
2. Nvidia in January 2025:
https://pages.stern.nyu.edu/~adamodar/pc/blog/NvidiaJan2025.xlsx
source
Thank you professor, for the lesson in common sense and yes I found this video useful.
Imo Nvidia is not even worth 1T.
You must be crazy to still hold it at this valuation.
Have to say the separation in the run of the mill and the high end usage of AI was insightful as well as the "cute" ones. I am interested in the innovative use cases of AI that can make a small but positive change in people's lives and make them think more. Maybe a red-pill/blue-pill AI is the need of the hour.
if this is a shift, then nvida will have a larger market. the current players will still suff their datacenters with chips because they still need to train the models. but additional players will show up because they use better models now on normal consumer cards. maybe this is also a tailwind to amd because there are also competitive chips aiming more on the consumer market
AI for big 7 is already making rewards, I know Big7 companies have been able to cut head count over 1000s because AI does what those head count were doing.
17:07 deep what??😂
Jk,love you Aswath,thank you for all these wisdom gems❤
Deepseek does thing at free of cost for which openai asks 200 dollars.
25:45 :
The value of the “rest” (which includes gaming) is up about 300 B vs the September valuation , partly compensating the ~40% (or 550 b) drop in the AI business.
Would anyone know why that is the case ?
love the remarks you left at the end of the video. thanks aswadh!
"Prof. ( "Dean of Valuation"), this was such an insightful and engaging session and the "reevaluating the story part"! The way you broke down DeepSeek's potential impact on the AI landscape was eye-opening. I really liked how you framed it as 'Story Break, Change, or Shift?'—it made me think deeply about how we often overlook the nuances of disruption.
Also, the idea of 'disruption for the disruptors' was spot on. It’s a reminder that no one is immune to change, no matter how innovative they are. Thanks for making these complex topics so relatable and for always tying them back to the fundamentals. Looking forward to more of your thought-provoking analyses!"
Very uninformed. What did gpt3 to gpt4o delta did ? 100x smarter 100x cheaper. It only pushed AI more. We need smarter models working 7/24. Datacenter need is just increasing!
9:00 slide, there is something wrong with Apple data
Please keep your views on big news coming, they are very instructive.
Great video, I really liked your ending thoughts!
Thank you, very interesting, clear thinking!
HOWEVER, about the closing remarks, I have to say that i prefer excel to paper, prefer google to going to the library, prefer facebook to sending weekly letters to 300 friends …and love to gossip with each one of them daily from '000s of miles away.
Importantly, the abundance of misinformation DOES have a very positive effect: teaches people to think critically, challenge assumptions, etc. Eventually we will emerge more savvy from this, just as we did when newspapers replaced a preacher in church.
Nice way to end this presentation. AI is a double edged sword.
should include some calculation about what other things that NVdia sells will increase from this change, I'd argue that even for a single user, an nvdia card or device will be have great advantage, that Nvdia already sells AI Dev kits and is likely to create a computer to run windows/linux running ARM processors and that NVDA's eco system is known to those who are interested and as data centre sales might fall their user products would increase. As for world changing – sure it will be – just no idea how.
it takes a lot of courage to go against conventional wisdom!
Thank you so much for sharing, it's always so interesting
@AswathDamodaranonValuation
How do I apply Approach 3 using the histogram of costs of capital from the valuation sheet for a currency other than USD? Can I simply add the risk-free rate of the relevant currency?
For example, if I'm valuing an Indian company and want to avoid a detailed approach, can I just select the ninth decile for emerging markets and adjust it by adding the INR risk-free rate, given that my company falls into a high-risk category?
Cost of Capital (in INR)=Cost of Capital (in USD)−Risk-Free Rate (USD)+Risk-Free Rate (INR Currency)
What I would be missing with this approach?
did you forget you say Nvidia’s adequate maximum stock price is 40$ assuming they get 100% ai share at 2033. why are you saying 78$? value was changed?
Please make a video about palantir and what you are thinking about
Csn someone explain why his DCF has so many variables? If free cash flow is the root of intrinsic value, then why not focus on that? It’s a discounted CASH FLOW model, please explain why revenue, taxes, EBITA, etc is helpful? More inputs create more sensitivity in the model. Charlie Munger talks at length about academics applying complex models when a simpler model will suffice.
Am I stupid?
Csn someone explain why his DCF has so many variables? If free cash flow is the root of intrinsic value, then why not focus on that? It’s a discounted CASH FLOW model, please explain why revenue, taxes, EBITA, etc is helpful? More inputs create more sensitivity in the model. Charlie Munger talks at length about academics applying complex models when a simpler model will suffice.
Am I stupid or his Aswaths teaching style just not for me?
The only man I can trust for this news
This was a very long way of saying he's upset that he sold in video 2 years ago 😂
Thank you for sharing your thoughts and the video it’s very informative and helpful! However, I have a slightly different perspective on electricity consumption.
Jevons Paradox suggests that as AI becomes more efficient and affordable, overall energy consumption could actually increase rather than decrease. With AI models becoming more accessible, their usage is likely to grow exponentially, which in turn could lead to a significant rise in electricity demand.
Please, let me know your thoughts on this 🙂
Jai Hind. The DeepShits all have a total freakout. Eric Schmidt now urges the trumpeter to speed up the processing of H1-B visas to get in all the Indian AI talents.
A strong voice of reason and wisdom as usual from Prof. Damodaran. Thank you Sir!
Nvidia, AMD, TSM in robotics and healthcare software and drug design?
Lots of respect, sir.
How does that one whistleblower (RIP) factor into this?
Worth considering that certain countries that shall not be named now have access to something as powerful as the distilled DeepSeek models which do not require fancy GPUs to run, only large enough amounts of RAM.
Let's pick a number!Himm 300B$ instead of 500B$. All your analysis depends on just picking the wrong number as potato chips expert… Sorry but this is the most made up fair value analysis I have ever seen recently…
Thank you very much for your insight Aswath.