Learning Library

← Back to Library

DeepSeek’s Market‑Disrupting Copycat Strategy

Key Points

  • DeepSeek’s core strategy is to flood the market with free, near‑identical copies of OpenAI’s entire product stack—including voice, coding, and upcoming video tools—to rapidly steal consumer and developer share.
  • By offering these services at no cost, DeepSeek expects millions of first‑time users to adopt its alternatives, using a relentless “drip” of media coverage to build mindshare and cement its position.
  • The surge in DeepSeek activity forces Nvidia and other chip manufacturers to keep investing heavily in AI hardware, driven by game‑theoretic pressure to stay competitive.
  • Meta’s Mark Zuckerberg has responded with a $60‑$65 billion AI budget, building a half‑Manhattan‑sized data center and pursuing an autonomous software‑engineer to counter DeepSeek’s advances.
  • The overarching incentive for all these massive expenditures is the high‑valued “prize” of achieving artificial general and superintelligence, which keeps spending cycles accelerating.

Full Transcript

# DeepSeek’s Market‑Disrupting Copycat Strategy **Source:** [https://www.youtube.com/watch?v=tg7Ajm9TYE8](https://www.youtube.com/watch?v=tg7Ajm9TYE8) **Duration:** 00:05:49 ## Summary - DeepSeek’s core strategy is to flood the market with free, near‑identical copies of OpenAI’s entire product stack—including voice, coding, and upcoming video tools—to rapidly steal consumer and developer share. - By offering these services at no cost, DeepSeek expects millions of first‑time users to adopt its alternatives, using a relentless “drip” of media coverage to build mindshare and cement its position. - The surge in DeepSeek activity forces Nvidia and other chip manufacturers to keep investing heavily in AI hardware, driven by game‑theoretic pressure to stay competitive. - Meta’s Mark Zuckerberg has responded with a $60‑$65 billion AI budget, building a half‑Manhattan‑sized data center and pursuing an autonomous software‑engineer to counter DeepSeek’s advances. - The overarching incentive for all these massive expenditures is the high‑valued “prize” of achieving artificial general and superintelligence, which keeps spending cycles accelerating. ## Sections - [00:00:00](https://www.youtube.com/watch?v=tg7Ajm9TYE8&t=0s) **DeepSeek's Aggressive Copycat Play** - The speaker predicts DeepSeek will flood the market with free, near‑identical versions of OpenAI’s offerings—voice, coding, and future video tools—to wrest market share from OpenAI and Anthropic, even at the cost of heavy cash burn. ## Full Transcript
0:00here's my prediction of what happens 0:01next on deep seek on Nvidia on this 0:05whole situation where we have a 0:06massively efficient model disrupting the 0:09market number one I expect as a 0:12strategic play deep seek to attempt to 0:16bring 0:17free very close to copy products to 0:21market across the entire open AI product 0:24surface area and I don't say close to 0:26copy because I'm saying they're just 0:27copying everything they do and they all 0:29they can do is copycat this is an 0:30Innovative team they've brought 0:32efficient things to Market I say copy 0:35because their deliberate strategy is to 0:37displace the market share that open AI 0:40has established with consumers and 0:42developers that's the play also with 0:46developers versus anthropic because 0:48Sonet is absolutely threatened by coding 0:51with deep 0:52seek 0:53so the play is bring operator to Market 0:58an operator clone whatever deep seek has 1:01that is like operator the play is bring 1:04a voice clone to Market where you have 1:06like advanced voice mode and these are 1:08all by the way going to be free and 1:11because they are free there are going to 1:12be millions of people who try something 1:14like operator for the first time not 1:16with open AI but with deep seek that is 1:21brilliant they are absolutely burning 1:23cash to do stuff like this and it is 1:25worth it to them if they can gain market 1:28share back 1:30they're going to release a video tool 1:33probably something that is designed to 1:34compete with Sora that may take them a 1:37little bit longer but I guarantee you 1:38they're working on it so they're going 1:40to go across the whole product Suite it 1:43will be relentless and now that the 1:45media has a narrative that deep seek 1:48driving releases means bad things for 1:51American companies there will just be 1:54this constant drip of news about it and 1:56deep seek is smart they will keep that 1:58constant drip of news going it helps 2:00them to gain mind share it helps them to 2:02gain market 2:03share okay so what happens to Nvidia 2:05what happens to the other players they 2:08will keep spending on chips and you know 2:11why because of Game Theory and that's 2:15not just me saying that by the way if 2:16you think oh Nate is sitting here saying 2:17oh it's Game Theory no Mark Zuckerberg 2:21has said I think about Game Theory all 2:22the time with AI it's not just me and 2:25Mark Zuckerberg by the way has a lot 2:27more billions than I do to spend on this 2:30Mark Zuckerberg announced 60 to 65 2:32billion in net AI investment for this 2:35year his goal is to have an autonomous 2:37software engineer he's building a data 2:39center that is half the size of 2:40Manhattan all of that was announced 2:43after deep seek do you think he didn't 2:45know about it he did know about it and 2:48the fact that he has apparently 2:51reportedly four different continuously 2:54running War rooms to go after like what 2:57deep seek dig did and how to bring it 2:59into llama that's fine that's just Zuck 3:02in founder mode he'll figure it out he's 3:04not going to spend less and the reason 3:06why he won't spend less is very very 3:08simple the prize value that these 3:13companies ascribe to artificial general 3:16intelligence and after that artificial 3:18super intelligence is so high they will 3:20pay anything just to be in the game so 3:23it is rational to spend more and more 3:27and more so that you stay in the game 3:30because all of your competitors are 3:33spending more what happens if you are 3:36the person who says Ah deep seek lets me 3:38spend less and you wind back your 3:41investment from $60 billion to20 billion 3:44and your competitors even one of them 3:47says I'm taking the other side of that 3:49bet I'm going to spend more and they 3:53win catastrophic for that company met 3:56meta immediately out of the AI race they 3:59lose out on from these CEOs perspective 4:02potentially trillions of dollars in 4:04gains they will not risk that kind of 4:08disaster they will spend more to make 4:10sure they are in the game they will 4:12overspend to make sure they are in the 4:14game and Nvidia will continue to benefit 4:18and if you think only American CEOs are 4:20playing game theory I have a I have news 4:22for you there is a reason that China has 4:26announced a 4:28128 billion 4:30package for AI if they really just 4:33trained this entire model on $5.5 4:36million you'd think why would they 4:39invest another $128 billion or $38 4:41billion the numbers vary on 4:44AI well because they also want to be in 4:47the game and they will spend a lot to 4:49get 4:50there it's pretty simple and so I think 4:54game theory has not gone away the 4:57advantages of artificial general 4:58intelligence have not gone away and the 5:01incentives all of these players face 5:02have not gone away and neither has the 5:05strategy that deep seek has to gain 5:08market share and it's a great strategy 5:10by the way and so I would expect that we 5:13are going to see more deep seek releases 5:15very soon I expect them to be free I 5:17expect them to be eye-opening and pretty 5:19good and I think they're going to put 5:21pressure on the business models that 5:23especially open AI has adopted where 5:25they're able to go to Consumers and get 5:26Mass payments from consumers I don't 5:29know if that sustainable if deep seek is 5:32willing to just give them free compute 5:34for a while someone's subsidizing that 5:37compute but they're giving it away for 5:38free for now to gain market share and it 5:39is working so we're going to have to 5:43see we're going to have to see what do 5:46you think is going to happen next