Get started

Deepseek

NVDA's Historical Bounce Data - This Is The Way.Forget all the nonsense about deepseek and evidence surrounding the NVDA chinese financial psyop that crushed the market today (primarily before the market even opened). It's all smoke and mirrors. Putting your money in the middle when the odds are in your favor is how you come out on top. That being said, there's a 90% rebound rate for NVidia's 10 biggest drops over the last - hence the reason I'm sitting on 75 calls with a strike of 125 that expire this Friday. The average next day rebound is 4.4% with the median being 5.3%. NVDA closed at $118.58, meaning there's a 90% chance that tomorrow we will see the price settle in the following ranges: Bull Case: 60% probability: $124.50 - $126.90 Base Case: 30% probability: $120.95 - $123.30 Bear Case: 10% probability: $115 - $117 The DeepSeek Red Herring: Speculating on the DeepSeek nonsense, the release of DeepSeek's R1 seems like an attempt to make the narrative fit the story rather than anything based on actual news: *As someone that works w/ AI every day, DeepSeek v3 has been out for a long time, and R1 was released over a week ago. There isn't anything new about this story. *This likely points to a coordinated dump of NVDA by 'whales' during premarket hours to push price action, and China has enough sway in the US markets to perform such a sway after hours. More than 12% of the 16.9% drop occurred in a short period before the market opened - limiting the influence/access of retail investors and thereby maximizing their leverage/power over the market. *This could be a preemptive move by China in a financial cold war that has been developing. Trump recently touted investing $.5T in stargate (ai), and has proposed tariffs of 10% on all chinese goods starting in just 4 days (Feb 1st) *NVDA is the perfect target to send a message. Most of their production is in Taiwan, and we know how China feels about that. The fact that China can't purchase their super chips is a big slap in the face. It'd be like China growing a bunch of crops in Idaho, only to not sell any food to the US while the US is starving. *It's a known fact that bots place the majority of trades on the US market these days. China is a master at reverse engineering tech (if not outright stealing it). Knowing what triggers market bots would be easier than supplying a fake narrative. Nothing about DeepSeek being the reason for the drop passes the smell test if for no other reason than from a logical standpoint...a couple If/Then scenarios: 1) If Deepseek did develop a model for $6M (which would be both insane and extremely unlikely) using outdated tech - Then NVDA's response that they should have their export restrictions removed and the 2nd largest AI market open to them is legitimate. Sales would skyrocket. 2) If this is Chinese misinformation and they're lying about using the A100 chips or the development costs, then why would they do that? 3) If China can't develop their own model without the A100s, what would they do to gain access to them? Then I think they steal the model - either the o1 (openAI) or llama (meta) model and tinkered with it just enough to optimize it as it's performance results are almost identical to openAI's o1 model - DeepSeek's Founder admits "there are no secrets in AI". While models can run on outdated hardware, you can't develop new models in a timely fashion on anything other than the A100s because they're 20x more powerful than the previous chips. The question is was this China's attempt to trigger a black swan event in the US markets prior to the tariffs being enacted - a financial cold war if you will.
NASDAQ:NVDALong
by DontListenToThisClown
30
DeepSeek: Interrupt, Reprice and RelaunchCME: Micro E-Mini Nasdaq 100 Futures ( CME_MINI:MNQ1! ) #Microfutures DeepSeek might have changed the landscape of artificial intelligence forever. Since the launch of OpenAI’s ChatGPT in 2022, A.I. ran on advanced computer chips and large language models, costing billions for anyone to get in the game. DeepSeek, a Chinese startup, made a competitive A.I. model for a fraction of the cost, using less advanced chips. With 8-bit instead of 32-bit data, and by using data relevant to the task at-hand rather than keeping the entire database active all the time, DeepSeek cut the training cost by 95% and completed the task with 2,000 GPUs instead of 100,000. U.S. Stocks were down sharply on Monday on fear of an A.I. stock bubble popping. The Dow dropped 122 points, or 0.3%. The Nasdaq shed 3.2%, and the S&P 500 slid 1.9%. Wall Street raises concern that the billions spent to build big AI models could be done with much less investment. AI darling Nvidia dropped 11%, Broadcom lost 12%, and AMD shed 4%. Microsoft lost 4%. Amazon and Meta shed 2.4% and 1.4%, respectively. This is an example of “selling first and asking questions later”. Investors felt valuations are stretched for technology companies and headed for the exit. This highlights the risk involved in high-tech investment. DeepSeek disrupts the huge competitive edge held by OpenAI and Nvidia, making them less valuable overnight. In balance, a high-tech benchmark like the Nasdaq-100 index (NDX) provides better risk-adjusted returns. NDX: Past, Present and Future On midday January 27th, the NDX is quoted 21,137, down 3.0% for the day. Once the selloff is settled, we want to ask the question: “Is this a normal correction in a bull market, or the beginning of a bear market?” Let’s have a quick look at the past bear markets. During the dot-com bubble, the Nasdaq Composite Index peaked on March 10, 2000, at 5,048.6. As the bubble burst, the index plummeted to 1,139.9 by October 4, 2002. This represented a staggering decline of around 76.8%. The collapse was driven by the realization that many internet companies were grossly overvalued and unprofitable. In the 2022 bear market, NDX logged in a huge loss of 33.0%, bigger than that of DJX (-8.8%) and SPX (-18.1%). High-tech companies relied heavily on financing to fund their research, while many of them were yet to be profitable. The Fed rate hikes pushed their borrowing costs up by 500 basis points, worsening their financial woes. OpenAI's ChatGPT saved the day. This A.I. chatbot redefined the standards of artificial intelligence, proving that machines can indeed “learn” the complexities of human language and interaction. In my opinion, DeepSeek did not cancel out the breakthroughs achieved by others. On the contrary, by massively lowering the barrier of entry, DeepSeek could quicken A.I. development and its widespread adoption. A new era of A.I.-driven industrial revolution, started by OpenAI and boosted by DeepSeek, has only just begun. Additionally, Tech giants in the Silicon Valley are not sitting idly. OpenAI responded immediately by making the $200-a-month ChatGPT premium product free for all. The major players will learn from DeepSeek and redirect their research and development. After some short-term declines, the market will reprice the NDX component companies, setting them up for the next phase of the A.I. revolution. A.I. and Robotic Applications Are a Reality Last year, I took three trips to China and visited a dozen cities. What I observed there shows you how A.I. technology could be applied right now, not years away. • In the past, when I ordered takeouts while at hotels, I needed to go to the lobby to pick up my food. Nowadays, the hotel front desk would put my order inside a robot, which would then run and ride the elevator on its own and deliver the food to my room. • In fact, delivery robots are widely used for hotel room-service in China. They are not just in fancy hotels, but many budget hotels have also adopted them. The hospitality industry is labor intensive. Think about how much the labor cost this could cut down. • Restaurant patrons in China can scan a QR code to review menu, order food and pay for the meal online. Many have done without waiters, cashiers and printed menus altogether. If you want to save the 20-30% service tips, this may be the way to go. • Other emerging A.I. and robotic applications include driverless Taxi and food delivery by drone. On the one hand, they threaten the jobs of millions of people. On the other hand, they would save so much money for businesses and help their bottom-line. The adaptation to A.I. and robotic applications is slow in the U.S. Sometimes, they are being blocked by labor unions, who value job preservation more than anything else. Another reason is the lack of investment in A.I. infrastructure and commercial applications. On January 21st, President Donald Trump announced Project Stargate, a joint venture promising to invest up to $500 billion for infrastructure tied to artificial intelligence. This is a new partnership formed by OpenAI, Oracle and SoftBank. Separately, on January 22nd, Saudi Crown Prince Mohammed bin Salman announced that the oil producing country would invest up to $600 billion in the U.S., after his telephone call with President Trump. To sum up my analysis, it’s my view that A.I. applications are well under way, and large investment would help shore up A.I. infrastructure and steadily deliver cost-saving and efficiency improving applications across every corner of the economy. Project Stargate, named after the popular sci-fi movie, has the potential to spur another industrial revolution. After the market correction, NDX could rise higher once again. Trading with Micro E-Mini Nasdaq 100 Futures Investors sharing my view could consider the CME Micro E-Mini Nasdaq 100 futures (MNQ). The MNQ contracts offer smaller-sized versions of the benchmark E-Mini Nasdaq 100 futures (NQ). Micro futures have a contract size of $2 times the Nasdaq 100 index, which is 1/10th of the E-Mini contract. Micro contracts are very liquid. CME Group data shows that 1,279,703 contracts were traded on January 24th. Open Interest at the end of the day was 100,680. Buying or selling 1 MNQ contract requires an initial margin of $2,306. With Monday midday quote of 21,156, each March contract (MNQH5) has a notional value of $42,312. Compared with investing in the underlying stocks, the futures contracts offer a built-in leverage of about 18 times (=42312/2306). Hypothetically, a trader waits for the Nasdaq futures price to drop to 20,000 then enters a long order. If MNQ rebounds to its previous high at 22,100, the price change of 2,100 points (22100-20000) will translate into $4,200 in profit for a long position, given each index point equal to $2 for the Micro contract. Using the initial margin of $2,306 as a cost base, the trade would produce a theoretical return of 182% (=4200/2306). The risk of a long MNQ position is that the Nasdaq goes into a bear market. To hedge the downsize risk, the trader could set a stop-loss in his buy order. For illustration, he would put the stop-loss at 19,500 when submitting the buy order at 20,000. If the Nasdaq declines 20% from its peak of 22,100 to 17,680, the long position would be liquidated well before that, and the maximum loss would be $1,000 (= (20000-19500)*2). Happy Trading. Disclaimers *Trade ideas cited above are for illustration only, as an integral part of a case study to demonstrate the fundamental concepts in risk management under the market scenarios being discussed. They shall not be construed as investment recommendations or advice. Nor are they used to promote any specific products, or services. CME Real-time Market Data help identify trading set-ups and express my market views. If you have futures in your trading portfolio, you can check out on CME Group data plans available that suit your trading needs www.tradingview.com
CME_MINI:MNQ1!Long
by JimHuangChicago
1
Hobo is Deepseek exposure: could run hard!The market has woken up to bad news from China with Deepseek. Not going into detail but its generally bad for semiconductors, especially NVDA - which is quite high priced (understatement). Been looking for opportunities on the other side of this trade and found HOLO. They will be using DeepSeek R1 as the model for their holographic AI engine. Thing is: Deepseek is the hype (just like AI and Quantum was before) and HOLO is probably one of the few stocks that offers exposure AND its only sitting at a 33 million market cap. Jokes, much higher! First target: $7, if rally can sustain for multiple weeks: $18. I literally just bought, i honestly dont care if it pumped 50% pre-market - its a sign of strength. Let's see, obviously a risky play but also makes a lot of sense. Chart looks good after a retest of previous resistance.
NASDAQ:HOLOLong
by cryptobullethbtcxlm
9