The spectacular rise of the stock markets through 2024 and early 2025 have been powered by a speculative AI mania. There can be little doubt that valuations have got somewhat ahead of themselves by any measure. Yet, as always in manias we hear, this time it is different, but is it? One just has to look at the market reaction to Chinese start-up DeepSeek’s artificial intelligence (AI) models being listed as a free open-source app. They were listed for free over the last weekend of January, went to the top of the Apple & Google app downloads and markets panicked immediately.
Panic ensued and NVidia finished down 17% the following Monday, the biggest loss of market capitalization in history with a $560 billion evaporation. The Semiconductor index sank 9%. NVDA shares performed a dead cat bounce at best and closed 15.8% lower at the end of the week. Looking back at the end of February, NVidia has delivered earnings that beat expectations, but the stock has not recovered, neither has the tech heavy NASDAQ or small cap Russel 2000 index.
Nvidia Shares had been well poised for what happened over the Lunar New Year Weekend, they are up roughly 28-fold in the past five years. They had completed a potential double top spitting the 4.618 fib extension. That said timing the impulse is one of patience, discipline and usually hindsight.
Was DeepSeek the beginning of the end of the AI Speculative Bubble?
There is no coincidence this all came about in the first weeks of U.S. President Trump’s Presidency. The move started the last week with a little tit for tat from China with AI. Over the Weekend, mind you on Luna New Year break the Chinese AI DeepSeek ‘market bomb’ was dropped. Oh my, suddenly, key aspects of the bullish narrative looked wonky.
The move also slammed- energy and utility stocks, which had been rising to the moon on the need for more energy to run data capture, abruptly headed back to earth. Nuclear energy related stocks were down 40-50% on the day. President Trump called it a “wake-up call.”
Tit for tat you say. The move came just after the $500 billion AI infrastructure initiative, Stargate, that involved OpenAI, Softbank, and Oracle (ORCL) was announced. We then had the FOMC & ECB in the middle and bookended with Tariffs kicking off February, which dominate the news come the end of the month with DeepSeek event seemingly forgotten.
The move is reminiscent of the subprime mortgage eruption in June 2007. We have loose conditions, speculative leverage, liquidity overabundance, and the AI mania. Back to the subprime eruption, the stock market posted record highs that October. “Core” AAA MBS initially benefited from lower policy rates and flight away from the risky “Periphery” (subprime derivatives, ABS, and Alt-A mortgages).
- January 27 – Financial Times: “Technology stocks tumbled on Monday after Chinese artificial intelligence start-up DeepSeek stunned Silicon Valley with advances apparently achieved with far less computing power than US rivals. Shares… Nvidia, one of the biggest beneficiaries of spending on AI chips, plunged almost 17%, wiping out almost $600bn of market value, a record loss for any company. DeepSeek last week released its latest large language AI model, which achieved a comparable performance to that of US rival OpenAI, even though the company has previously claimed to use far fewer Nvidia chips. Venture capital investor Marc Andreessen called the new Chinese model ‘AI’s Sputnik moment’, drawing a comparison with the way the Soviet Union shocked the US by putting the first satellite into orbit. The results sent a shockwave through markets on Monday, as investors reassessed the likely future investment in AI hardware.”
The surprise here was it came over the Luna New Year break. The Chinese AI DeepSeek was listed for free and went top of the Apple & Google app stairs. Panic ensued and NVidia finished down 17% Monday, the biggest loss of market capitalization in history with a $560 billion evaporation. The Semiconductor index sank 9%. NVDA shares performed a dead cat bounce at best and closed 15.8% lower at the end of the week.
- January 30 – New York Times (Andrew Ross Sorkin, Ravi Mattu, Bernhard Warner, Sarah Kessler, Michael J. de la Merced, Lauren Hirsch, Edmund Lee and Vivienne Walt): “Wall Street has been on tenterhooks about how Silicon Valley would respond to DeepSeek, the Chinese start-up whose low-cost artificial intelligence software threatens to undercut the pricey American approach to the technology. So far, the answer appears to be: full steam ahead. Meta and Microsoft… said they each planned to keep spending billions on A.I. And news reports about SoftBank’s talks to inject billions more into OpenAI suggest that deep-pocketed investors are still bullish on the ChatGPT creator. Continuing to spend heavily on A.I. will be a ‘strategic advantage over time,’ Mark Zuckerberg, Meta’s C.E.O., told analysts…, defending plans to invest up to $65 billion… And Amy Hood, Microsoft’s C.F.O., told analysts that her company — which plans to invest about $80 billion in A.I. this fiscal year — will grow such spending next year, though at a slower rate.”
The impact was amplified given trading had boomed in products that allow investors to amp up bets on the world’s most popular equities, with investors pouring a record of more than $6.5 billion into such “single-stock ETFs”, vehicles that track just a single company but use derivatives to intensify bullish or bearish wagers. A major winner in this category was a fund from GraniteShares, which provides two times the daily returns of market powerhouse Nvidia Corp. The strategy, which trades under the ticker NVDL, saw assets explode to a peak of $6.7 billion in late November 2024, with returns in excess of 350% that year. Funds tracking MicroStrategy, Tesla Inc. and Coinbase Global Inc. have followed similar paths.
The move tracked by the KnovaWave extension model followed the announcement of NVDA 4/1 split come levels off the energy break NVidia hadn’t looked back with many gaps below until it hit the 2.618 target of the exhaustion phase. We saw another power move off the $200 retest (old $800) & earnings off $300 which retested. It is a clear leader of SOX SMH look for cues there and ABC failures for changes.
Speculative Bubble aren’t new, we had the nineties tech bubble, followed by a phenomenal mortgage finance and housing Bubbles. China’s apartment Bubble took bubbles to another level. Now we have the AI Bubble for the potential for unmeasured financial excess and resource misallocation.
NVidia and it’s influence on Technology and Speculative Markets
NASDAQ 100
Nasdaq fell just short of retesting the ATH after it bounced hard after the August rout.
Background: Fuel was after it broke and held the weekly Tenkan to see a spit of a spit fail which is completive of 5 of some degree with Chikou rebalancing. From there we sold off right to Tenkan (as did SPX) and bounced hard Support Tenkan to Kijun. Watch Chikou for divergence for continuation or failure. Divergence with Russell also a clue.
Russell 2000
The small cap Russell RUT broke out of multi year flag and then spat the previous high. Unlike SPX and NDX we could not get to new ATH until near the end of this cycle after it got through Tenkan and Kijun on the bounce. This is the index showing more of the fast money crowd and is trading like it. Needs to get traction in here for bulls. Support +1/8 through 7/8 (cloud base)
What is DeepSeek AI?
DeepSeek AI is designed to offer open-source LLMs, efficient architecture, advanced reasoning, multimodal learning.
More accessibility and efficiency: DeepSeek is designed to be less expensive to train and use than many competing large language models (LLMs). Its architecture allows for high performance with fewer computational resources, which is designed to lead to faster response times and less energy consumption.
Open-source availability and rapid development: DeepSeek is under active development with new models and features being released regularly. Models are often available for public download (on Hugging Face, for instance), which encourages collaboration and customization.
Advanced capabilities: Reasoning and multimodal learning Models like DeepSeek-R1 are designed with a focus on advanced reasoning capabilities, aiming to go beyond simple text generation. DeepSeek is expanding into multimodal learning, handling diverse input types such as images, audio and text for a more comprehensive understanding.
Limitations: Bias and context Like all LLMs, DeepSeek is susceptible to biases in its training data. Some biases may be intentional for content moderation purposes, which raises important ethical questions. While efficient, DeepSeek could have limitations in handling extremely long texts or complex conversations.
Architecture and performance: DeepSeek uses a “mixture of experts” architecture, employing specialized sub models for different tasks, enhancing efficiency and potentially reducing training data needs. DeepSeek has demonstrated competitive performance, comparable to established models in certain tasks, especially mathematics and coding.
We now head into the rest of the year with turmoil in technology markets, the impact of geopolitical gamesmanship in a highly leveraged environment. Where will we sit at year’s end?
Is this the top of NVDA, is AI a fad to be replaced, will there be a new leader in technology or is this simply a consolidation in the NVidia story? So many questions for investors and traders.
Focus on yourself and what YOU CAN INFLUENCE, set your trading plan and goals in be set for 2025.
-comment section below data-
Subscribe and Follow
Find us at www.traderscommunity.com
Follow our contributors on Twitter @traderscom @thepitboss16 @knovawave @ClemsnideClem
Note these charts, opinions news and estimates and times are subject to change and for indication only. Trade and invest at your own risk.
Trade Smart!