Monday Market Open

Good morning gents,

I hope you had a class weekend, but it’s time to get back on the grind.

This edition of Monday Market Open is a bit longer, with a special report down below on how much DeepSeek AI will really affect Nvidia (spoiler, it won’t).

So enjoy and get ready for what I can sense will be a banger week.

Market Indices & Commodities

News
The Recap

You thought a lot happened last week? Well the weekend was quite a thing to behold with the confirmation of tariffs making everything look a bit wobbly…

• Trump confirms tariffs on Canada and Mexico 🏦

• Canadian dollar (CAD) hits lowest point since 2003 🇨🇦

• Pro-Russian paramilitary leader killed in Moscow bomb blast 💥

• $250bn wiped off crypto markets over the weekend 📉

• Musk says Trump has agreed to shut down $40b US foreign aid agency 🙅‍♂️

Graph of the Day
Apple Services Now A $100b Business

In light of Apple’s ($AAPL) decent earnings last week, the graph above shows how the services Apple offers alone, is a $100 billion business.

News
This week

The middle of the week is gonna be big this week. But don’t worry Monday and Friday will be quiet so you can ease yourself in and out of this week smoothly (preferably with a beer too).

• Monday - Palantir ($PLTR) earnings.

• Tuesday - PayPal ($PYPL), Spotify ($SPOT), Pfizer ($PFE), Pepsico ($PEP), Ferrari ($RACE), AMD ($AMD), Alphabet ($GOOGL), Snap ($SNAP), Chipotle ($CMG) and Electronic Arts ($EA) all announce earnings.

• Wednesday - Uber ($UBER), Walt Disnep ($WALT), Novo Nordisk ($NVO), Toyota ($TM), Ford ($F) and Qualcomm ($QCOM) share their earnings. US Balance of Trade for December.

• Thursday - Roblox ($RBLX), Peloton ($PTON), Amazon ($AMZN), Cloudflare ($NET) and Pinterest ($PINS) earnings. Phillip Morris ($PM) also announce their earnings. Given the massive rise in popularity of Zyns, it’ll be interesting to see how this reflects in their earnings. (To be honest, I reckon my personal consumption will make its own fucking impact lol.) US Initial / continuing jobless claims also come out.

• Friday - No big earnings. Unemployment / Labor Force Rates for January is released.

Now You Know
Deep Dive Into DeepSeek: How Fucked is Nvidia?

China’s back-to-back AI launches—DeepSeek R1 and Alibaba’s Qwen 2.5-Max—could very well be nothing but a geopolitical attempt to mog the West by using sus claims of "cutting costs" and" increasing efficiency".

- DeepSeek R1: Claims to be 10x cheaper than GPT-4, trained on a shoestring budget.

- Qwen 2.5-Max: Boasts 89.4% accuracy on Arena-Hard, crushing OpenAI and Meta’s models.

Both models flaunt “efficiency” thanks to Mixture-of-Experts (MoE) architectures, which use fewer computational resources. Alibaba even bragged about training Qwen on 20 trillion tokens with homegrown Hanguang 800 chips.

But is China lying about not using Nvidia chips?

Despite U.S. export bans, reports suggest Chinese tech giants hoarded $5B+ of Nvidia GPUs before sanctions hit. In layman's terms, China’s “efficiency” might rely on smuggled H100s and A100s retrofitted into domestic systems. Beyond the potential purchases of pre-sanctioned GPUs, rumors are spreading that the Chinese are actively doing whatever they can to get their hands on/purchase American chips.

The motive behind this? (possibly) Power—by hyping cheap AI, it could be China’s attempt to smoke out Nvidia’s stock (already down 17% this week) and bait the U.S. into loosening chip restrictions. This could very well be a high-level psyop with trillion-dollar stakes.

The issue for Nvidia:

China’s AI spam fest isn’t just about having the best models—it’s may be about decreasing demand for Nvidia’s GPUs. Here are the numbers:

- Qwen 2.5-Max’s MoE design slashes infrastructure costs by 40-60%, meaning enterprises need fewer GPUs.

- DeepSeek’s “low-cost” AI could flood the market with budget models, decimating the need for $40K H100 clusters.

Why China’s new AI models are actually fucking useless:

‍Large Language Models (LLMs) like OpenAI’s GPT-4 or Alibaba’s Qwen 2.5-Max are essentially autocomplete on roids. They’re trained on heaps of text to predict the next word in a sequence, spitting out “human-like” responses by mimicking patterns—not logic.

The result? They make up facts, parrot pre-approved biases, and fail catastrophically at tasks requiring real reasoning. They’re the McDonald’s of AI: quick, cheap, and deeply unsatisfying if you need actual nutrition.

Summary:

China’s LLM win is like a bump—it’ll feel good for a bit, but it all comes crashing back. Whether they’re using smuggled Nvidia chips or not, they’re still stuck in the “bigger, faster, dumber” AI loop. 

Nvidia may be down, but it’s unlikely it’s gonna be out any time soon…

^{_{Thank You}}
That’s all Folks

Hey, thank you for checking out Monday Market Open. I hope you enjoyed it, learned something new, and are feeling ready to attack the week. Keep an eye on your inbox for more Alpha coming your way this week.

Cheers,