The DeepSeek Surprise

Why it's impossible to predict moments like this and much more...

Hey everyone,

Welcome back to Stocks To Space, where I curate the best ideas, tools and resources I’ve found each week as I explore my curiosities.

One of my goals for Stocks To Space in 2025 is personalising your experience.

I’d love to know what themes interest you most. What topics fascinate you and make you want to explore them further each week?

Is it:

  • AI?

  • Finance?

  • Tech startups?

  • Personal growth?

Let me know in the feedback form at the bottom of this email.

And as always, if you find something thought-provoking in this week’s edition, forward it to a friend.

IDEAS
The DeepSeek Surprise

Created with Midjourney

History's most profound moments aren't the ones we see coming.

They're the black swans that blindside us and rewrite the rules overnight.

Humans spend countless hours trying to predict the next big thing, but it’s mostly folly.

The most transformative events in history have always been surprises—the iPhone, the fall of the Berlin Wall, and the rise of social media.

None were predicted by the experts of their time.

Today’s example of this phenomenon is DeepSeek.

Not a single person expected R1 to have the impact it’s had on the AI industry.

Because of that surprise, it's moved markets, upended preconceived notions, and redefined the art of the possible regarding AI models.

Who knows what's next? The funny thing is, that's the point.

INSIGHTS
1 Article

Source: statista

The chart above is the most direct visualisation of why DeepSeek upended the AI market this week.

The implications:

  • DeepSeek charges 5x less than OpenAI for input and output processing.

  • It’s cheaper than most comparable models from the hyperscalers, fundamentally challenging the assumptions of the AI cost curve.

  • DeepSeek demonstrated that frontier AI performance could be achieved with last-gen chips at 1/50th the training cost.

This is why NVIDIA plunged 17% on Monday, January 27th, wiping $589 billion from its market cap—the largest single-day loss for a single stock in history.

1 Post

1 Video

TOOLS
Perplexity x DeepSeek-R1

Source: Perplexity

Today’s insights were researched using Perplexity’s Pro Search functionality and distilled by Reasoning with R1—using DeepSeek’s newest revolutionary model.

For over a year now, Perplexity has been my go-to search tool, slowly replacing Google.

I love the LLM-based search experience for digging deeper into topics and building context.

Adding R1 on top now allows me to build expert analyst reports backed by citations with the hypothesis, intro, body, and conclusion reasoned out and written faster and more effectively than ever.

The controversy aside, I can see why there’s such a fuss around DeepSeek-R1.

Unfortunately, Pro Search and Reasoning with R1 are only available on Perplexity’s Pro Plan.

However, since one can access the best models from OpenAI, Claude, Grok, and now DeepSeek, the $20 a month to Perplexity might be worth it.

THOUGHTS
Quote I’m Pondering

“DeepSeek-R1 is AI's Sputnik moment”

— Marc Andreessen

Was this email forwarded to you? If you liked it, sign up here.

If you loved it, forward it along and share the love.

Thanks for reading,

— Luca

What theme do you want me to focus on in 2025?

This will help me build the best newsletter for you.

Login or Subscribe to participate in polls.

*These are affiliate links—we may earn a commission if you purchase through them.

Reply

or to participate.