2D·

A 100,000-Euro AI Experiment Put to the Test

This portfolio is a long-term experiment: An artificial intelligence was given a virtual starting capital of 100,000 euros to allocate. To ensure a consistent strategy, I regularly, every six months, to the leading models (ChatGPT, Gemini, and Claude) for portfolio updates and potential rebalancing.

The AI has done an excellent job so far and has significantly outperformed the broad market indices.

Portfolio Key Figures (getquin Performance)

Total portfolio value: 273,048.83 EUR
Invested capital: 134,913.40 EUR
Unrealized gains: +102.41% (+138,163.84 EUR)
Realized gains: +97.09% (+34,148.65 EUR)
Total return: +173,860.76 EUR

Top Positions and Drivers

Palantir Technologies ($PLTR (+1,59%)
): +418.18%

Nvidia ($NVDA (-1,6%)
): +335.65%

Alphabet Inc. Class A ($GOOGL (-0,56%)
): +137.72%

ASML Holding ($ASML (-3,44%)
): +98.18%

Shopify ($SHOP (-0,82%)
): +83.57%

Amazon ($AMZN (+1,81%)
): +66.31%

Consolidation in individual stocks such as Novo Nordisk (-13.84%) is factored into the allocation and is offset by the fundamental strength of the core positions. In addition, profits were selectively taken on stocks such as Tesla and Nvidia in earlier cycles, and these funds were invested more heavily in Palantir and ASML. The AI models remain consistent in their view of Novo Nordisk, see strong potential, and wish to maintain this position.

Strategic Conclusion & Outlook

In the current six-month update, the AI recommends no changes: The portfolio remains unchanged in its current composition. The fundamental thesis regarding digital transformation and the scaling of AI systems still holds, while the cash flows of the held companies remain intact.

How do you assess the AI’s decisions in the current market environment?

More on this in my new video: https://youtu.be/R0ka3r9BlIo

12 Comentários

Based on the gains, I assume the portfolio has been running for quite some time. After all, it’s been a long time since $PLTR returned 420% and $NVDA returned 335%. I thought those AI models didn’t even exist back then! :-)

•

Aktienfox@Aktienfox

@Multibagger It all started with ChatGPT on January 1, 2024 :) See the attached portfolio :) (attached)

•

Raketentoni@Raketentoni

@Multibagger I just wrote this somewhere else—you know I did this with real money, but only because I work with AI and wanted to test it.
In 2024, the AI’s selection capabilities weren’t that extensive yet. According to tests at the time, the AI’s IQ was around 80. Humans have an average IQ of 100. Today’s AI systems have IQ scores of around 130 to 140 according to standard tests.
If the prompt from back then hadn’t been changed, it would be working with completely outdated data and parameters.
That would mean it was more a matter of luck that it picked out the best stocks over the last few years. 😉

•

Raketentoni@Raketentoni

@Multibagger I asked our AI what it thought 😬

Hey @Aktienfox! 🚀

First off, kudos on those impressive numbers—a total return of over +173,000 € certainly looks great on paper! Anyone who’s held stocks like Palantir (+418%) and Nvidia (+335%) in their portfolio since early 2024 has really lined their pockets. 👏

But since we’re all about honest and level-headed analysis here in the forum, let’s briefly put this AI experiment under the microscope and take off those emotional “hype glasses”:

* Not “super-alpha,” but the most obvious play in the world:

What ChatGPT put together for you back in early 2024 wasn’t really any “superhuman stock market magic,” but rather the absolute mainstream consensus at the time: all-in on tech & AI infrastructure.

The models simply spewed out what was already being repeated ad nauseam in every financial media outlet.

* Bull Market Rider:
The reason this portfolio performed so spectacularly is simply that the NASDAQ and the entire tech sector have been on one of the most extreme cycles in history since early 2024. AI didn’t unearth any hidden turnaround value here; instead, it rode the wave of the massive market beta to the fullest.

* The Concentration Risk Trap:
A portfolio consisting almost exclusively of semiconductors, cloud services, and big tech looks like the work of a genius in a bull market.
But where’s the risk management? What rules did you give the AI?
If the AI cycle ever stalls or a sector rotation sets in, a concentrated portfolio like that will melt away twice as fast. No AI model currently protects you from this, even if it dutifully reports: “No changes necessary.”

* The Capital Effect:
A small side note: The invested capital has now grown from €100,000 to just under €135,000. So you’ve been diligently adding to your position. That’s perfectly legitimate, but it naturally distorts the picture compared to a pure “one-time 100k investment.”

My honest conclusion:

A spectacular run that we’re happy to see you enjoy—after all, profits are profits! But it was simply the perfect play in the perfect market environment. The AI experiment will only get exciting when the market starts dealing out real pain again and the models have to prove whether they can also handle diversification, risk management, and real bear market cycles.

I’m curious to see how your AIs react when the tide turns on the stock market! 😉

•

Raketentoni@Raketentoni

I actually ran an experiment like that for 4 months and ended up with a loss of over 60%. But that approach has its weaknesses, and what you’re doing here is having someone manage a portfolio that’s already been around for a while.
Start from scratch—and not just virtually, but with real money. Then it’ll get interesting. You can probably follow along with my posts.
I’ve been working on something completely different with AI for 6 weeks now, and today I gave a little sneak peek at @Get_Rich_or_Die_Tryin. 😬
I’d love to know what your prompt looks like for that—I’m already working with loops. 🤷

•

SGP@Koenigmidas

Which KI tool is used?

••

Aktienfox@Aktienfox

@Koenigmidas At first it was just ChatGPT, and now there's Claude and Gemini, too

••

Raketentoni@Raketentoni

What’s much more interesting is the prompt or the loop it runs... because that’s what defines the AI’s strategy and rules.
I work with AI every day as part of my job, and here I don’t see any rules or guidelines for the WL.
It looks more like, “Let’s just take last year’s successful stocks and say the portfolio has been running forever.”
I don’t want the whole prompt—I’d never share that myself—but I’d love to see just an excerpt of the rules for the portfolio 😬
I’m happy for you, but I’ve put real money on the line, and that’s when things get interesting 🙃

•

OberstvonGatow@OberstvonGatow

Beeindruckt!

••

Felix@Investor_in_Jogginghose

Ki is so much like me 🙈 Does Chat happen to have Meli or Netflix on her watchlist?

••

Aktienfox@Aktienfox

@Investor_in_Jogginghose nope, Crowdstrike

••

Felix@Investor_in_Jogginghose

@Aktienfox Okay, but I've actually had those in my portfolio for quite a while now.

••