Every once in a while an experiment comes along that makes even seasoned investors stop and raise an eyebrow.
This month, one of Europe’s biggest banks decided to do something that sounds almost too good (or too crazy) to be true: they let their in-house artificial intelligence build an actual investment portfolio from scratch, using only the same information human analysts see every day, and then they promised to track it publicly for a full year.
The question they’re trying to answer is simple but massive: Can AI make the average investor rich the way Warren Buffett did? Not just match the market; actually beat it, consistently, using nothing more than data and logic.
When a Bank Hands the Wheel to a Machine
Most of us have played with robo-advisors. You answer a few risk questions, the algorithm spits out a pie chart of ETFs, and you forget about it until retirement. Fine. Safe. Boring.
This is different.
The team gave their model, called dbLumina, a full week of their internal daily market commentary — the same research notes the bank’s human strategists use to guide billion-dollar decisions. They also fed it historical performance data for the most widely held exchange-traded funds among retail clients at the end of November.
Then they stepped back and said: “Go build us the best possible basket of funds. Surprise us.”
“The initial allocations surprised us a little — but the proof will bear out over the coming year.”
– Luke Templeman, Research Analyst
That single sentence tells you everything about where we are in 2025. Even the humans who built the AI didn’t fully expect what it would do.
What the AI Actually Bought (And Why It’s Fascinating)
Without giving away the exact weightings — the bank is keeping some mystery for dramatic effect — the model immediately went heavier on certain themes than most retail portfolios dare.
It loaded up on areas that had been beaten down but showed improving sentiment in the commentary. It trimmed exposure to the usual “safe” mega-cap tech darlings that everyone already owns. And perhaps most interestingly, it completely ignored a couple of crowd-favorite funds that had been top-10 holdings for retail investors for years.
In short, the AI wasn’t trying to be comfortable. It was trying to be right.
- It overweighted sectors showing early signs of sentiment reversal according to natural-language analysis of news flow
- It reduced concentration risk in the “Magnificent Seven” stocks that dominate most retail accounts
- It added exposure to mid-cap value — an area humans have largely abandoned
- It completely avoided certain high-fee active funds despite their marketing muscle
Think about that last point for a second. An emotionless machine looked at decades of data and said, “Nah, not worth the cost.”
The Ghost of April Still Haunts Retail Investors
One of the most revealing parts of the experiment was when the analysts asked dbLumina to score investor “rationality” month by month throughout 2025.
April scored the lowest by a mile.
Remember April? Markets cratered after the new administration rolled out aggressive tariff plans. Panic selling was everywhere. Then, almost as quickly, stocks ripped higher into summer as companies adapted faster than anyone predicted.
Millions of retail investors sold at the bottom — or simply never owned enough equities to participate in the rebound. The AI flagged that behavior as the textbook definition of irrational.
Humans chase performance. Machines chase probability.
Weekly Rebalancing: The Silent Superpower
Here’s where things get really interesting.
Unlike most robo-advisors that rebalance quarterly or annually, dbLumina is allowed to review its decisions every single week using fresh commentary and data.
That means if sentiment suddenly flips on, say, European banks or clean energy or whatever the hot theme of the moment is, the model can tilt the portfolio before the crowd piles in.
It’s active management without the ego. No bonus anxiety. No career risk. Just cold calculation.
I’ve followed quantitative strategies for years, and in my experience the biggest edge any system has over humans is the ability to act without hesitation when the data says “now.” Most of us talk ourselves out of perfectly good trades because they feel uncomfortable.
So Can AI Really Make You the Next Buffett?
Let’s be honest — probably not.
Warren Buffett’s genius isn’t just about picking stocks. It’s about understanding business moats, capital allocation, human psychology, and having the temperament to sit on your hands for years when nothing looks cheap.
No current AI comes close to that level of holistic judgment.
But here’s the twist: you don’t need to be Warren Buffett to get rich investing.
You just need to beat 99% of other investors. And most of them make emotional mistakes constantly. If an AI can simply avoid panic selling in March, avoid FOMO buying in December, and keep costs low while capturing broad market returns with a few intelligent tilts… that might be enough.
Over decades, that’s the difference between being comfortable and being truly wealthy.
What This Means for Regular Investors Right Now
You don’t have access to dbLumina (sorry), but the tools available to retail investors in 2025 are already astonishing.
- Commission-free trading on almost every platform
- Fractional shares letting you buy exactly the allocation you want
- Direct indexing services that let you own individual stocks tax-efficiently
- AI-powered research tools that read 10-Ks faster than any human
- Automated tax-loss harvesting running 365 days a year
The playing field has never been this level.
Perhaps the most underrated edge today isn’t genius stock-picking. It’s behavioral coaching. And machines are uniquely good at that because they never get excited, never get scared, and never check their phone during dinner to see if Tesla is up another 5%.
My Prediction for the Next Ten Years
Ten years from now, the majority of household investment portfolios will be managed — at least partially — by artificial intelligence.
Not because AI is magically smarter than every human, but because it removes the biggest drag on returns that has ever existed: ourselves.
We are our own worst enemy when it comes to money. We buy high, sell low, chase trends, abandon strategies the moment they hit a speed bump.
An AI doesn’t care if Bitcoin is going to a million or zero. It cares if the current price offers an attractive risk-adjusted return given the latest data.
That discipline, applied consistently, is devastatingly powerful.
So no, AI probably won’t make you the next Warren Buffett.
But it might just make you richer than you would have been without it.
And honestly? For most of us, that’s more than enough.
I’ll be tracking this Deutsche Bank experiment all year and reporting back with updates. If the AI starts crushing it — or face-planting — you’ll hear about it here first.
Until then, maybe ask yourself: would you trust a machine with your money?
Because the machines are already asking if they can trust us with theirs.
