Tilman

18 yo investor | Documenting my journey from €3.8k | Concentration is key | Compounding through asymmetric bets

[Tilman @tilmancompounds]

I spent 3 weeks building a 4 prompt AI pipeline that analyzes and rates stocks using Gemini.

Screener > Deep Research > Technical Analysis > Verdict

This is based on S&J’s asymmetric investing framework. Full prompts included.

Here’s how it works:

https://t.co/SUsz0kK47A

[Tilman @tilmancompounds]

Operating efficiency is becoming $NBIS real moat.
#1 on MLPerf shows they are maximizing their unit economics.
The operational edge is real and the fundamentals keep getting stronger.

[Daniel Koss @daniel_koss]

🚨 Just in: Nebius $NBIS just got third-party validation that its AI cloud can deliver world-class performance on NVIDIA’s latest AI systems.

For investors, this matters because MLPerf is one of the hardest independent AI infrastructure benchmarks. It helps prove Nebius is not just buying GPUs, it can operate them at elite efficiency for real customer workloads.

- Nebius ranked #1 among comparable NVIDIA HGX B300 submissions in 2 AI training tests

- On NVIDIA GB300 NVL72, Nebius finished within 3.1% of the fastest result across all 3 submitted benchmarks

- Bottom line: customers now have independent proof that Nebius can deliver top-tier AI training performance on cutting-edge NVIDIA infrastructure

$NVDA

[Tilman @tilmancompounds]

@josephinvesting Really good point. Wierd to see SpaceX pulling them down, while Elon is projecting 1 Trillion in revenue. Would mean that Planet (an actually REALLY profitable Space company) would benefit massively from this inflow.

[Tilman @tilmancompounds]

@anthonyiseman @Marketbeating20 @wealthmatica I used that exact value approach & held $UNH last year. But healthcare is defensive. SW multiples don't have to revert because AI is a structural threat. I just think a weak SW basket drags everything down. Still, I wish you good luck.

[Tilman @tilmancompounds]

@anthonyiseman @Marketbeating20 I don't know $ZETA well enough to judge. I just prefer setups supported by massive long term demand like AI SCAs. SW is dead money rn. I don't buy the vibe coding eats SW narrative, but if the market thinks that, there is little you can do. It’s a high opportunity cost play.

[Tilman @tilmancompounds]

@Marketbeating20 @remindmetweets Fair. Let's revisit this in a year.

[Tilman @tilmancompounds]

@Marketbeating20 CapEx is the wrong metric. The metric that matters is unit economics like PPT & compute efficiency. If the cost to generate tokens drops exponentially while demand explodes, the ROI scales up. The absolute spend becomes sustainable. The market doesn't price shifts on day one.

[Tilman @tilmancompounds]

@Marketbeating20 Fair point, it comes down to time horizon. If you are looking out 5 years, this approach definitely makes sense. I focus more on a 6 & 12 month time horizon. Just two different styles, both valid.

[Tilman @tilmancompounds]

@Marketbeating20 Hyperscalers don't sign 3 to 5 year SCAs for fun, they do it because they can't risk a supply deficit in memory, because it’s essential for every GPU. This supercycle is a structural paradigm shift over the next years & a way longer cycle than expected, not just a hype bubble.

[Tilman @tilmancompounds]

@Marketbeating20 Just because a company has intrinsic value doesn't make it a good investment if momentum is dead. I learned this the hard way with $DUOL. $MU isn't a hype story. HBM demand is locked in until 2030 through SCAs. Ignoring this supercycle is massive opportunity cost.

[Tilman @tilmancompounds]

@Marketbeating20 Calling $MU a bubble ignores this supercycle. HBM demand is locked in & memory makers are keeping capacity rational for pricing power. On $ZETA, pumping to $25 & dropping back to $19 doesn’t mean it’s undervalued. Relying just on valuation in a downward trend is a money trap.

[Tilman @tilmancompounds]

@Marketbeating20 I respect your capital size argument, but we are underwriting different things. Hyperscalers have to ramp CapEx or risk getting left behind, so this cycle lasts. $NBIS isn't a zero revenue bubble. It's a cloud leader with huge contracts, a clean balance sheet & execution record.

[Tilman @tilmancompounds]

@capitalzeph Getting close to the 1000£ mark!

[Tilman @tilmancompounds]

@Marketbeating20 Fair, but I feel like mega caps aren't the play. Real gains go downstream to supply chain leaders that benefit from CapEx like $NBIS. Holding dead momentum like $ZETA & $ADBE is catching a falling knife & missing this generational opportunity when $MU can easily do another 50%.

[Tilman @tilmancompounds]

@MB_Hogan @J5jayfive Thanks Mark!

[Tilman @tilmancompounds]

@MB_Hogan @daniel_koss

[Tilman @tilmancompounds]

@daniel_koss This is great stuff Daniel! Ai is a future proof skill.

[Tilman @tilmancompounds]

@josephinvesting

[Tilman @tilmancompounds]

🚨 Just reverified all my $NBIS data & the technical setup is crazy. We are sitting right on the 38.2% Fibonacci level at $224. Short term EMAs are broken from the pullback, but don't be surprised if this absolutely rips when the market recovers. Potentially great entry zone.