Five stocks. Palantir, PDD, NIO, Meta, Cloudflare.
2024 was when I had conviction and executed. Concentrated AI/tech positions, active trading, taking profits and re-entering. That year generated most of the 463%.
2025 was different. I went defensive—held 30-40% cash waiting for a crash that never came. Result: 4.16% alpha over two years, but underperformed by 12.34% in 2025 alone.
The question for 2026: was the thesis wrong or just early?
But first, the five trades that worked.
Palantir: 150% (Then I Came Back for Another 45%)
October 2023. AI was everywhere. Everyone wanted exposure, but most people were chasing the application layer—consumer chatbots, productivity tools, AI writing assistants.
I looked at Palantir and saw something different. They weren’t building toys. They were building the infrastructure governments and enterprises would need to actually deploy AI at scale. Defense agencies, intelligence operations, commercial data platforms. The picks-and-shovels play.
I started buying around $15-16. Small position initially—500-1000 shares. The thesis was simple: if AI was real, Palantir’s 20 years of relationship capital with the U.S. government would print money.
But I didn’t just buy and hold.
I traded it. Actively. Bought at $15-16, sold at $19-20. Watched it pull back, bought again. Sold at $21. The AI narrative kept accelerating through early 2024, and I kept scaling in and out.
By February 2024, when I sold at $21.74, the stock was up 40% from my initial entry. The valuation screamed overheated—15x revenue when comparable software companies were at 5-8x.
But I kept coming back. The thesis hadn’t changed. Government AI adoption was accelerating. I bought options at $22 strike, exercised them in June, sold the shares at $30 in August.
By August 2024, I closed out the entire position. +150.79% realized.
Then I stayed out for 6 months. Just watched it from the sidelines.
March 2025. Stock had pulled back from the highs. Same thesis, fresher valuation. I re-entered.
I traded it several more times through August 2025—smaller sizes this time, 100-200 shares instead of 1000. As the price went higher, I reduced position size. Risk management.
Final exit: August 2025. Position 2: +45.54% realized.
Here’s what I keep thinking about: I took a 150% gain, stepped aside completely, then came back and made another 45%. Same stock. Same thesis. But I didn’t marry the position. Both times I took profits and walked away.
PDD: 105% on Temu’s Global Expansion
Late 2023. PDD Holdings—parent company of Temu—caught my attention for reasons most investors missed.
Everyone focused on the “China risk” narrative. I looked at the actual business: Temu had become a legitimate competitor to Amazon and Alibaba. The numbers were staggering—explosive customer growth in the US and Europe, strong unit economics, and a business model that was actually working.
By late 2023, Temu had over 130 million US users and was rapidly expanding across Europe. The platform was second only to Amazon in e-commerce traffic globally. This wasn’t a flash-in-the-pan Chinese app—this was a structural threat to Western e-commerce.
The thesis was simple: if Temu could sustain this growth trajectory, PDD was massively undervalued. The market was pricing in geopolitical risk while ignoring the fundamental business momentum.
I started building in November 2023 around $110. First exit in February 2024 at $145—quick 32% gain in three months.
The stock kept running to $147. I re-entered in March—not because I was wrong to sell, but because the thesis was still intact. Then China regulatory headlines hit. I flipped it immediately at $147.55 for a quick exit. Small gain, but preserved capital when the news turned negative.
That’s the tradeoff with active management. I don’t try to ride the full move—I take profits, reassess, and re-enter when the setup resets.
Mid-2024, PDD crashed to $90-100 range as China fears intensified. Third opportunity.
Bought at $100 in August, added more at $92 in September. Rode it back to $155 in October for the final exit.
Total return across three round trips: +104.96%.
Why I exited in October 2024: Trade tensions were escalating. Trump was leading in polls, and his campaign rhetoric around China tariffs was getting louder. I’d made my gains—time to derisk before policy risk materialized.
Four months later, April 2025, Trump ended the “de minimis” exemption that allowed Temu to ship packages under $800 duty-free. The stock crashed 22% that month, hitting a 52-week low of $87.11. Temu had to add 145% “import charges” to customer orders, effectively doubling prices overnight.
The thesis was right—Temu was a legitimate competitor. The trade worked because I recognized when policy risk outweighed fundamental momentum. Exit before the hammer drops, not after.
NIO: Three Attempts, Three Different Results
I was involved with NIO through my work prior to its IPO. I understood the business, the technology, the competitive landscape. That familiarity made me think I had an edge.
I’ve traded NIO three separate times. Same company, three completely different approaches.
2019 – The Overexpansion Disaster
Bought at $8. Chinese EV revolution thesis.
Then the liquidity crisis hit. NIO had overexpanded—burning through $3.67 billion while delivering fewer than 32,000 vehicles. Analysts were calling them the “king of cash burn.” The stock fell from $8 to $2.
I averaged down the whole way. “I know this space. The technology is real. This is temporary.”
Eventually capitulated at -39%.
2020 – The Redemption
NIO secured a $683 million bailout from Hefei government. Stock was at $4.
The thesis had fundamentally changed: Chinese government just demonstrated it wouldn’t let a strategic EV company fail. The overexpansion problem was solved with capital injection.
Time to re-enter. Position size: 2,000-3,000 shares. The opportunity was there, but I wasn’t going to oversize it.
Rode it from $4 to $20. +97%.
Exited at $20 in late 2020. Disciplined profit-taking on a position that had already 5x’d from the lows. As of late 2025, NIO trades around $5—right back where I sold.
2024 – Pattern Recognition, One Last Time
NIO crashed back to $3.80 in August 2024. Same fear pattern. Different year.
I recognized it immediately. Bought 2,000 shares at $3.86-3.89. Sold at $5.02 three weeks later.
+32%.
But I could see the writing on the wall. By 2025, Chinese EV production capacity had hit 36 million vehicles against just 14 million in annual demand. That’s a 22 million unit surplus. Companies were flooding global markets with exports just to move inventory. The sector’s median net profit margin had collapsed to 0.83% from 2.7% in 2019. Analysts projected over 10 automakers would face bankruptcy in 2025, with hundreds of brands already gone under in brutal price wars.
I made one last trade at $3.80, but after that? I shifted focus to XPeng—better execution, faster growth trajectory (331% YoY vs NIO’s 40%), clearer path to profitability in 2026. NIO was still burning $3 billion annually. XPeng had narrowed quarterly losses to $90M with products like the P7+ and Mona M03 gaining real traction.
Same sector thesis. Different horse. When the fundamentals shift, switch companies—don’t abandon the sector entirely.
Meta: 75% Buying Before the AI Narrative Shifted
December 2023. The market thought Meta was dying:
- TikTok competition threatening engagement
- Apple’s privacy changes destroyed the ad model
- Billions wasted on metaverse
Stock was at $350. I bought.
The contrarian case wasn’t about the metaverse. It was about Meta’s track record of innovating when threatened. They’d rebuilt for mobile. Acquired Instagram. Built Reels to compete with TikTok. Deployed AI across WhatsApp, Messenger, and ad targeting.
I focused on two things: AI infrastructure buildout and Reels monetization gaining traction.
February 2024: Sold at $495. +41% in just over two months.
July 2024: Bought again at $450. Sold September at $550. +22%.
Multiple smaller trades through 2025 as the stock continued higher.
Total return across all Meta trades: +74.75%.
I didn’t hold through the whole rally. I traded it. That active management captured multiple 20-40% moves instead of trying to time one perfect entry and exit.
I still maintain a small position. The AI bubble fears that kept me defensive in 2025 haven’t gone away, but Meta’s the type of company that can grow through it—real revenue, real profits, not just hype. Small enough to sleep at night, large enough to matter if I’m wrong about the correction.
Cloudflare: 54% From Three Consecutive Trades
2023-2025. Edge computing and CDN infrastructure during an AI-driven attack explosion.
By 2024-2025, AI had weaponized cyberattacks—28 million AI-driven incidents projected globally, with phishing success rates jumping from 12% to 54%. Companies needed better infrastructure protection. Cloudflare was taking market share from legacy players with superior edge computing technology.
I didn’t try to find the perfect entry or hold for a 10-bagger. I traded it three times over 18 months:
Nov-Dec 2023: $64 → $80 = +24%
Feb 2024: $79 → $102 = +29%
Mar-May 2025: $117-130 → $126 = +2%
Total: +54.44%.
That third trade tells the 2025 story. I bought during the March-May tech selloff when Cloudflare crashed from $176 to $120—the same broad market correction that hit all high-growth stocks on bubble fears and trade tension concerns. The stock had run from $80 to $176 in three months. The selloff felt like a reset.
I took a tiny 2% profit in May at $126. Then the stock doubled to $253 by October.
Why didn’t I buy back in June when it broke out? I watched it happen. The rally looked like pure FOMO—no fundamental catalyst, just momentum. I wanted to wait for quarterly results to confirm whether it was real growth or speculation. By the time earnings came, the stock had already ripped 100%. At that point, chasing felt dangerous.
That’s the cost of being defensive in 2025. I was right about the March-May correction. Wrong about it continuing.
Three trades, 54% total. Each one captured a different move. That’s the tradeoff with active management—you get paid multiple times, but you cap the upside on each round trip.
The Counterfactual: What If I Just Held?
Here’s the comparison: what if I’d bought these five stocks on day one and held them through December 31, 2025?
| Stock | First Entry | Entry Price | Dec 31, 2025 | Buy & Hold | My Returns | Difference |
|---|---|---|---|---|---|---|
| Palantir | Oct 25, 2023 | $15.90 | $177.75 | +1,018% | +196% | -822pp |
| PDD | Nov 8, 2023 | $110.07 | $113.39 | +3% | +105% | +102pp ✓ |
| NIO | Feb 1, 2024 | $8.35 | $5.02 | -40% | +32% | +72pp ✓ |
| Meta | Dec 21, 2023 | $350.00 | $660.09 | +89% | +75% | -14pp |
| Cloudflare | Nov 7, 2023 | $64.35 | $197.15 | +206% | +54% | -152pp |
| TOTAL | +1,276% | +463% | -813pp |
Buy-and-hold calculated from first entry date to Dec 31, 2025. My returns represent total realized gains across all round trips on each stock.
Buy-and-hold would have delivered 2.44x more. Palantir alone accounts for 822 percentage points of that difference.
But buy-and-hold also means holding NIO from $8.35 to $5.02 (-40%) and riding PDD’s full roller coaster including the -40% drawdown periods. Active management captured profits on PDD and NIO that buy-and-hold would have given back.
The tradeoff: I locked in gains and rotated defensive in 2025. That protected against downside risk but capped participation in the continued rally. Different approaches, different outcomes.
2024 vs 2025: Same Market, Different Mindset
In 2024, I saw AI infrastructure buildout and went all-in on Palantir. I saw China panic and bought PDD. I saw Meta doubters and bought the dip. Concentrated bets, active management, willingness to take profits and re-enter.
In 2025, I saw something different. Nvidia hit $5 trillion market cap. Palantir traded at 15x revenue. Analysts were building 30-year DCF models to justify valuations. Every bubble indicator I knew was flashing red.
I went defensive. Rotated to 30-40% cash. Built a VIX framework. Bought options for protection. Waited for a crash that hasn’t come yet.
Same market. Same strategies available. Different risk assessment.
The question I’m still working through: wrong thesis, or just early?