This page tracks every stock Couch Investor has discussed on YouTube, extracted from AI transcription of their 23 most recent videos covering 36 tickers. They are currently bullish on SOFI, NU, MELI, DLO, NBIS, NVDA and others. Bearish on ORCL, MDB. Each take below includes their stated reasoning and, where given, a price target.
Meta's first negative FCF since 2012 stems from massive AI CapEx ($135-145B 2026 guide), fueling debate over whether AI spending is a money printer or an expensive habit; creator stays neutral, awaiting proof AI can run 10x cheaper.
| Ticker | Action | Why | Target | When |
|---|---|---|---|---|
SOFI●●● | add | Down ~50% despite a history of beats and 35%→25%→20% guided growth; PEG 0.5, DCF base case $22.8 (~51% upside). Will add if management assures targets hold despite high rates. | $22.8 | today |
NBIS●●● | buy | His poster child for new AI cloud players: triple-digit growth, a disciplined execution-focused CEO, and plans to bring 1GW of Vera Rubin capacity online in 2027. | — | today |
META●●● | add | Strong quarter (26% revenue growth, AI-boosted ad pricing, huge Muse adoption, shrinking Reality Labs losses) yet still undervalued under $700; the headline-driven drop isn't a business problem. | — | today |
GOOGL●●● | buy | Google is overlooked again: Cloud is capacity-constrained and could grow even faster, the TPU business should explode next year, and a cross-product agent is coming — a very big future winner. | — | today |
NBIS●●● | buy | $40B+ customer commitments, a gigawatt of Vera Rubin capacity ramping through 2027, and $28–40B exit-ARR scenarios vs a $40–43B market cap point to severe undervaluation, with margins improving at scale. | — | today |
META●●● | buy | Core ads are a 'juggernaut' (impressions +10%, price per ad +14%, headcount -15%) and Muse shows strong product-market fit; ~$700 still looks cheap despite the ~$10B legal-charge EPS miss and the Class C split seen as setting up future dilution. | — | 5d ago |
SE●●● | buy | Oversold after a 31% YTD drop despite ~35-40% revenue growth, improving profitability, and consistent revenue beats; cheap at 21x forward P/E and 12x forward P/FCF, and run much better than 3-4 years ago. | — | 6d ago |
GOOGL●●● | buy | Still undervalued: super-profitable core with a big cash pile, Google Cloud grew ~82% with expanding margins (100% growth possible), and it's a name that lets you sleep well through a correction. | — | 6d ago |
NBIS●●● | add | Pullback is overblown: revenue accelerating from $1.3B LTM toward ~$23B by FY28 with ~$3B operating profit, best-in-class fundraising/buildouts, and rising GPU spot prices help unsold capacity. He bought more last week. | — | 6d ago |
AXON●●● | add | Unrelated to the AI narrative turmoil; he likes the business and is using the market dip as a chance to buy more shares. | — | 1w ago |
NBIS●●● | add | AI neocloud exposure he already owns; he will add more if the stock dips back under $200 amid the selloff. | — | 1w ago |
AVGO●●● | add | Existing AI-related holding he expects to size up at cheaper prices during the headline-driven drawdown. | — | 1w ago |
META●●● | buy | Still undervalued at 22.4x forward earnings (PEG 1.1) with a strong, profitable core business; the Muse release changed the AI narrative and he expects Meta to monetize its ecosystem by offering models for others to build on. | — | 2w ago |
MELI●●● | add | One of the best-run businesses on the planet—an 'Amazon without AWS' for Latin America—growing 30%+ for ~30 straight quarters; current margin compression is just another investment cycle that has always rebounded. | — | 2w ago |
NU●●● | add | Cheap at ~16x forward earnings with 50% revenue growth, cheap customer acquisition and rising ARPU; proven management is expanding into the US and globally, where even small wins are huge. Monzo acquisition rumor was denied. | — | 2w ago |
Down ~50% despite a history of beats and 35%→25%→20% guided growth; PEG 0.5, DCF base case $22.8 (~51% upside). Will add if management assures targets hold despite high rates.
Price target: $22.8
Lula's re-election triggered a ~22% pre-market drop, but NU has executed through many LatAm political cycles; the selloff doesn't change his long-term view.
Same Lula-driven selloff; MELI has navigated repeated South American political cycles and keeps executing, so his view is unchanged.
Down on the Brazil election result, but like NU/MELI it has executed through LatAm cycles; view unchanged, just let the overreaction cool first.
World-class company and management that keeps executing; Musk's neocloud bashing on X is self-serving for SpaceX and nothing new.
Data center revenue more than doubled last quarter, GPU prices are rising with demand exceeding supply, and it generates substantial free cash flow — the sell-off defies fundamentals.
AI revenue up 221% and expected to double twice to ~$230B within two years, with tens of billions in FCF ahead — fundamentals fully on track.
Revenue up 41.2% with Q4 guidance of ~58% YoY growth at the midpoint — exceptional results while the stock is punished with the sector.
One of the hyperscalers he expects to do quite well on the ~$2.3T contracted AI demand, with FCF expected to stay positive; reports next week.
Expected to do quite well despite negative FCF from huge capex; Pichai says demand exceeds supply and they are leaving money on the table.
Among the big-four hyperscalers he thinks will do quite well on contracted AI demand, though FCF is currently negative; reports next week.
Down nearly 50% from 52-week highs despite a semi sector beating expectations and growing fast — part of the unfairly punished semi trade.
Down almost 50% from 52-week highs while the semi sector beats expectations — he sees the drawdown as disconnected from fundamentals.
Oversold after a 31% YTD drop despite ~35-40% revenue growth, improving profitability, and consistent revenue beats; cheap at 21x forward P/E and 12x forward P/FCF, and run much better than 3-4 years ago.
Tiny $1.5B space/defense name with huge TAM and partnerships (Palantir, Anduril); triple-digit revenue growth toward ~$1B by FY28, but execution is key and profitability is far off. He holds a small position.
Lots of upside per his DCF; generates strong free cash flow today and, despite autonomous-vehicle fears, he still thinks Uber is a winner.
Unrelated to the AI narrative turmoil; he likes the business and is using the market dip as a chance to buy more shares.
Beat across the board with FY gross margin raised to ~56% vs initial 51-53% guide; demand is sold out into 2028-2030 and planned +30% capacity hikes in '27/'28 still can't meet it, though the stock isn't cheap.
Undervalued per his DCF (PEG 0.7, forward P/E 26) with rising ARPU and improving margins, but he holds only a starter position and waits for AI data-licensing news and monetization catalysts before adding.
Different profile from capex-heavy peers: spends little on AI infrastructure yet keeps growing steadily.
$2B company partnering with Anduril on advanced weapons and propulsion. CEO targets $10B enterprise value by early 2030s, representing a potential 5x. Comparable to early Rocket Lab opportunity.
Used as comparison for Voyager; creator previously saw path to $100B market cap given launch, space systems, and applications layer progress.
The weak link among hyperscalers in his view — an insane amount of debt and no free cash flow make its share of the AI backlog riskier.
CEO departed for Meta's Enterprise Platform, causing a 17-20% stock drop. Loss of key leadership signals uncertainty.
Negative FCF (-$2.1B) despite 26% revenue growth as AI CapEx balloons; creator undecided whether Meta is the next money printer or a very expensive habit.
Stock already under heavy pressure; it simply has to deliver on its guidance, and over-delivering would be even better. No strong view until results.
Solid print (gross margin beat, FSD attach >75%, robotaxi miles scaling) but no material near-term catalysts; panelists mostly sold or never owned, and Optimus customer sales are seen as unlikely before 2028.
Triple beat with a Red Hat rebound and raised free cash flow guide; software strength is lifting SaaS names after hours.
Beat on EPS and revenue; together with IBM's print it is giving SaaS/software names an after-hours boost.
Loves the business—cloud revenue +830%, already profitable, unmatched speed—but guidance implies hundreds of billions in capex and numbers too big to underwrite, so it stays out of his portfolio.
Profitable, cheap and undervalued with pricing power, but growth is decelerating and the 2026 ad target was cut; holding the remaining half until 2027 guidance shows reacceleration.
Memory names had a tremendous run; he questions whether the market gives them another leg up—cheap on paper can stay cheap if momentum fades.
At $35 vs his $30 fair value it's ~15% overvalued on conservative assumptions; the bull case (+37%) gets 30% weight, so it's wait-and-see on re-acceleration.
Price target: $30
~22% overvalued even on his aggressive assumptions, but Palantir always upside-surprises—another 100% growth print could flip the math.
Likes the AI/cybersecurity tailwinds and expects beats, but at ~$25B cap, 250x forward earnings and PEG ~12, success is priced to perfection; he took profits and will only add on a big pullback or faster profitability.
Believes AMD's growth is underestimated with a massive FCF ramp ahead, but after a 184% YTD run the price is too high to add; he cut ~25-30% and would rather park money in Nvidia or Broadcom until acceleration signs appear.
Meta posted its first negative free cash flow quarter since 2012 (-$2.1B) despite 26% revenue growth to $64.6B, as CapEx hit $34.9B and 2026 CapEx is guided to $135-145B. The creator weighs whether Meta is a money printer or a very expensive AI-spending habit, citing the $9.9B legal charge crushing margins, the new Class C stock for capital raising, and a Q4 guide of $73-76B.
Ahead of SoFi's earnings, the creator calls the stock extremely undervalued after a ~50% drawdown (DCF base case $22.8, ~51% upside) and plans to buy more shares if management assures it can still hit targets in the high-rate environment. He dismisses the Lula-driven selloff in NU, MELI and DLO and Musk-driven neocloud fears as sentiment overreactions, staying bullish on Nebius while CoreWeave simply needs to deliver on guidance.
The creator argues the AI-semiconductor pullback is an overreaction: NVIDIA, Broadcom and TSMC are beating expectations with demand still exceeding supply, and valuations are nowhere near dot-com or 2021-22 extremes. He is bullish long-term, calling Nebius the poster child of the new AI cloud players, but refuses to call the bottom. He suggests either starting to look at opportunities now or waiting for next week's FOMC meeting and mega-cap earnings for clarity.
The creator argues the AI/growth selloff on Middle East headlines is illogical and creates buying opportunities, staying 'very bullish on this AI trade.' He calls Meta undervalued under $700 after a strong quarter and will add on weakness, sees Google as overlooked with Cloud/TPU upside, and frames Nebius as potentially severely undervalued on 2027 exit-ARR scenarios.
Couch Investor frames Nebius as an AI 'token factory' monetizing GPUs two ways—renting out H100 capacity and selling tokens—citing a $3B run rate, ~50% adjusted EBITDA margin, and ~22-month deal paybacks. He argues the money is in keeping GPUs busy rather than owning them, and leaves open whether renting compute or selling tokens wins long term.
Couch Investor relays Nebius CEO comments that AI demand is not a constraint, capital is a solvable task despite temporarily closed debt markets, and chips could become the next bottleneck next year. The creator hints the market may be mispricing Nebius with the question 'Has the market gotten it all wrong?' No explicit buy or sell call is made.
Live Q3 earnings coverage: the Meta-heavy panel calls META a cheap juggernaut despite a legal-charge EPS miss, cheering Muse adoption and the new Class C split (seen as a prelude to a non-voting equity raise), while staying on the sidelines with Tesla despite its margin beat. IBM and SAP double-beats lift SaaS names after hours, and the hosts expect the AI capex arms race to surprise higher.
The creator argues SpaceX's blowout quarter and massive guidance lifted AI cloud names and may have saved the AI trade, though he won't own SpaceX because the numbers are too big to underwrite. He's constructive on fast-executing neoclouds like Nebius, more cautious on CoreWeave's pace, and is holding his remaining half Netflix position while waiting for growth to reaccelerate.
Earnings-night live show: SpaceX delivers a blowout Q3 (revenue +161%, first GAAP operating profit) while Netflix misses on Q4 guidance and cuts its ad-revenue outlook. The hosts respect both prints but won't pay SpaceX's ~$2T premium, preferring Nebius and NVIDIA as cheaper AI-compute exposure, and aren't buying Netflix despite the low multiple. They call the recent neocloud/AI sell-off confusing since compute demand hasn't changed.
Ahead of a pivotal, AI-focused earnings season in a choppy market he calls 'not healthy,' Couch Investor highlights four names he sees as fairly cheap before they report: Sea Limited, Alphabet, Voyager Technologies, and Nebius. He is bullish on Sea and Alphabet as oversold quality, is adding to Nebius despite its drawdown, and frames small-cap Voyager as an asymmetric watchlist bet. He warns many stocks are already in bear markets despite bull-market optics, but says the recent pullback means names are no longer priced to perfection.
Based on AI analysis of this creator's recent videos. Theses are paraphrases of the creator's own words, not verbatim quotes. Not financial advice.