Ninja

[WSBgod @WSBgod]

We’re in biggest bull market of all time and the most successful man of all time is IPOing his AI space exploration company with 95% shares locked, and yet we STILL can’t pump for more than 2 days??

Doubled down.

[Bilaal- BD investing @bdinvestingg]

7 ETFs Dominating AI, Power, Space & Robotics

$DRAM — Invests in memory chip companies. Bets on the companies storing and moving AI data faster than ever.

$NLR — Invests in nuclear energy. Bets on uranium miners and reactors powering the AI electricity boom.

$NASA — Invests in space companies. Bets on satellites, rockets, and the businesses building the space economy.

$HUMN — Invests in robotics and AI. Bets on companies building the physical robots that will replace human labor.

$EUV — Invests in semiconductor equipment. Bets on the machines that print the world’s most advanced chips.

$SMH — Invests in semiconductors broadly. Bets on the companies designing and manufacturing the chips powering everything.

$GRID — Invests in electricity infrastructure. Bets on the companies upgrading power grids to handle surging energy demand.

[Mike @MikeLongTerm]

$AMD| Fwd P/E is 15-25x, still super cheap ✍️
Based pm $551/share today.
Not Financial Advice! DYOR!

Glad to see more research firms and analysts using my researches/threads.

AMD has been outperforming the broader market, where almost everything selling off.

We will see more research paper coming out on how to train agents better and more efficient. This will be mostly dependent on @AMD CPUs with highest core.

Not too long from now, we will get EPYC Verano next Gen after Venice. AMD will maintain leadership on highest core for 4-5 years+ and most of Inference market.

The stock is extremely cheap on fwd P/E for its growth and potential.

Fwd P/E(12-18 months) is 15x-25x(from Q3 2026)

Revenue is not going to slow after 12-18 months. Even with the most optimistic projection with no delay from TSMC 12 2nm Fabs Ramp, AMD would only get 4-8m EPYC Venice on top of Existing Helios Orders in 2027 where to rebalance 2023-2025 CPU:GPU Ratio would require 40m+ EPYC. The CPU shortage is going to last minimum of 3-5 years.

The drama on Tokenmaxxing is all abt in-house chips @awscloud @googlecloud and $NVDA chips are unable to produce lower token cost, led to @AnthropicAI overcharging enterprises(mostly for coding). The cost between Helios Rack and others is significant.

It is $0.0003-$0.0005/M tokens vs $0.01-$0.5/M Tokens on Inference( Agentic AI)

Token cost will continue to collapse on $AMD Annual Cadance system. That will help drive adoption and increase token consumption. Dont foget 3 smart AI Agents are more effective than 30 average Agents. Reinforcement Learning is all about CPUs.

Recent Upgrade on CPUs TAM to $223B by 2030 is going to change later this year or next year to $500B by 2030. My $500B CPUs TAM by 2030 is based on $2T+ TAM for US AI CapEx. Where the $223B was based on $3.5T TAM for US AI CapEx. If Total TAM is $3.5T by 2030, CPUs TAM would be $0.9-$1T. Total TAM is included CPUs GPUs NPUs ASIC Arm-based & others.

Alright, that is it.

Not Financial Advice! DYOR!

[Antonio Linares @alc2022]

You're not thinking about your mitochondria, bone marrow and thymus gland enough

[1000xSTOCKS @1000xStocks]

5 stocks that could double in 5 years…

$AMD → $544 today. Needs to add around $900B in market cap to double.
Capturing nearly half of x86 server revenue. The AI supercycle is just starting.

$SOFI → $17 today. Needs to add around $22B in market cap to double.
Lowest mathematical hurdle on this list. The stock was trading at $33 less than a year ago.

$NOW → $94 today. Needs to add around $96B in market cap to double.
AI customers spending $1M+ annually grew 130% last year.

$RDDT → $174 today. Needs to add around $33B in market cap to double.
Every AI company needs Reddit's human conversation data to train models.

$PLTR → $123 today. Needs to add around $295B in market cap to double.
43.6% net profit margins. Government and enterprise AI contracts compounding.
The easiest double mathematically maybe?

$SOFI → $22B market cap, already profitable, growing 40% annually.
The hardest?

Which stock would you own for the next 5 years?

[Antonio Linares @alc2022]

Peptides

[Antonio Linares @alc2022]

$PLTR is trading at just a few times 2028 earning assuming AI scaling laws don’t continue accelerating

This stock is extremely cheap

[Yiannis Zourmpanos @yianisz]

I'm actually more confident in $HIMS today than I was six months ago.

do you realize how many catalysts are now stacked on top of each other??

$NVO. $LLY. Eucalyptus. Peptides. Accelerating traffic. Improving engagement. falling short interest..

My view is simple: the bear case worked when growth was decelerating and regulatory risk was rising.

Today the debate is shifting toward whether HIMS is becoming a global health platform.

If execution continues, I think $40 is a waypoint, not a destination. The real battle starts once the stock pushes into the low-$40s, where the remaining high-conviction shorts begin running out of time.

[Antonio Linares @alc2022]

Hot peptide summer

Don’t crack under pressure ;)

[Mike @MikeLongTerm]

$AMD| People that trimmed @AMD or trimming AMD in the last 6-12 months are the people that didnt read @MikeLongTerm threads.

Traders gonna trade. I was told $AMD is overvalued since $10-$20 or $50-$100 Price Target last year from some of the most well respected wall street analysts.

Not Financial Advice!

[Mike Investing @MrMikeInvesting]

Nobody is talking about this stock right now, but soon everyone will be…

$CRWV which currently trades at $118 is setup for a run into $340+ before end of year.

With 167% YoY revenue growth, $3.6B in cash, & multi billion dollar contracts from $NVDA, $META, OpenAI, & more you cannot be bullish enough on $CRWV.

I’ve called this weeks ago now, but the AI infrastructure theme will dominate the second half of 2026.

You can mark my words on that…

[Mike @MikeLongTerm]

$PLTR is no longer expensive 🧵

TTM PEG: 0.47-0.49
2yr PEG : 0.7-0.85
5y PEG 0.3-0.5

All are under 1.0 AKA unvervalued!

1. P/S is collapsing FAST

End of 2021: ~22x
End of 2022: ~7x
End of 2023: ~17x
End of 2024: ~63x
End of 2025: ~89–102x
TTM : 59x
Fwd P/S : 35-40x
FY2027 P/S : 19-21x
Average 5 year P/S is 58x

2. P/E is collapsing FAST

End of 2021: Negative
End of 2022: Negative
End of 2023: ~171–265x
End of 2024: ~378–383x
End of 2025: ~257–416x
TTM : 134-136x
Fwd P/E : 40-60x
FY2027 P/E : 25-35x

Average 5 year P/E is 150-200x

3. Revenue Per Employee

2020: ~$448K (lowest)
2021: ~$528K
2022: ~$497K
2023: ~$596K
2024: ~$728K
2025: ~$1.01M
TTM : $1.3-1.4M
FY2027: $2-$2.5M

This is god-like Operational efficiency, scalability, and productivity!

What will be bears crying abt this time?

Not Financial Advice! DYOR!

[Antonio Linares @alc2022]

“I think some of the wildest stuff I've seen is around biology and medicine.

We're gonna get these huge, enormous benefits.

Life is gonna get better.”

[Mike @MikeLongTerm]

$AMD| @perplexity_ai CEO on Token Value/Watt/User🧵

It is fascinating how Dr. Su saw this 3-4 years ago, while everyone was chasing raw throughput or higher TDP the better. Aravind Srinivas noted people are "tired of tokenmaxxing" (burning huge budgets on frontier model tokens) and wants to push more efficient orchestration, including potentially more on-device/local inference to cut costs and improve privacy.

Dr. Lisa Su was spot on with her long-term bet on power efficiency (TDP), TCO, and inference economics around 2022–2023. Dr. Su highlighted early that inference would become 90-95% of AI compute long-term. She positioned AMD’s stack (EPYC CPUs + Instinct GPUs + unified systems like Helios racks) for exactly this shift.

Aravind Srinivas stresses maximizing economic value from output tokens per watt consumed per user, balancing cost, latency, accuracy..., while moving away from "tokenmaxxing" (over-spending on frontier tokens). @AMD focuses on this by delivering competitive or superior tokens-per-watt and cost-per-token in inference, especially for large models and agentic/orchestrated workloads.

TDP & Power Efficiency: AMD's Instinct MI300X/MI325X/MI355X/MI455X series and EPYC families emphasize high memory and bandwidth, allowing larger models to run efficiently on fewer GPUs without excessive power draw. Benchmarks show strong tokens-per-watt in high-concurrency scenarios (MI355X delivering higher throughput per GPU and better per-user performance than some NVIDIA equivalents at scale).

As Host CPUs for GPU Clusters: EPYC processors ( 5th-gen Turin/9965 and EPYC Venice) boost overall system throughput, reduce latency (TTFT and inter-token), and improve GPU utilization by 3–14%+ vs. competitors like Intel Xeon. This directly lowers effective $/M tokens at scale by maximizing output per watt and per dollar spent on the full node/rack.

This perfectly matches the shift to agentic AI, more CPU-heavy orchestration, hybrid local/server setups, and maximizing value per watt rather than pure GPU tokenmaxxing. EPYC enables efficient hybrid systems (CPU for control plane + GPU for heavy inference), reducing power waste and effective token costs while improving privacy/latency in orchestrated flows.

While GPUs (AMD MI355X/MI455X or NVIDIA) often lead raw throughput for large models, EPYC is a key enabler for the lowest end-to-end inference TCO in real-world agentic deployments, exactly the efficiency layer Srinivas highlights as the long-term winner. Real results depend on workload, quantization, software (ROCm/ZenDNN), and scale, but AMD's integrated CPU+GPU approach targets this sweet spot aggressively.

Not Financial Advice! DYOR!

Source Full Video here: https://t.co/4E9b93KOXW

[Antonio Linares @alc2022]

$PLTR is an IQ test right now

Either you understand that it’s a Singularity Scaler or you don’t

[Antonio Linares @alc2022]

Artificial longevity

[Melvin @MelvinInvests]

Nebius will be the first $1 trillion neocloud company (Save this).

Most people have never heard of Nebius and they should have because it is the fastest growing infrastructure company on the planet right now

AI cloud revenue grew 841% year over year in Q1 2026 to $389.7 million, and the company exited Q1 with an annualized run rate of $1.9 billion that management expects to reach $7 to $9 billion by year end.

For context, t took AWS over five years to grow from $1 billion to $7 billion in annual revenue while Nebius is doing it in a single year.

The neocloud category itself is the key to understanding why Nebius's ceiling is so high.

Neoclouds are clean slate AI cloud providers, built from the ground up for training and inference instead of retrofitting legacy infrastructure like AWS, Google Cloud, or Azure.

This matters because AI workloads are fundamentally different from traditional cloud workloads because they require extremely high bandwidth GPU interconnects, low latency storage, optimized networking fabrics, and custom cooling, none of which legacy hyperscalers were built for.

The neocloud market generated roughly $25 billion in revenue in 2025, up 223% year over year and Synergy Research projects it will approach $400 billion by 2031, compounding at 58% annually, one of the fastest sustained growth rates ever recorded for an infrastructure category of this scale.

And Nebius is the best positioned pure play in that market.

Now look at the Wells Fargo capacity chart.

Nebius was running 172 MW of active capacity at the end of 2025 and by 2030, Wells Fargo projects 7.5 GW, a 43x expansion in five years.

The two flagship buildouts driving that number are enormous in scale and ambition.

The Independence, Missouri campus is a 400-acre AI factory campus approved under Chapter 100 incentives, the largest of its kind in the Midwest scaling to 1.2 GW when fully built.

The Pennsylvania campus, announced in Q1 2026, is an identically sized 1.2 GW site that will begin contributing revenue in the first half of 2027 with 250-300 MW online by year-end.

Both campuses have a combined planned capacity of 2.4 GW and those are just the flagship self-build sites.

The Wells Fargo model has Nebius adding another 3.4 GW of other/unallocated capacity through 2030, which represents future site announcements not yet disclosed.

The financial mechanics of why $1 trillion is achievable are not complicated.

Wells Fargo's model implies that at 7.5 GW of active capacity and using the industry standard of approximately $10 million per MW per year in revenue, Nebius would be generating roughly $75 billion in annualized revenue by 2030.

At a conservative 5x revenue multiple ,well below what infrastructure scale AI businesses command today that is a $375 billion market cap.

At the revenue multiples Nebius currently trades at, which reflect the scarcity of pure play AI infrastructure assets, you reach $1 trillion.

I’m extremely bullish on Nebius.

Follow me @MelvinInvests for more on $NBIS and other underrated AI infrastructure gems.

[Antonio Linares @alc2022]

You can’t not know biology now

[Antonio Linares @alc2022]

Focus on fundamentals $hims investors

[Antonio Linares @alc2022]

Three words

Predictive Biology Ontology