SpaceX has announced their new Gigasat factory in Bastrop, Texas, which will start producing AI satellites by the end of 2027.
• 1,000+ acres of land owned/under contract
• Over 11 million square feet of building potential
• Solar ingots and wafer, solar cells, and AI sats will also be produced here.
My top 5 best AI stocks to invest in long-term:
These are my biggest positions, and I believe each of these will 10x or more over the next 5 years:
$NBIS: The man who built Russia’s Google, lost it to Putin, and rebuilt from nothing. $46 billion in contracted backlog from Microsoft and Meta. 684% revenue growth year over year. $9.3 billion in cash. ARR guided to $7 to $9 billion by end of 2026.  The fastest growing AI infrastructure company most people still haven’t heard of.
$AAOI: Every AI GPU cluster needs optical transceivers. $AAOI makes them. Revenue went from $210 million in 2023 to $456 million in 2025 to a $1 billion+ guide for 2026. $LITE is now $850. $COHR is at $370.  $AAOI makes the same products, growing faster than both, at a fraction of the valuation.
$CRDO: Revenue up 205% in fiscal 2026. Earnings up 805%. FY2027 guidance projecting 80%+ growth with optical products contributing over $600 million.  622 employees printing $40 billion in market cap. The AI connectivity play nobody is talking about loudly enough.
$MXL: Q1 2026 revenue up 43% year over year. Infrastructure sales surged 136% and became the core business.  Panther V targets AI inference bottlenecks across a $5 billion serviceable market. Benchmark just initiated with a Buy and a $125 price target.  Early innings of a full re-rating.
$MRVL: Jensen Huang called it the next trillion dollar company at Computex. They just joined the S&P on Friday. Q1 FY2027 revenue $2.42 billion, up 28% year over year. FY2028 revenue target of $16.5 billion.  Custom ASICs and optical interconnects. The custom silicon wave has barely started.
Five different companies. One common purpose. All of them are building the infrastructure the AI era runs on.
I am long all five. Not financial advice. Do your own research.
AI is moving faster than copper can handle. That's not a take — that's physics. Nvidia put $4.5B behind the solution.
$LITE (Lumentum) — Makes lasers and optical components that transmit data as light inside data centers replacing slow copper cables.
$AAOI (Applied Optoelectronics) — Builds transceivers that convert electrical signals into light so data travels faster across AI infrastructure.
$AXTI (AXT Inc.) — Grows specialty semiconductor substrates — the raw crystal material that high-speed optical and wireless chips are built on.
$MRVL (Marvell Technology) — Designs custom AI chips and networking silicon for hyperscalers like Amazon and Google — deep Nvidia ecosystem partner.
$CRDO (Credo Technology) — Makes high-speed SerDes chips and cables that connect AI accelerators inside data centers with less power and heat.
$ANET (Arista Networks) — Builds the ethernet switches and networking software running inside the world's largest cloud and AI data centers.
Before the week ends, let's acknowledge one of the most INSANE week ever for open AI, with 25+ notable open-weight drops across every modality:
🧠 LLMs
→ NVIDIA Nemotron 3 Ultra: 550B hybrid Mamba-MoE, only 55B active, 1M context, MMLU 89.1. NVFP4 variant claims ~5x throughput on Blackwell. First openly-weighted 550B hybrid Mamba-Transformer, closing the gap with frontier closed models.
→ Google Gemma 4 12B: fully open dense any-to-any (text/image/audio/video), 256k context, encoder-free, 140+ languages, AIME 2026 at 77.5. Shipped with a 23-checkpoint QAT wave (mobile ONNX + MLX). Most deployable model of the week.
→ StepFun Step-3.7-Flash: 198B sparse MoE VLM, ~11B active, SWE-Bench PRO 56.3. Apache 2.0.
→ Liquid AI LFM2.5-8B-A1B: edge MoE, just 1.5B active, 128k ctx, MATH500 88.8, MLX-ready. Best on-device option this week.
→ JetBrains Mellum2-12B-A2.5B-Thinking: their first open MoE, near-Qwen3-14B coding at 2.5B active. Apache 2.0.
🎨 Image gen (the surprise of the week)
→ Ideogram 4: their FIRST-EVER open weights. 9.3B flow-matching DiT trained from scratch. #2 overall behind GPT Image 2, top open-weight model on Design Arena + LMArena. Strongest open checkpoint for text-rich images, full stop. It has taste. Still can't believe this is open weights.
🔊 Audio & Speech (a breakout week for open TTS, 4 labs shipped)
→ Boson Higgs Audio v3 4B: 102 languages, 21 emotions, singing/whispering/shouting, sub-second TTFA.
→ RedNote dots.tts: the only fully continuous (no codec) open TTS pipeline, Apache 2.0.
→ Google Magenta RealTime 2: real-time music gen, <200ms latency, text+audio+MIDI. multimodalart ported it to PyTorch within hours with live ZeroGPU demos.
→ NVIDIA Nemotron-3.5 ASR: 600M streaming, 17x more concurrent streams vs Parakeet RNNT 1.1B.
👁️ Vision & VLMs
→ PaddleOCR-VL-1.6: SOTA document parsing at 1B params, Apache 2.0.
→ Baidu NAVA: 6.3B joint audio-video gen, best-in-class A/V sync, Apache 2.0.
🎬 Video, 3D & World Models
→ NVIDIA Cosmos3-Super: 64B omnimodal world model coupling action trajectories with video+audio gen, for Physical AI.
→ JD JoyAI-Echo: up to 5-min multi-shot text-to-video on LTX-2.3.
→ ByteDance Bernini-R + VAST TripoSplat (single-image-to-3D Gaussian splats, MIT).
BREAKING: Meta, $META, is considering raising "tens of billions of dollars" in a stock offering as it seeks new sources of capital following the launch of Google’s record $85 billion equity raise this week, per FT.
Bitcoin Capitalist: Bitcoin reaches its full potential by integrating with the global economy: currencies, credit, securities, companies, banks, institutions, governments, families, and individuals. Bitcoin is an open monetary network for everyone.
🚨 BREAKING: A Democrat congresswoman just YELLED at Marco Rubio and then STORMED OUT of the hearing because he destroyed her
RUBIO: "Why is she leaving? I'm gonna answer her questions! Well, THANK YOU FOR COMING!"
We love Marco 😂😂
"Can I answer? Oh, she gets to scream now too? What is this? You ask questions for 5 minutes and I don't get to answer, what is this a DUNK TANK?"
🚨 OMG. SEC. SCOTT BESSENT JUST WENT THERE 🔥
SEN WYDEN (D): We don't want ramblings about the most corrupt regime in history, we want facts
BESSENT: "And we'd like to hear what Adam Wyden and Jeffrey EPSTEIN talked about. Your son's largest investment position was Rick's Cabaret. So, did your son and Jeffrey Epstein talk about pole dancing as he begged him for money using your limited credibility?"
ASSASSIN.
IREN has announced a planned 800MW data center campus in Bundey, South Australia.
This marks IREN’s first announced Australian data center project and one of the largest in the Asia-Pacific region announced to date.
Learn more: https://t.co/3bOYCUG3pk
Tom Lee expects the rally to continue through June, led by software companies and AI beneficiaries.
But from late June to October, he sees a correction that could feel like a bear market.
Why?
• Potential IPOs from SpaceX, OpenAI, and Anthropic could absorb significant liquidity.
• The market will test Kevin Warsh’s inflation framework.
• Midterm election seasonality has historically been a headwind.
Tom compares it to the February-April pullback:
The S&P 500 fell just 9%, but many individual stocks dropped 50%+.
His view: the stocks that have run the hardest are the most vulnerable.
Long term, he’s still extremely bullish:
‘From October, I think that’s when a very strong rally starts.’
🚨 BREAKING
🇺🇸 FED WILL INJECT $6,576,000,000.00 INTO THE MARKETS TOMORROW AT 9 AM ET, RIGHT BEFORE THE MARKET OPEN!
KEVIN WARSH IS OFFICIALLY CONTINUING T-BILL PURCHASES AND TURNING THE MONEY PRINTER BACK ON!
GIGA BULLISH FOR MARKETS!
$IREN co-founder Daniel Roberts said if someone wanted to build a 1GW AI factory TODAY, they likely would not get first compute online until 2030, reinforcing how severe the power bottleneck has become across AI infrastructure.
That is exactly why energy and power infrastructure in $BE and $FPS continue becoming increasingly important to support the expanding compute footprint tied to the builders $NBIS, $WULF, and $APLD.