ManuLion.eth retweeted
introducing Gemini 4 Hydrogen
this new model brings cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows
start building with Hydrogen via the Gemini API and in AI Studio: aistudio.google.com/prompts/…
Tesla just unveiled Powercore, its first-ever transformer, and it completely changes how data centers get their power
Powercore is a solid-state medium voltage transformer, inverter and breaker, all in one unit. A traditional grid setup spreads those jobs across separate pieces of heavy equipment. Tesla puts them all into a single box
The numbers:
• Peak efficiency of 98.86%
• 7.4 MW / 9 MVA continuously at 34.5 kVac
• Grid-forming support for 10 MVA of peak power
• AC voltage from 10 kV to 35 kV
• DC voltage from 800 to 2000 Vdc
• Up to 25 years warranty
Why it matters:
Transformers are one of the biggest bottlenecks in building new power and data centers. Tesla just removed it. Powercore powers 800 Vdc data centers directly, so they run fully off-grid on Tesla solar and Megablock storage with no years-long wait for a utility connection
It speeds up 1500 to 2000 Vdc solar installations and makes Megablock even better. Tesla now builds the entire chain: generation, storage and the power electronics that connect them
Grid-forming means Powercore doesn't just follow the grid. It creates and stabilizes power on its own, which is exactly what an off-grid site needs
Made in America. Operating 2027
Tesla isn't just a car company or a battery company anymore. It's now also a full energy infrastructure company
ManuLion.eth retweeted
$NVDA deciding not to sell H200 chips to China isn't some groundbreaking event.
Look at the history and context.
🖥️Dec-25, President Trump announced that $NVDA could sell H200s to approved Chinese customers, with 25% of the revenue going to the US government.
🖥️Jul-26 $NVDA confirmed the first H200 shipments to China. These accounted for less than 1% of data-center revenue (then ~$89 billion). The company also took a $400 million inventory charge related to excess H200s, partly tied to weaker China demand.
🖥️Oct-26 Licensed Hopper products (the H200 family) shipped to China reached $2.4 billion, roughly 2% of $NVDA total quarterly revenue.
🖥️Nov-26 Interim CEO Colette Kress stated that $NVDA is no longer shipping licensed Hopper/H200 data-center products to China. The company’s outlook assumed zero China data-center compute revenue going.
Whoopee doo!
ManuLion.eth retweeted
$PLTR Apollo’s time to shine. This ATLAS line from July hits different now:
“Apollo is the layer that makes this multi-model, multi-environment, continuously-patched reality possible. As Al-powered cyber threats accelerate, the organizations that lack Apollo-equivalent infrastructure will be unable to maintain security posture across their software fleet. The gap between organizations with continuous, autonomous security patching and those without will become a gap between organizations that can operate and those that cannot.”
$PLTR ATLAS breakdown on the importance of Palantir’s Apollo product for the threat that models like Anthropic’s Mythos impose in the future:
“The gap between organizations with continuous, autonomous security patching and those without will become a gap between organizations that can operate and those that cannot.”
“In the Mythos era, where Al models can discover and exploit vulnerabilities in minutes, the only viable defense is an infrastructure layer that can detect, patch, and deploy fixes at comparable speed across the entire fleet. Apollo provides exactly this.”
A few random thoughts on $NVDA, their quarter, and what it means for AI, especially $MU and $SNDK ...
1) $NVDA crushing its quarter $115 Billion (vs. expectations of just $108 Billion) and profits of ~$64 Billion ($2.68/share) ARE absolutely mind-blowing! 🤯 Nvidia is officially the most profitable company in the world, making $AAPL's $30 Billion and $MSFT's $36 Billion of profits in the latest quarter look like the kids league. Both Apple & Microsoft are currently valued more than Nvidia
2) $NVDA guided 2027 for 75% growth, which is even more mindblowing! 🤯 If it hits its targets, Nvidia is poised to be tech's highest revenue company, by a mile!
3) $NVDA's board chairman decided to reiterate on this call that the board is not looking to bring back @JensenHuang. IF it turns out that Jensen is not charged with a crime, I think this was a huge error by the board and will cause the board to be replaced within 12 months. I don't think the board will last a major shareholder meeting, which could even happen as a result of a special meeting much sooner. Jensen is extremely popular with both shareholders and employees and I think the board is miscalculating how badly people want him back.
4) $NVDA's quarter AND future guide basically shut down any notion that the AI bubble has popped or is about to pop. Companies like $MU and $SNDK are trading with a forward PE of ~3 to 4 as if armageddon has already happened. Not only has the AI spending not slowed down, but it's increasing, which is extremely good news for AI-related companies. It's especially good news if you happen to make a bunch of critical components (such as memory & storage) that are in extremely high demand for AI, computers, phones, robots, self-driving cars, etc. and you are trading at a PE of just 4 like $MU is. I fully expect $MU to have a stellar quarter when it reports next month!
5) Lastly, in the short-term, I wouldn't be surprised if there is an exodus of talent from $NVDA, which will likely be good for $SPCX, $META and others in the AI space who are trying to design their own chips.
ManuLion.eth retweeted
BREAKING: Nasdaq 100 futures extend gains to +1.3% in after hours trading in reaction to Nvidia's, $NVDA, earnings report.
ManuLion.eth retweeted
BREAKING: Nvidia, $NVDA, raises its revenue growth guidance for FY2028 to +75% year-over-year, up from +70%.
This implies annual revenue of ~$728 billion, up over +1,100% from levels seen just 4 years ago.
That's an average of $182 billion per quarter.
ManuLion.eth retweeted
$NVDA just beat by ~$8B, guided $130B with zero China and still has ~$227B left on the biggest buyback authorization increase ever.
$NVDA CRUSHED THEIR Q3 EARNINGS
• Revenue $115.8B vs Est. $108.7B
• EPS $2.68 vs Est. $2.47
• Data Center $107B vs Est. $102B
• Gross Margin 75% vs. Est. 74%
Q4 Guidance
• Revenue $130B vs Est. $125B
ManuLion.eth retweeted
JUST IN: Nvidia $NVDA surges over 5% after reporting $115.84 billion in revenue for Q3, beating expectations.
ManuLion.eth retweeted
Nvidia, $NVDA, Q3 Earnings Summary:
1. Record quarterly revenue of $115.8 billion, above expectations
2. Record data center revenue of $107.6 billion, above expectations
3. Record net income of $65.3 billion, above expectations
4. Record Q4 revenue guidance of $130 billion, above expectations
5. Announces new $150 billion stock buyback
This is objectively one of the most impressive quarters ever posted by a public company.
The AI Revolution is in full-swing.
ManuLion.eth retweeted
oh look the thing everyone knew was happening is confirmed
The #FBI issued a FLASH advisory warning that cybercriminals and state-aligned hackers are exploiting ARTEX, a super intelligence-powered penetration testing framework, to automate attacks against businesses and organizations worldwide. ARTEX can accelerate intrusions from initial targeting to privileged access in under three hours. It was used in the October ConnectWise supply-chain compromise, which exposed Microsoft 365 customer environments and involved attempted authentication against more than 2,250 downstream tenants.
The FBI is urging businesses and organizations to strengthen access controls, monitor suspicious automated activity, and report potential compromises to ic3.gov.
Read more here: ic3.gov/CSA/2026/261113.pdf
ManuLion.eth retweeted
"I Own My Mistakes": Jensen Huang Breaks Silence, Talks About Return To Nvidia zerohedge.com/political/i-ow…
ManuLion.eth retweeted
This is f*cking insane. Google Brain founder Andrew Ng just dropped a free 3-hour course on becoming an AI engineer in 2026:
"The most valuable skill this year is building with AI. Not chatting with it, building with it
Agentic workflows will drive more progress than the next generation of models"
In 3 hours, he goes from agentic systems to a full prompting course to shipping a working AI app in 30 minutes
Prompts → Agents → Agentic Workflows → A Shipped App → An AI Engineer Job
An agent is a model that plans, uses tools and checks its own work
An agentic workflow chains those agents so the job gets done without you in the loop
Everything else you need is free:
Andrew Ng's Agentic AI course
deeplearning.ai/courses/agen…
AI Python for beginners, taught by Andrew Ng
deeplearning.ai/short-course…
Anthropic's interactive prompt engineering tutorial
github.com/anthropics/prompt…
Anthropic's courses: from your first API call to tool use
github.com/anthropics/course…
Karpathy's Neural Networks: Zero to Hero
karpathy.ai/zero-to-hero.htm…
Practical deep learning for coders
course.fast.ai
Microsoft's 21 lessons on building with generative AI
github.com/microsoft/generat…
Google's machine learning crash course
developers.google.com/machin…
Hugging Face LLM course, start to finish
huggingface.co/learn/llm-cou…
The OpenAI cookbook: code for every common task
cookbook.openai.com
Hundreds of AI apps you can copy and ship
github.com/Shubhamsaboo/awes…
The prompting guide everyone shares
promptingguide.ai
Total: $0
95% will keep paying for bootcamps that teach last year's stack. The 5% who start with this course will ship their first AI app this month
Bookmark this and send it to someone who wants into AI this year
Then read how to wrap that first app in a harness that runs on its own in the guide below ↓
ManuLion.eth retweeted
I have to say, we live in the most interesting of times. We're trying to pick future winner of the AI race.
> Anthropic has the frontier model, momentum -- but rents datacenter from hyperscalers and neoclouds.
> Nvidia controls 90% of the price of most advance AI compute.
> Anthropic/OpenAI/Amazon/Broadcom all racing to build cheaper TPU solution
> Government don't want exports to China, cheering for the Nvidia leadership collapse, and diversify AI compute faster. (AI supply chain actually depends a lot on Asia)
> Opportunists insiders at NVDA and MSM scheme their way to take power the most advanced ai compute ecosystem in the world.
> Politician endorsed by by Anthropic/closed ai lobbyists to tighten open model ai, china.
> Politician want to rein in policy to slow down AI.
> SpaceX is building their own chips, memory, and launching datacenter into space.
> Downstream supply chain become collateral damage.
Who will come on top of the AI race?
Will it be Anthropic/OpenAI, fully integrating TPUs into their model and owning their datacenter one day?
Will it be hyperscaler, developing their own frontier models, leveraging their customer base, and datacenter infrastructure?
Will a lot of future AI compute capacity be in space?
Will TPU win? Will it be the the chip maker who are building deeper connection within the ecosystem.
Will heavy datacenter based business be commoditized like electricity/utility? Is the value chain higher up?
Which Ai company will bankrupt, get acquired, or grow stronger?
ManuLion.eth retweeted
Everyone has drawn this chart in their head and is making a generational bet on which line is true.
🤖 Made with AI
ManuLion.eth retweeted
🚨 Rumors I’ve been hearing. First OpenAI, then Anthropic and Google.
OPENAI
OpenAI is in a “safety lockdown” on its frontier work, tightening safety measures and containment protocols to avoid another Hugging Face incident. Training continues under much stricter controls.
Most gains come from video games and extreme sandboxes where models must complete clear, difficult long-term goals. These force planning, experimentation, recovery from failure, and strategy discovery, driving intelligence and creativity.
Public research roles match this: Synthetic RL uses self-play and simulators; agent teams focus on sustained execution and multi-model coordination; the Recursive Self-Improvement team builds systems that generate hypotheses, run experiments, and automate research.
The internal model is still training but nearing completion. The jump is enormous. narrow superintelligence in specific areas and overall AGI-level capability. Its ability to independently handle difficult work sets it apart.
ANTHROPIC
Model 5 is training and is expected to solve extremely hard problems in mathematics, biology, and other sciences. Anthropic is pushing models toward original research contributions, complex experiments, and sustained reasoning.
Fable 5.5 is Model 4. Fable 6 is Model 5; Opus 6 will be based on Model 5. Internal capabilities will feed the next generation of both families.
Training environments are central. The Universes team creates realistic situations with ambiguity, interruptions, and delayed consequences. The model must track the objective, recognize failing approaches, and find alternatives.
Public roles align: scaling RL across model size, compute, and task duration, plus environments for chip design, verification, and complex visual work. Multi-agent research explores how larger model teams can handle major codebases and difficult mathematics.
Model 5 advances Claude toward owning entire research and engineering workflows. Its role in developing the next model grows more significant.
Anthropic expects RSI March 2027.
GOOGLE
Google is pushing Gemini 5 for a major leap in reasoning, multimodal understanding, and independent long-horizon work. Potentially surpassing all labs by end of year. Carbon topping Artificial Analysis signaled the speed of their frontier progress.
World models are a key advantage. Genie 4 generates interactive environments that evolve with agent actions, enabling vast practice in planning, testing approaches, and learning from outcomes.
Public roles highlight persistent context and personalization: agents that retain useful information, adapt to users, and execute complex workflows without losing track.
Gemini 5 delivers stronger intelligence across text, images, audio, and video. Combined with Google’s infrastructure and simulation work, the scale is enormous.
All three labs are also advancing robotics and world models. intelligence that understands physical environments, predicts action consequences, and operates through real machines.
Meanwhile, their models increasingly contribute to creating their successors. Each improvement accelerates the research behind the next.
The next three months are going to be insane.
In 2014, NASA awarded Boeing $4.2 BILLION to develop Starliner
SpaceX received just $2.6 BILLION for Crew Dragon
Now look at the results:
• Boeing Starliner: Just one crewed flight and got two astronauts stuck on the ISS for over 9 months in space before SpaceX sent Dragon to rescue them
• SpaceX Crew Dragon: 21 crewed missions since 2020, carrying astronauts to orbit for NASA and private customers, saving NASA billions of dollars
Boeing received $1.6 BILLION more in the original NASA contract
Yet SpaceX ended up bringing Boeing's own astronauts back to Earth







































