Razim

[Razim @Mostlyy_Human]

The legendary and infamous cost-cutting ideas by CEO Micheal O'Leary, that didn't make it to execution:

1/ To charge £1 to use onboard toilets

2/Reducing toilets from 3 to 1 per plane to add 6 more seats

3/To introduce standing areas for short flights

4/Charging passengers who forgot to print their boarding passes.

5/Asking cabin crew to pay for their training and uniforms

The CEO once said “We might charge for air if we could find a way to do it.”

[Razim @Mostlyy_Human]

Of course there is a catch. Such a battery will be atleast 30% bigger than lithium ion battery and energy density gap is big.

Also it needs higher operational temperatures and atmospheric oxygen is not as pure. Super complex to achieve consistent efficiency.

Still the tech is super nice and niche. It's just that it's not a lithium ion killer. This belongs to a warehouse or underground.

[Razim @Mostlyy_Human]

@PeterSweden7 UK recently hit their lowest army personnel count. They now want a higher strategic reserve.

[Razim @Mostlyy_Human]

NVIDIA have severe shortage of GDDR7 and DRAM and they are redirecting their limited stocks to their most profitable AI Enterprise chips.

Micron also killed its consumer brand and every single one of their wafers are for the data centers.

Higher RAMs are gonna be pricey in the coming years

[Razim @Mostlyy_Human]

This will create a pressure cooker for purpose. Without mandatory labor, we'll confront the void of what's the point of being human if not to build, explore, create?

Boredom becomes the new black death, spiking birth rates in some corners while dooming others to digital opium dens.

[Razim @Mostlyy_Human]

The $500 Billion OpenAI Delusion!

People defending OpenAI’s invincible status are ignoring the brutal reality on the ground in late 2025. A $500B valuation at 38x revenue isn't based on fundamentals but mostly reliant on the nostalgia from 2023.

OpenAI isn't dying but structurally dismantled by competitors playing a different, more efficient game. The strategic advantage of OpenAI has disappeared.

> OpenAI Got Smoked on Benchmarks. Google’s Gemini 3 Pro dropped and scored nearly twice as high as GPT-5 Pro on ARC-AGI 2. xAI’s Grok-4 is beating them on coding. The "gold standard" remains only for their older o3 models.

> Broken Unit Economics as OpenAI’s $8.5B annual burn rate is unsustainable. They are dependent on expensive NVIDIA H100s, while others like Google is scaling Gemini on custom TPUs for a fraction of the energy cost. They cannot win a long-term compute war with this disadvantage.

> Data Starvation is the killer. Google has infinite YouTube/Search data. xAI has the real-time Tesla/X feed. OpenAI has only synthetic data and shaky partnerships. Frontier LLMs without unique data pipes depreciate instantly.

> Importantly OpenAI can't build the future when half of their top researchers shifted to DeepMind in 2025. This is hemorrhaging their ability to ship faster than the competition.

Microsoft is their only real lifeline. Unless GPT-6 (Q2 '26?) is an absolute miracle in on-device reasoning, the "independent AGI company" dream is over. Expect a massive down-round or a total absorption by MSFT by 2027.

[Razim @Mostlyy_Human]

@unusual_whales The biggest hidden driver is Reinsurance. Insurance for insurance companies which surged heavily after the California wildfires, hurricanes in Florida etc.

[Razim @Mostlyy_Human]

To reach Type II civilization, we have to burn Type 0 fuels first.

xAI’s Colossus is already leaning on gas turbines. The grid simply can’t scale fast enough for Gigawatt-class clusters.

This is why Tesla is deploying Megapacks at grid-scale to stabilize this transition.

We will see a massive pivot to nuclear (baseload) and solar-in-space (for compute) within 5 years.

[Razim @Mostlyy_Human]

@denatalguy1975 @SawyerMerritt @boringcompany Basically a giant RC car.

[Razim @Mostlyy_Human]

@cb_doge @grok Investigate and report what really happened

[Razim @Mostlyy_Human]

@amitisinvesting Meta buying from Google proves that for Big Tech, Unit Economics is more important than rivalry. They will do anything to escape the Nvidia tax.
The fracture is happening live https://t.co/uv67W5nJZF"

[Razim @Mostlyy_Human]

The $NVDA Tax is Ending 📉 Custom Chips Are the AI Endgame!

NVIDIA's dominance is cracking with Google's TPU Edge and xAI's TeraFab Bet.

The hard truth is that if you don’t own the silicon, you don’t own your margins.

OpenAI's GPT series is bleeding cash on NVIDIA H100s and is spending billions on NVIDIA hardware, with no quick escape route.

The giants are quietly leaving the ecosystem.

> xAI bridges the gap with NVIDIA's Colossus cluster today but eyes full independence via TeraFab, a massive fab aiming to produce chips at 1/10th the cost.

> Google sidesteps NVIDIA's high costs by training Gemini entirely on in-house TPUs, saving 50-70% on compute expenses compared to GPU rivals.

> Broader trend we see is that Meta, Amazon, and Microsoft are racing to custom silicon, signaling a shift from NVIDIA dominance as AI power demands skyrocket.

[Razim @Mostlyy_Human]

@barrese_chris 100%. And that Tier 2 is exactly where Nvidia stays dominant. Nvidia loses the Hyperscalers but keeps the long tail of the entire global startup ecosystem.

The rest who can't afford to write custom kernels or build compilers, need CUDA to just work.

[Razim @Mostlyy_Human]

The $NVDA Tax is Ending 📉 Custom Chips Are the AI Endgame!

NVIDIA's dominance is cracking with Google's TPU Edge and xAI's TeraFab Bet.

The hard truth is that if you don’t own the silicon, you don’t own your margins.

OpenAI's GPT series is bleeding cash on NVIDIA H100s and is spending billions on NVIDIA hardware, with no quick escape route.

The giants are quietly leaving the ecosystem.

> xAI bridges the gap with NVIDIA's Colossus cluster today but eyes full independence via TeraFab, a massive fab aiming to produce chips at 1/10th the cost.

> Google sidesteps NVIDIA's high costs by training Gemini entirely on in-house TPUs, saving 50-70% on compute expenses compared to GPU rivals.

> Broader trend we see is that Meta, Amazon, and Microsoft are racing to custom silicon, signaling a shift from NVIDIA dominance as AI power demands skyrocket.

[Razim @Mostlyy_Human]

@ShafronTom @ApoStructura I think the infrastructure needed to power the magnetic shield will be not worth enough to try it. The current regular shielding blocks up to 80% of moderate radiation dose.

[Razim @Mostlyy_Human]

@elonmusk @MichaelAArouet Moving AI to Space will solve the 300GW problem and only SpaceX can do it first. OpenAI wants 250GW of power by 2035. The entire US grid is only 500GW. Mathematically, the grid breaks.
https://t.co/l9pUIZmgwx

[Razim @Mostlyy_Human]

The US consumes about 4,000 terawatt-hours (TWh) of electricity annually & 100 GW would represent almost 22% of that US total.

1/ Starlink V3 satellites are massively scaled for compute. Estimated power output is 10-50 kW per asset (5 x V2 version output)

2/ Each V3 could host Tesla Dojo/AI chips similar to Tesla's supercomputers. A single satellite could pack 1-10 petaflops of compute networked into clusters.

3/ High-speed optical links ,up to 1 Tbps throughput, connect satellites, creating a mesh for real-time AI inference/training across the constellation.

4/ SpaceX will launch 10,000+ V3s over time, potentially generating 100+ GW collectively, enabled by Starship.

[Razim @Mostlyy_Human]

@aelluswamy @Tesla_AI The tight coupling of SW/HW is the ultimate moat. General purpose hardware (Nvidia) hits a ceiling that purpose-built silicon smashes through, especially at 1ms latency.
https://t.co/ubCgvQ2jLO

[Razim @Mostlyy_Human]

The NVIDIA Tax is Ending 📉 Custom Chips Are the AI Endgame!

NVIDIA's dominance is cracking with Google's TPU Edge and xAI's TeraFab Bet.

The hard truth is that if you don’t own the silicon, you don’t own your margins.

OpenAI's GPT series is bleeding cash on NVIDIA H100s and is spending billions on NVIDIA hardware, with no quick escape route.

The giants are quietly leaving the ecosystem.

> xAI bridges the gap with NVIDIA's Colossus cluster today but eyes full independence via TeraFab, a massive fab aiming to produce chips at 1/10th the cost.

> Google sidesteps NVIDIA's high costs by training Gemini entirely on in-house TPUs, saving 50-70% on compute expenses compared to GPU rivals.

> Broader trend we see is that Meta, Amazon, and Microsoft are racing to custom silicon, signaling a shift from NVIDIA dominance as AI power demands skyrocket.

[Razim @Mostlyy_Human]

@XFreeze Until Blue Origin can reach the infrastructure to match the launch cadence, Elon has got the ball.
https://t.co/ysuHuwbcl6

[Razim @Mostlyy_Human]

@XFreeze Biggest hurdle is the launch capacity which ofcourse Elon can achieve with Starship.

The system would require around 500,000 tons to LEO. That means its that’s 3,300 falcon launches per yearr. and when its Starship it will be 1700 flights per year.

https://t.co/l9pUIZlIGZ

[Razim @Mostlyy_Human]

@XFreeze The Terafab is key to solving production at exascale for space AI.

xAI might tap Tesla's designs down the line, hinting at convergence between his companies.

Also XAI in the future won't be depended on NVIDIA chips and pay the NVIDIA tax , same like Google.

[Razim @Mostlyy_Human]

Everyone circling Elon for a Starship slot like it’s the last lifeboat out of Titanic, meanwhile Jeff is just trying to convince the group that Blue Origin is the Lyft to SpaceX's Uber.

Jeff getting the realization that he needs to 10x Elon's launch cadence or get left back on Earth.