AI Update: NVIDIA’s Record Revenue, Elon’s Data Centers in Space & Gemini 3’s Insane Performance
The podcast discusses Nvidia's record-breaking revenue growth in AI compute, potential competitors like Google's TPUs, and the long-term infrastructure buildout and singularity. The podcast discusses the formation of AI power blocks, the importance of ethical considerations in tech, and the trend towards vertical integration among tech companies. Discussion on the AI market, the shift in computing form factors, NVIDIA's business strategy, and the establishment of AI data centers in Saudi Arabia. The podcast discusses the use of Nvidia chips for AI, particularly XAI, and Elon's plans for space-based data centers and mining the moon for silicon and solar power, emphasizing the rapid advancement and potential economic impact of these technologies. Discussion on energy production, including the gap between the US and China, the need for Project Warp Speed, predictions on fusion and AI energy
Talk about mining the moon for resources to support space-based data centers and the vision of launching 100 terawatts of capacity per year from the moon.
Predictions made
openAWG: Nvidia's AI compute capex party can continue indefinitely if automation and new markets continue to generate revenue.
“I would be mildly, if not significantly, shocked if 10 years from now, conservatively, we don't see low hundreds, if not many hundreds of gigawatts of AI compute not on the Earth's surface.”
Your call:
openAlex: The long-term trend line for energy production is towards fusion and other advanced technologies, barring any shocking discoveries.
“I think this is going to happen with time. Barring some shocking ontologically shocking discovery about the nature of our universe, it seems like we're on trend.”
Your call:
openPeter Diamandis: There will be a 5 to 7-year difficult period in energy production due to the gap between current and future energy technologies.
“AI is about to become the biggest consumer electricity in the US. It's going to be bigger than steel, bigger than crypto, bigger than every industry combined.”
Your call:
openAlex: We're going to speedrun the Star Trek universe in the next decade.
“I think in the intermediate term, we'll see drone delivery and all of these masses start to to to move around in Earth's atmosphere.”
Your call:
Numbers that matter
57 billionNvidia's revenue growth.
62%Year-on-year growth in Nvidia's revenue.
10 gawattsAI infrastructure deals announced by Nvidia.
$65 billionExpected revenue for Nvidia next quarter.
$15 billionInvestment from Microsoft and Nvidia in Anthropic.
$30 billionInvestment from Enthropic on Microsoft Azure cloud.
$15 billion plus another $30 billionTotal investment in Anthropic from Microsoft and Enthropic.
460 gawattAverage annual electricity consumption in the United States
16 to 45 billion dollarsSamsung manufacturing contract for Nvidia chips
100 terowattsEstimated energy output from mining the moon for silicon and solar power, five times Earth's total energy output
15 yearsExpected timeline for robotics refactoring construction, considered long compared to the space-based computing project
100 gigawatts per year of capacity being launched in the next decade or soEquivalent to 100 nuclear power plants, each typically one gigawatt capacity.
China has gone all out on energy production.Compared to the US.
Fusion is 4% efficient, while fision is less than 1% efficient.Comparison of energy efficiency between fusion and fision.
The gap between 2030 and 2035 in energy production is 15 years.Predicted timeline for fusion energy coming online and existing fision plants coming back online.