Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

AI is fueling a massive wave of spending on chips, data centers and power. Ahead of #BloombergTech, Ed Ludlow explains why training advanced models has become so expensive. Conversations kick off tomorrow at 7 PM PT ⏯️

Bloomberg Live

59,488 subscribers

15,043 views • 3 months ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Will AI data centers become more consolidated — or more distributed? jsun, Founder of FLock.io, sees inference pushing compute outward. Unlike large-scale training, many inference workloads don’t require massive, tightly connected clusters. That could make smaller sites with underused power viable as “mini data centers.” Desmond Lim, Co-founder & CEO of Workstream, sees consolidation happening at a different layer. AI remains an intensely capital- and resource-heavy game, requiring talent, funding, chips and infrastructure at scale. In his view, those advantages compound — meaning the biggest players may only get bigger.

Will AI data centers become more consolidated — or more distributed? jsun, Founder of FLock.io, sees inference pushing compute outward. Unlike large-scale training, many inference workloads don’t require massive, tightly connected clusters. That could make smaller sites with underused power viable as “mini data centers.” Desmond Lim, Co-founder & CEO of Workstream, sees consolidation happening at a different layer. AI remains an intensely capital- and resource-heavy game, requiring talent, funding, chips and infrastructure at scale. In his view, those advantages compound — meaning the biggest players may only get bigger.

YZi Labs

27,123 views • 3 days ago

The CEO of $ARM explains where the bottlenecks for AI data centers are right now: - power & power regulation, - $TSM "which is the only game in town" for 3nm and 2nm - HBM memory There is a reason why $TSM, $MU, and SK Hynix are a significant part of my portfolio.

The CEO of $ARM explains where the bottlenecks for AI data centers are right now: - power & power regulation, - $TSM "which is the only game in town" for 3nm and 2nm - HBM memory There is a reason why $TSM, $MU, and SK Hynix are a significant part of my portfolio.

Rihard Jarc

94,253 views • 1 year ago

🚨 WE’RE NOT BULLISH ENOUGH! 56% of my portfolio is in $IREN, $CIFR, $NBIS, $TE = AI/HPC, GRID & ENERGY. Spend 45 seconds watching this clip, and you’ll know why. 1. $CIFR has an elite team of ex-hyperscaler specialists from $GOOG, $META, and $AAPL optimizing every site. 2. The next couple of years will see more acquisitions, more sites, more deals in AI data centers. 3. $IREN secured a $9.7B AI cloud contract with $MSFT. $CIFR inked a $5.5B lease with $AMZN AWS and has a stake from $GOOG. 4. $NBIS landed $3B with $META and multi-billion with $MSFT. 5. J.P. Morgan forecasts AI infrastructure capex for data centers hitting $1.4 trillion annually by 2030, tripling near-term levels. 6. Jon Gray, President and COO of Blackstone Inc., stated in his latest presentation: “It starts with the picks and shovels, with things like chips, data centers, and power at massive scale. We have really focused on the right two circles here.” (Refering to data centers and power) 7. $NVDA’s Jensen Huang: “We wish we had more compute energy.” 8. The current administration pushes AI, energy, and defense with plans for massive power plants to fuel the AI boom. Onwards, -BP Not financial advice. Turn on notifications 🔔 for alerts on new posts.

🚨 WE’RE NOT BULLISH ENOUGH! 56% of my portfolio is in $IREN, $CIFR, $NBIS, $TE = AI/HPC, GRID & ENERGY. Spend 45 seconds watching this clip, and you’ll know why. 1. $CIFR has an elite team of ex-hyperscaler specialists from $GOOG, $META, and $AAPL optimizing every site. 2. The next couple of years will see more acquisitions, more sites, more deals in AI data centers. 3. $IREN secured a $9.7B AI cloud contract with $MSFT. $CIFR inked a $5.5B lease with $AMZN AWS and has a stake from $GOOG. 4. $NBIS landed $3B with $META and multi-billion with $MSFT. 5. J.P. Morgan forecasts AI infrastructure capex for data centers hitting $1.4 trillion annually by 2030, tripling near-term levels. 6. Jon Gray, President and COO of Blackstone Inc., stated in his latest presentation: “It starts with the picks and shovels, with things like chips, data centers, and power at massive scale. We have really focused on the right two circles here.” (Refering to data centers and power) 7. $NVDA’s Jensen Huang: “We wish we had more compute energy.” 8. The current administration pushes AI, energy, and defense with plans for massive power plants to fuel the AI boom. Onwards, -BP Not financial advice. Turn on notifications 🔔 for alerts on new posts.

Black Panther Capital

58,387 views • 7 months ago

South Korea’s stock market has nearly tripled since the start of 2025. This has caused a feeding frenzy among retail investors trying to ride the AI wave. But the financial products many are buying are incredibly risky investments. Ethan Wu, The Economist’s Asia business and finance editor, explains why South Korea’s market has become a casino. Read how SK Hynix became the king of advanced memory chips:

South Korea’s stock market has nearly tripled since the start of 2025. This has caused a feeding frenzy among retail investors trying to ride the AI wave. But the financial products many are buying are incredibly risky investments. Ethan Wu, The Economist’s Asia business and finance editor, explains why South Korea’s market has become a casino. Read how SK Hynix became the king of advanced memory chips:

The Economist

44,887 views • 1 month ago

China’s objective isn’t complicated: They want America to shut down every data center so we instead have to rely on Chinese data centers, Chinese AI, and the Chinese economy. Why? Because China wants to control us! We break down what’s happening in the wake of a report of the Chinese spending big bucks to make you hate data centers. All that and more on the latest episode of Verdict:

China’s objective isn’t complicated: They want America to shut down every data center so we instead have to rely on Chinese data centers, Chinese AI, and the Chinese economy. Why? Because China wants to control us! We break down what’s happening in the wake of a report of the Chinese spending big bucks to make you hate data centers. All that and more on the latest episode of Verdict:

Ted Cruz

78,482 views • 13 days ago

$Power is 5-7% of data center revenues. Some geothermal and nuclear stocks are trading at 50-70x earnings to supply it. Jim Chanos on iConnections: the AI halo has reached its furthest extension into alternative energy. The bull case is that data centers need power and there's a shortage. Chanos's counter: the US doesn't have a power shortage. It has a permitting bottleneck and turbine delivery delays. Both resolve in 2-3 years. When the permitting clears, the scarcity narrative collapses. And even if these companies win a power contract - power is 5-7% of a data center's revenues. Companies at 30-40x EBITDA for potentially supplying a fraction of a 5-7% cost line are priced for a constraint that is transient. Full breakdown of the alternative energy overhang: Source: iConnections -$

Power is 5-7% of data center revenues. Some geothermal and nuclear stocks are trading at 50-70x earnings to supply it. Jim Chanos on iConnections: the AI halo has reached its furthest extension into alternative energy. The bull case is that data centers need power and there's a shortage. Chanos's counter: the US doesn't have a power shortage. It has a permitting bottleneck and turbine delivery delays. Both resolve in 2-3 years. When the permitting clears, the scarcity narrative collapses. And even if these companies win a power contract - power is 5-7% of a data center's revenues. Companies at 30-40x EBITDA for potentially supplying a fraction of a 5-7% cost line are priced for a constraint that is transient. Full breakdown of the alternative energy overhang: Source: iConnections -

Podcast Alpha

75,076 views • 2 months ago

"AI is a leap into the future.” From helping our engineers work differently to supporting the growing power needs of data centers, AI is creating new opportunities. Watch Patrick Pouyanné share his perspective on AI, innovation and electricity at VivaTech.

"AI is a leap into the future.” From helping our engineers work differently to supporting the growing power needs of data centers, AI is creating new opportunities. Watch Patrick Pouyanné share his perspective on AI, innovation and electricity at VivaTech.

TotalEnergies

303,528 views • 2 months ago

As Larry Ellison explains, there are two types of AI models: real-time models, also called AI on the edge, where you need ultra-low latency decisions (think autonomous driving or robotics), and models that don’t need an immediate response, like ChatGPT, where the system has time to reason. Real-time models require local compute inside the device, and these chips have nothing to do with the GPUs running in data centers. The AI-on-the-edge market is still largely untapped — and it’s going to be a multi-trillion-dollar opportunity. I believe $AMD will be one of the biggest beneficiaries

As Larry Ellison explains, there are two types of AI models: real-time models, also called AI on the edge, where you need ultra-low latency decisions (think autonomous driving or robotics), and models that don’t need an immediate response, like ChatGPT, where the system has time to reason. Real-time models require local compute inside the device, and these chips have nothing to do with the GPUs running in data centers. The AI-on-the-edge market is still largely untapped — and it’s going to be a multi-trillion-dollar opportunity. I believe $AMD will be one of the biggest beneficiaries

Daniel

151,698 views • 9 months ago

China is our adversary. They are building more power than we are, and then using it to power data centers to advance both economic interests and of course military as well. We can’t let the Chinese beat us in AI. We’ve gotta build more data centers, we’ve gotta build more power. We’ve got a lot ahead of us here because they’re beating us on this. We cannot let this happen.

China is our adversary. They are building more power than we are, and then using it to power data centers to advance both economic interests and of course military as well. We can’t let the Chinese beat us in AI. We’ve gotta build more data centers, we’ve gotta build more power. We’ve got a lot ahead of us here because they’re beating us on this. We cannot let this happen.

Kevin O'Leary aka Mr. Wonderful

19,405 views • 4 months ago

Is uranium the future of AI? As demand for cloud computing and AI models continues to explode, the real bottleneck is becoming power. That’s what led me to start looking at nuclear energy and uranium. Small modular reactors could become a major part of the solution, and if AI keeps demanding more electricity, the commodities that power that infrastructure could become just as important as the chips themselves.

Is uranium the future of AI? As demand for cloud computing and AI models continues to explode, the real bottleneck is becoming power. That’s what led me to start looking at nuclear energy and uranium. Small modular reactors could become a major part of the solution, and if AI keeps demanding more electricity, the commodities that power that infrastructure could become just as important as the chips themselves.

Kevin O'Leary aka Mr. Wonderful

35,073 views • 25 days ago

Think bigger than data centers. Much bigger. Elon Musk just unveiled SpaceXAI's first AI satellite, a spacecraft designed specifically for running AI workloads in orbit. The satellite features 150 kW of solar power, a 70-meter wingspan and massive radiators to dump heat into space, solving one of the biggest challenges in scaling AI. But this is just the beginning. Musk says truly large-scale orbital AI could eventually require hundreds of gigawatts to a terawatt of power, meaning millions of tons of infrastructure in orbit. The long-term vision isn't AI running in data centers. It's AI running in space at a scale that's impossible on Earth. Elon Musk

Think bigger than data centers. Much bigger. Elon Musk just unveiled SpaceXAI's first AI satellite, a spacecraft designed specifically for running AI workloads in orbit. The satellite features 150 kW of solar power, a 70-meter wingspan and massive radiators to dump heat into space, solving one of the biggest challenges in scaling AI. But this is just the beginning. Musk says truly large-scale orbital AI could eventually require hundreds of gigawatts to a terawatt of power, meaning millions of tons of infrastructure in orbit. The long-term vision isn't AI running in data centers. It's AI running in space at a scale that's impossible on Earth. Elon Musk

Tesla Owners Silicon Valley

19,203 views • 3 months ago

$China spends a fraction of what the US does on data centers, and lacks the most advanced AI chips. Why, then, does its AI industry appear to be catching up, and rivaling Silicon Valley? I asked The New Yorker staff writer Evan Osnos who lived in China for eight years, and just returned from another reporting trip there with some fresh insights:$

China spends a fraction of what the US does on data centers, and lacks the most advanced AI chips. Why, then, does its AI industry appear to be catching up, and rivaling Silicon Valley? I asked The New Yorker staff writer Evan Osnos who lived in China for eight years, and just returned from another reporting trip there with some fresh insights:

Fareed Zakaria

107,361 views • 21 days ago

AWS has activated Project Rainier, one of the world’s largest AI compute clusters with nearly 500,000 Trainium2 chips brought online in under a year. Anthropic is already training and serving Claude on the system and is slated to scale past 1 million Trainium2 chips by the end of 2025. The UltraServer/UltraCluster design (NeuronLinks + EFA) speeds massive distributed training, while AWS’s vertically integrated stack and data-center efficiency targets aim to cut cost and environmental impact.

AWS has activated Project Rainier, one of the world’s largest AI compute clusters with nearly 500,000 Trainium2 chips brought online in under a year. Anthropic is already training and serving Claude on the system and is slated to scale past 1 million Trainium2 chips by the end of 2025. The UltraServer/UltraCluster design (NeuronLinks + EFA) speeds massive distributed training, while AWS’s vertically integrated stack and data-center efficiency targets aim to cut cost and environmental impact.

Wes Roth

50,699 views • 10 months ago

“Taiwanese companies make 90% of the data servers and the AI servers that run the AI realm.” “We make 95% of the advanced chips in Taiwan.” Taiwanese Ambassador Alexander Yui revealed to me just how reliant the entire U.S. AI system is on Taiwan. “Taiwan has become an important partner in terms of technology, in terms of investment, and in terms of education.” “United States is intending on preserving AI supremacy… and asking partners, including Taiwan, to reach that.” “To achieve that, Taiwan is a key partner.”

“Taiwanese companies make 90% of the data servers and the AI servers that run the AI realm.” “We make 95% of the advanced chips in Taiwan.” Taiwanese Ambassador Alexander Yui revealed to me just how reliant the entire U.S. AI system is on Taiwan. “Taiwan has become an important partner in terms of technology, in terms of investment, and in terms of education.” “United States is intending on preserving AI supremacy… and asking partners, including Taiwan, to reach that.” “To achieve that, Taiwan is a key partner.”

Jan Jekielek

69,008 views • 4 months ago

Mining Bitcoin and training AI models look different. Underneath, they run on the same input: clean power. HIVE has been building that foundation since 2017, supporting a Bitcoin network nobody can shut off, and AI compute that the next decade of discoveries depends on.

Mining Bitcoin and training AI models look different. Underneath, they run on the same input: clean power. HIVE has been building that foundation since 2017, supporting a Bitcoin network nobody can shut off, and AI compute that the next decade of discoveries depends on.

HIVE Digital Technologies

38,783 views • 11 days ago

Elon Musk explains how data centers in Space could help the US compete with China in AI “Given that Chinese AI companies are doing as well as they are with a relatively small amount of compute, it seems that if they had a lot of compute, there’s a good chance they would be the leaders. The constraints on AI are essentially electricity and AI chips. China is very strong in digital AI, and it has far more electricity than the United States. In fact, China already has more electricity than the United States, Europe and India combined. The US government can prevent US companies from using Chinese models, but it can’t stop the rest of the world from using them. That won’t stop China from becoming the leader in AI. Digital intelligence comes down to how much AI compute, how many chips and how much electricity you have. The current constraint is electricity outside China. In China, the constraint is chips. China is also closer than most people realize to solving the lithography problem, which would enable it to make AI chips in very large volumes. Once we address the power constraint with AI data centers in space, the constraint will once again be chips outside China.”

Elon Musk explains how data centers in Space could help the US compete with China in AI “Given that Chinese AI companies are doing as well as they are with a relatively small amount of compute, it seems that if they had a lot of compute, there’s a good chance they would be the leaders. The constraints on AI are essentially electricity and AI chips. China is very strong in digital AI, and it has far more electricity than the United States. In fact, China already has more electricity than the United States, Europe and India combined. The US government can prevent US companies from using Chinese models, but it can’t stop the rest of the world from using them. That won’t stop China from becoming the leader in AI. Digital intelligence comes down to how much AI compute, how many chips and how much electricity you have. The current constraint is electricity outside China. In China, the constraint is chips. China is also closer than most people realize to solving the lithography problem, which would enable it to make AI chips in very large volumes. Once we address the power constraint with AI data centers in space, the constraint will once again be chips outside China.”

DogeDesigner

245,936 views • 1 month ago