Bloomberg Tech: Meta to Become the Biggest Nuclear Buyer Among Hyperscalers
Episode Date: January 9, 2026
Host: Bloomberg Tech Team
Overview
This episode centers on Meta’s (Matter) landmark investment in nuclear energy to supply its massive AI-driven data centers, positioning the company as the largest buyer of nuclear power among hyperscalers. The show also explores Minimax’s blockbuster public debut in Hong Kong, the implications of Snowflake’s acquisition of Observe for AI observability, controversy around X AI’s Grok image tool, and major M&A news in streaming and chip industries. The episode features in-depth reporting, expert commentary, and exclusive interviews, unpacking the business and technology ramifications of these major headlines.
Key Discussion Points & Insights
1. Meta’s Nuclear Power Megadeal
[01:25 – 05:02]
- Announcement: Meta (Matter) secures three varied nuclear energy agreements, totaling up to 6.6 gigawatts, becoming the largest nuclear buyer among hyperscalers.
- Rationale: The deals focus both on supporting existing U.S. nuclear plants (preventing potential shutdowns due to underinvestment) and investing in future nuclear capacity.
- Strategic Context: Unlike typical hyperscalers who provide rentable cloud capacity, Matter directly powers its own data centers—primarily for AI training and inference—by aggressively locking in energy supply.
- Regional Breakdown: Some deals center around Ohio and Pennsylvania, targeting the Prometheus data center cluster. Matter is also engaging natural gas plants for more immediate needs.
- Industry Future-Proofing: Tech giants are expressing anxiety over insufficient energy supply and risk to existing nuclear infrastructure, prompting proactive investments.
“There’s an insatiable demand for energy and this cements their low carbon play.”
— Riley Griffin, Bloomberg Analyst/Reporter [02:47]
“This is a future proofing, as you say.”
— Riley Griffin, on Meta’s investments [04:45]
2. Energy as AI’s Bottleneck
[05:13 – 11:09]
- Expert Views: Nvidia CEO Jensen Huang (previously interviewed) and Paul Meeks (Freedom Capital Markets) emphasize energy supply—not demand—as the most critical bottleneck for AI data center growth.
- Long Lead Times: Meeks points out that new nuclear capacity likely won’t be operational before 2030–2032, making natural gas a short-term bridge.
- Industry Anxiety: Public optimism from AI sector leaders belies private concern over energy supply constraints impacting financial forecasts and operational planning.
“Of all the bottlenecks, [energy] is the most most important.”
— Paul Meeks [06:01]
“When I’m talking to companies about financial models... do you have the power? Yes, they are anxious and I think they’re actually very anxious.”
— Paul Meeks [07:41]
3. The Changing Shape of “Tech” Coverage
[08:12 – 12:01]
- Utilities Go High-Tech: Tech sector analysts now monitor utilities and infrastructure issues alongside traditional technology metrics due to the convergence of energy and data.
- Physical Bottlenecks: Construction logistics (pouring concrete for data centers) are now front-and-center for tech growth projections.
- Regulatory Lag: There’s currently no adequate federal framework dictating how the U.S. tech sector’s surging energy needs should be met, risking a repeat of past regulatory delays seen with the wider Internet.
“Now [utilities] have come to the forest. And... they’re your new era utilities.”
— Paul Meeks [08:35]
“We are still suffering from a lack of solid regulation for the Internet... Here is the next wave. We need to embrace it and at least come up with some guardrails.”
— Paul Meeks [09:48]
4. China’s AI IPO Boom: Minimax Goes Public
[14:50 – 21:59]
- IPO Success: Minimax, backed by Alibaba and Abu Dhabi’s SWF, doubles post-IPO in Hong Kong ($619M raised), signaling surging investor appetite for China’s AI sector.
- Minimax’s Model: CEO Ye Yen highlights cost innovation, “capital efficiency,” and a focus on global product adoption—not just competition with U.S. or domestic rivals.
- Competitive Landscape: Chinese “AI Tigers/Dragons” like Minimax, Deep Seek, and Jeep compete through fast go-to-market and cost-effective models, gaining global traction especially in developing markets.
- Tech Sourcing: Minimax expresses agnosticism about chip sourcing (domestic/Nvidia), focusing on ROI and accessibility for users worldwide.
“Actually I don’t think it’s a competition. It’s more about all the top talents... bring the technology breakthrough to the society, to the end users.”
— Ye Yen, Minimax CEO [15:31]
“We focus more on make the best technology and the best user experience that people will love to pay for the best performers models for their use cases.”
— Ye Yen [17:26]
5. Scandal at Grok/X AI: Deepfake Image Crisis
[21:59 – 26:06]
- Investigation: Bloomberg partners with researchers to show that Grok (X AI’s image generator) was producing thousands of explicit deepfake images per hour (over 6,700/hr at its peak).
- Platform Design: The merging of X (social media) with X AI allows users to easily create and distribute explicit or sexualized images of women and children, sparking public outrage and tighter access controls for Grok.
- Company Response: Elon Musk posts a non-specific warning; X/X AI otherwise remain unresponsive to requests for comment.
“Grok is one of the biggest deepfake producers on the Internet today.”
— Cecilia Danastasio, Bloomberg Reporter [24:18]
6. Snowflake Buys Observe: The AI Observability Race
[26:15 – 34:49]
- Acquisition News: Snowflake confirms the purchase of Observe, a leading AI-powered observability platform, in a deal reportedly around $1B.
- Observability Explained: CEO Sridhar Ramaswamy says observability helps organizations ensure applications, websites, and AI agents run smoothly by centralizing and analyzing massive telemetry data.
- Strategic Synergy: Observe was already built atop Snowflake, easing integration. Cost efficiency, rapid troubleshooting, and seamless platform upgrades are highlighted as customer wins.
- Broader Vision: Snowflake sees observability, data clean rooms, and adjacent features as core to its strategy for becoming an end-to-end data platform, citing rapid customer adoption.
“The leverage that you get is the game changer... I can get stuff done in a matter of a couple of hours... that would’ve taken me two to three weeks just last year.”
— Sridhar Ramaswamy, Snowflake CEO [32:03]
7. Media and Streaming M&A: Warner Bros/Netflix/Paramount
[38:44 – 45:44]
- Deal Dynamics: Netflix’s offer for Warner Bros Studios and streaming ($27/share) faces competition from Paramount/Skydance’s all-in $30/share bid, which controversially assigns “zero value” to legacy cable networks.
- Investor Sentiment: Netflix shares have slumped amid cost/integration/regulatory concerns and possible overpayment fears.
- Analyst Perspectives: Paramount has more to gain operationally from the full acquisition, while Netflix’s incremental benefit may be less substantial.
- Deal Timelines: January 21 is the next key date (Paramount tender offer expiration).
“There’s been some fear that Netflix would even pay more. There’s also some fear they’ve overpaid...”
— Matthew Dolgan, Morningstar [45:08]
8. Intel and the U.S. Government: A New Stakeholder Era
[46:51 – 49:24]
- Visit Recap: Intel CEO Lipp Bhutan updates President Trump and Commerce Secretary Lutnick at the White House on semiconductor progress; government now owns ~$11B in Intel stock from earlier investments.
- Performance Goals: While the holding has doubled in value, tens-of-billions gains require further outperformance.
- Product Pipeline: Intel is launching next-gen chips as promised; the industry awaits real-world performance data following CES.
“We own currently about $11 billion worth of intel stock and that’s a nice return, roughly twice what we owned when we made the investment as a nation last year, but not in the tens of billions.”
— Bloomberg’s in King [47:49]
Notable Quotes & Moments (with Timestamps)
-
“There’s an insatiable demand for energy and this cements their low carbon play.”
— Riley Griffin [02:47] -
“Of all the bottlenecks, [energy] is the most most important.”
— Paul Meeks [06:01] -
“We are still suffering from a lack of solid regulation for the Internet... Here is the next wave. We need to embrace it and at least come up with some guardrails.”
— Paul Meeks [09:48] -
“We focus more on make the best technology and the best user experience that people will love to pay for the best performers models for their use cases.”
— Minimax CEO Ye Yen [17:26] -
“Grok is one of the biggest deepfake producers on the Internet today.”
— Cecilia Danastasio [24:18] -
“The leverage that you get is the game changer... I can get stuff done in a matter of a couple of hours... that would’ve taken me two to three weeks just last year.”
— Sridhar Ramaswamy, Snowflake CEO [32:03] -
“There’s been some fear that Netflix would even pay more. There’s also some fear they’ve overpaid...”
— Matthew Dolgan, Morningstar [45:08]
Important Timestamps by Topic
- Meta’s Nuclear Energy Strategy: [01:25 – 05:02]
- Energy as a Bottleneck / Hyperscaler Insight: [05:13 – 11:09]
- Analyst/Regulatory Perspective: [09:36 – 12:01]
- Minimax IPO and China’s AI Landscape: [14:50 – 21:59]
- Grok/X AI Deepfake Controversy: [21:59 – 26:06]
- Snowflake’s Observability Acquisition: [26:15 – 34:49]
- Streaming M&A (Netflix, Warner Bros, Paramount): [38:44 – 45:44]
- Intel/White House Stakeholder Update: [46:51 – 49:24]
Tone and Style
The host and guests bring an analytical, incisive tone. Discussions mix financial reality checks, technology insights, and frank assessment of sector bottlenecks and opportunities. There’s a sense of urgency around energy, convergence of tech and infrastructure, and competitive dynamics on both sides of the Pacific.
This summary covers all the major league developments, candid expert analysis, and memorable exchanges from this episode—ideal for anyone in tech, business, energy, or policy looking to understand the forces shaping AI, cloud, infrastructure, and next-gen digital business in 2026 and beyond.