Bloomberg Tech — “Meta to Spend Billions on AMD Gear, AI Scare Trade Continues”
Date: February 24, 2026
Hosts: Caroline Hyde (New York), Ed Ludlow (San Francisco)
Guests/Contributors: Riley Griffin, Ian King, Hillary Frisch, Mandeep Singh, Brody Ford, Michael McDermott, Ryan Pope, Matt Hart
Episode Overview
This episode centers on Meta’s massive, multi-billion dollar deal to buy AMD chips and data center hardware, marking a new surge in the AI infrastructure arms race. It also dissects the resulting market turbulence, the challenges and opportunities facing software vendors amid rapid AI disruption, and features interviews with key industry insiders and startup founders trying to compete with the tech giants. Additional segments touch on pending major media mergers, the regulatory and business implications of prediction markets, and a look ahead to President Trump’s State of the Union, particularly regarding energy costs for AI.
Key Discussion Points & Insights
1. Meta’s Multibillion-Dollar AMD Deal: Scope and Strategy
[02:59–09:49]
- Deal Details:
- Meta has agreed to purchase AMD chips and data center gear in a deal worth tens of billions—6 GW of compute capacity over time.
- The contract includes share warrants for Meta, similar to AMD’s structure with OpenAI, involving operational and financial milestones.
“It's a real endorsement…They're playing catch-up with Nvidia…This is a real strong affirmation and it also includes share warrants…”
— Riley Griffin, [03:36] - Motivation & Market Context:
- Meta is aggressively “front-loading” compute capacity to fuel its AI ambitions, targeting hundreds of gigawatts as it pursues artificial superintelligence.
- Spending is projected at $135 billion in CapEx this year, much of which focuses on chips for both AI objectives and Meta’s core social products.
“Mark Zuckerberg last month announced MetaCompute with ambitions to get hundreds of gigawatts to fuel its data centers and ultimately reach that super intelligence goal ... spending won’t stop.”
— Caroline Hyde, [03:59] - Diversification & Competitive Dynamics:
- Meta diversifies its hardware bets: custom chips, Nvidia and AMD hardware, each for different AI workloads and applications.
- The move doesn't signal immediate anxiety for Nvidia's dominance; demand is so high that “all boats are floating.”
- AMD’s inclusion signals tighter relationships through equity-sharing, validating AMD’s growing presence in the market, though it still trails Nvidia in revenue and installed base.
“Ultimately, when the market slows down ... then we'll see the market share shift because then it'll really matter.”
— Riley Griffin, [07:46]- Regulatory approvals and energy delivery bottlenecks are gating factors; specific chip deployments remain to be determined.
2. AI Disrupts Legacy Software Markets
[10:17–16:29]
- Market Impact:
- IBM shares plunge after Anthropic’s “Claude” AI model showcases ability to modernize the COBOL programming language, challenging IBM’s core infrastructure.
“You've got to feel powerful if you're Anthropic right now, right? ... Markets are really jumpy. The potential that Claude or another AI tool can disrupt the leaders in a given software category is really frightening investors right now.”
— Brody Ford, [11:02] - AI’s Dual Role:
- AI can both disrupt and enhance established platforms. Cloud and SaaS companies must urgently show that AI is additive, not just offsetting declines.
“The idea that software vendors lose the kind of pricing leverage they've enjoyed for so long because of that potential disruption—it's really something that markets are struggling to grapple with right now…”
— Brody Ford, [12:04] - How Companies Defend Against AI-Driven Panic:
- Companies must demonstrate real adoption and value-add from AI, not just claim association.
“What the vendors really have to do is show that AI is additive to the business, to beat and raise and show that AI isn't just offsetting declines in the core business, but it's actually adding ...”
— Hillary Frisch, [14:55]
3. Research, Market Reactions, and AI’s Real Disruption Potential
[16:29–18:46]
- Market Jitters:
- Reports (e.g., from Citrine, Harvard) amplify risk concerns but also point to new outperformance from AI-infused financial analysis.
- AI’s impact is uneven—while some subsectors are defensible (security, data infrastructure), others (application software) face existential threats.
“AI represents a risk, and it also represents an opportunity for many of the incumbent vendors ... there are reasons why a lot of subsectors should be relatively more defensible...”
— Hillary Frisch, [17:34] - Security & Data Platforms as Defensible Moats:
- Security vendors integrate AI for vulnerability scanning but retain an edge due to physical enforcement—and AI itself increases attack surfaces.
- Next-gen data platforms (Snowflake, MongoDB, Databricks) support new AI applications, automate complexity amid the explosion of data.
4. AI Hardware: The Race Beyond Nvidia
[27:37–30:09, 33:26–39:13]
- Meta’s Leverage and AMD’s Play:
- Meta’s deal structure with AMD gives it leverage over Nvidia, potentially reducing chip costs as a percentage of CapEx and balancing future supply.
- Co-engineering and focus on inference (MI 450) signal AMD’s intent to differentiate via custom solutions for Meta’s AI workloads.
“This could potentially double, you know, every year in terms of just by adding one customer at the scale that they plan to buy the chips from AMD.”
— Mandeep Singh, [28:02] - Startup Competition:
- Matt X raises $500M to build hybrid (HBM+SRAM) AI chips, targeting ultra-high throughput (flops per mm²) and low latency for LLM workloads.
- Their approach breaks strict compatibility requirements to maximize performance for AI tasks—a “blank slate” design.
“If you want to absolutely nail the LM workload, you have to be willing to break compatibility with previous chips.”
— Ryan Pope, [37:15]
5. AI’s Verticalization: Domain-Specific Solutions
[39:39–44:37]
- Basis (Accounting AI) Story:
- Basis distinguishes itself from general-purpose tools (like Anthropic’s cowork agent) by specializing in accounting: full business tax return automation, audit-ready accuracy, user experience tuned for experts.
“You want domain specific capabilities ... we announced today the first example of a long running agent completing an entire business tax return workbook. That's something that you can't really do in any other AI tool.”
— Matt Hart, [39:39]- AI is positioned as a co-pilot rather than a job destroyer: it empowers accountants to accomplish more, tackling accounting tasks that were previously infeasible due to resource limits.
6. Macro Context: Power, Regulation, and Prediction Markets
[45:22–51:13]
- US Policy Response:
- Anticipation around President Trump’s State of the Union speech: expected focus on tech giants’ energy consumption, with new commitments from big tech to share higher electricity costs of data centers.
“President Trump is set to announce a negotiated commitment from big tech companies to pay more when it comes to electricity costs related to AI data centers in a bid to remove some of that burden from U.S. consumers.”
— Caroline Hyde, [45:23] - Prediction Markets:
- Segment explores the proliferation and regulation debate around online prediction markets (Kalshi, Polymarket), their uses in politics and events, and their surprising predictive accuracy—especially with high liquidity.
“Prediction markets are more accurate than polling and certainly more accurate than you or me or any individual trying to prognosticate about elections.”
— Brody Ford, [47:55]
Notable Quotes & Memorable Moments
- On AI Hype & Software Panic:
“Investors shoot first and ask questions at some very distant date. So we're seeing a pretty dramatic compression in terminal multiples of many of the names.”
— Hillary Frisch, [13:58] - On the Scale of Meta’s Bet:
“They're going to try to get as much as they can, gobble it up while there's still availability ... could be applied not just for AI purposes, but for that core social media business which still drives more than 98% of their revenue.”
— Caroline Hyde, [08:34] - On Hardware Demand:
“Demand for LLM Compute is just insatiable. All of the Frontier labs are looking at where this space is going and they're all concerned: I'm going to run out of silicon.”
— Ryan Pope, [33:54] - On AI as a Copilot:
“From our perspective, this is going to allow firms and the accountants at those firms to take on even more work and get more things done, get more things done in the same way that software engineering has been able to take off over the course of the last 12 months.”
— Matt Hart, [44:37]
Timestamps for Key Segments
| Timestamp | Topic / Segment | |---------------|--------------------------------------------------------------| | 02:06–09:49 | Meta’s AMD deal: context, mechanics, market implications | | 10:17–16:29 | AI-driven disruption in legacy software & SaaS, Anthropic’s impact | | 16:29–18:46 | Research on AI risk, ‘AI panic’ vs. real change | | 23:03–23:56 | News headlines: Anthropic, OpenAI, market moves | | 24:28–26:18 | Paramount’s increased bid for Warner Brothers | | 27:37–30:09 | Analyst insight: AMD vs Nvidia, Meta’s CapEx, deal dynamics | | 33:26–39:13 | Startup insight: Matt X’s $500M round and chip strategy | | 39:39–44:37 | Domain AI: Basis transforms business tax filing | | 45:22–47:49 | State of the Union preview: Tech energy policy | | 47:11–51:13 | Prediction markets: regulation, accuracy, societal impact |
Tone & Narrative
The episode is fast-paced and energetic, toggling between bullish, anxious, and occasionally irreverent as the hosts parse market mayhem and hype cycles. The language is sharp, often skeptical, and frequently lightened by in-jokes (“Matter is traded flat as a pancake”), with a focus on giving investors and tech watchers a strategic lens on both headline-grabbing deals and underlying systemic shifts.
Summary
This Bloomberg Tech episode skillfully unpacks the dual boom of AI infrastructure—exemplified by Meta’s huge AMD hardware deal—and the resulting volatility in software markets as AI agents threaten established players, create new platforms, and drive fierce competition for chip dominance. By mixing high-level analysis, timely news scoops, and direct input from both established analysts and emerging founders, the show offers a comprehensive, nuanced, and vivid snapshot of the ever-evolving technology landscape in early 2026.