The Exchange – Episode Summary
Episode: OpenAI Optimism, Microsoft's AI Malaise, and Bitcoin $38K?
Date: February 9, 2026
Host: Kelly Evans (CNBC)
Overview
This episode of The Exchange dives into the sharply divergent currents shaping today’s tech and market landscape: OpenAI-fueled optimism (and how it’s buoying Oracle), Microsoft’s stumbles in the evolving AI race, and a bearish case for Bitcoin’s future price. The conversation features in-depth analysis from leading technology and market experts, practical investment debates, and a look at changing trends in software, crypto, metals, and the wider stock market.
Key Discussion Points & Insights
1. Microsoft & The AI Narrative – “Malaise” and Strategic Dilemmas
Guest: Ben Wright, Head of Technology Research, Melius Research
-
Microsoft Downgraded: Melius Research downgrades MSFT to “Hold,” citing narrative loss and strategic struggle with AI investments.
- Copilot’s Costly Burden: Microsoft’s push to build out Copilot may pressure both Azure’s profitability and overall free cash flow.
- Free vs. Paid AI: The analyst predicts Microsoft may need to give Copilot away for free to drive adoption as competition from upstarts like Anthropic and OpenAI accelerates.
- Innovator’s Dilemma: Microsoft is “damned if they do, damned if they don’t”—they must protect massive SaaS business lines while facing the risk of cannibalizing them via free AI tools.
- Channel Inertia Problems: Unlike newer AI companies, Microsoft’s sales depend on traditional ISVs and reseller channels, slowing their ability to innovate and pivot.
Notable Quotes:
"We don’t think paying extra for AI is a thing... they may have to give away Copilot." — Ben Wright (04:12)
“They're just going to have trouble getting paid for it … that takes away from Azure upside.” — Ben Wright (07:23)Timestamps:
- Microsoft’s cash flow and Copilot dilemma (02:47–08:22)
- Strategic question: abandon Copilot or persist? (06:47–09:21)
- OpenAI partnership potential (09:21–10:18)
2. OpenAI’s Resurgence & Oracle’s Boom: Narrative Reversal
Guest: Gil Luria, Head of Technology Research, DA Davidson
-
Oracle Upgraded on OpenAI Optimism: DA Davidson upgrades Oracle as faith in OpenAI’s turnaround and capital-raising rebounds.
- OpenAI’s Focus: OpenAI recommits to ChatGPT, advertising, and re-energized growth after strategic missteps.
- Massive “War Chest”: Expected to raise $100B+ with backing from Microsoft, Amazon, Nvidia, fueling continued AI advances.
- Azure & Oracle Benefits: OpenAI’s workloads benefit Microsoft Azure and, by extension, Oracle, as their compute backbone.
- Microsoft Bull Case: Despite narratives fixating on Copilot, the real story is Azure’s hyper-growth and dominant AI infrastructure.
Notable Quotes:
“Microsoft is winning more than anybody else in AI … Copilot is an insignificant rounding error.” — Gil Luria (11:00)
“OpenAI is actually winning ... with another $40 billion of cash on hand, OpenAI now has the war chest to continue to compete.” — Gil Luria (11:40)Timestamps:
- Oracle’s upgrade and OpenAI’s benefits (10:27–13:19)
- Comparison of Microsoft, Google, Amazon AI positioning (11:00–13:19)
- Azure’s capital spending and business strategy (13:20–14:36)
- Discussion on tech stock valuations, future reversals (15:44–16:40)
3. Bitcoin Bearishness: Why $38K Is on the Table
Guest: Barry Bannister, Chief Equity Strategist, Stifel
-
Bitcoin as Speculative Tech Asset: Bitcoin is trading more like high-beta tech—no longer acting as “digital gold.”
-
Liquidity & Macro Headwinds: The Fed’s less accommodating stance and waning liquidity threaten Bitcoin and the broader speculative trade.
-
Price Target Logic: Previous major drawdowns suggest a floor around $38,000–$40,000.
-
Wider Market Implications: Potential “K-shaped” economy—rich consumers tied to stock market wealth, but broad growth limited; caution urged on betting for a seamless industrials rebound.
Notable Quotes:
“Bitcoin is not digital gold... it’s behaving more like an overextended tech stock.” — Barry Bannister (19:32)
“As interest rates are not as cooperative, you would see some pressure on the most speculative elements of the market. And Bitcoin’s in that bucket.” — Barry Bannister (22:21)Timestamps:
- Bitcoin macro behavior (19:32–22:17)
- Rationale for $38K target (22:17–23:20)
- Broader economic/market rotation skepticism (23:30–24:54)
4. Metals & The Retail Mania
Reporter: Pippa Stevens
-
Precious Metals Surge: Retail “buy gold/silver” searches at record highs; gold and silver prices rocketing on weaker dollar.
-
Broader Metals Rally: Lithium and uranium prices jump, driven by low inventory, battery sector demand, and speculation.
Timestamps:
- Gold, silver, and battery/lithium moves (28:17–29:41)
5. Software Sector: Still Life Beyond AI Anxiety
Guest: Anastasia Amoroso, Chief Investment Strategist, Partners Group
-
Software Shakeout Not Universal: Many software companies (especially with vertical expertise and proprietary data) remain strong, despite AI disruption fears.
-
Public vs. Private: Private market software offers opportunities—domain experts embedding “agentic AI” are attractive bets.
-
Differentiation Is Key: Firms with shallow moats or only basic tools face existential threats; others, with embedded AI and data leverage, can thrive even in AI era.
-
Macro View: Strong consumer support (tax cuts, stimulus), high corporate productivity, and robust margins support stock markets generally.
Notable Quotes:
“If there is the death of software, it is not now, it’s not tomorrow, and it’s not of all software.” — Anastasia Amoroso (32:27)
“We really kind of dissect the incumbents into... little to no competitive moat... likely going to be disintermediated by agentic AI.” — Anastasia Amoroso (35:29)Timestamps:
- Software stock anxieties and fundamentals (31:47–32:32)
- The case for software as a buying opportunity (32:55–33:46)
- Private market perspectives and software specialization (33:46–36:48)
- Macro case: consumer, margins, productivity (37:20–38:40)
6. Databricks CEO on AI Disruption & Secret to Software’s Future
Guest: Ali Ghodsi, CEO, Databricks (interviewed by Deirdre Bosa & Kelly Evans)
-
Rise of AI Agents: 80% of databases on Databricks now built by AI agents—signals accelerating automation in software infrastructure.
-
Legacy Moats Shrinking: With AI interfaces, technical barriers and stickiness of old databases is eroding; true differentiation depends on data and agility.
-
Winners & Losers: Entrenched SaaS firms with proprietary, “system of record” data are in best shape (e.g., SAP); tools with narrow, undifferentiated offerings face the greatest risk.
-
IPO Strategy: Databricks not in a hurry to go public; heavy fundraising now to stay “well capitalized” and pursue long-term AI investments—cautious after the lesson of 2022’s market crunch.
Notable Quotes:
“This moat of … learning a new interface … is evaporating because humans can now just talk to AI.” — Ali Ghodsi (41:00)
“Not all SaaS is created equal… some may be disrupted sooner than others.” — Ali Ghodsi (42:16)Timestamps:
- AI agents transform database usage (39:52–41:00)
- How generative AI disrupts software moats (41:00–42:16)
- What kinds of SaaS are most vulnerable (42:39–43:48)
- Capital raising, IPO outlook, and market volatility (44:13–44:58)
7. Meta & Social Media Trials: Legal & Platform Risks
Reporter: Julia Boorstin
-
Two major trials: One in New Mexico (on app design contributing to child sexual exploitation), and one in Los Angeles (addictive algorithms—autoplay/infinite scroll).
-
Section 230 Loophole: These cases challenge Meta/YouTube not for content, but for platform designs themselves—a new front in social media regulation.
-
Key Testimonies Expected: Instagram’s Adam Mosseri and Mark Zuckerberg to testify in coming weeks.
Timestamps:
- Meta legal updates, platform safety debate (44:58–[end])
Notable Quotes & Moments (with Timestamps)
- "We don’t think paying extra for AI is a thing... they may have to give away Copilot." – Ben Wright, Melius Research (04:12)
- "Microsoft is winning more than anybody else in AI … Copilot is an insignificant rounding error." – Gil Luria, DA Davidson (11:00)
- "Bitcoin is not digital gold... it’s behaving more like an overextended tech stock." – Barry Bannister, Stifel (19:32)
- "If there is the death of software, it is not now, it’s not tomorrow, and it’s not of all software." – Anastasia Amoroso, Partners Group (32:27)
- “This moat of … learning a new interface … is evaporating because humans can now just talk to AI.” – Ali Ghodsi, Databricks (41:00)
- “Not all SaaS is created equal… some may be disrupted sooner than others.” – Ali Ghodsi, Databricks (42:16)
Segment Breakdown with Timestamps
| Segment/Theme | Timestamp | |--------------------------------------|-------------------| | Microsoft AI Dilemma (Ben Wright) | 02:47–10:18 | | Oracle, OpenAI, & Cloud Arms Race (Gil Luria) | 10:27–16:50 | | Bitcoin Bear View (Barry Bannister) | 19:03–24:54 | | Metals Mania (Pippa Stevens) | 28:17–29:41 | | Private & Public Software Outlook (Anastasia Amoroso) | 31:36–38:40 | | Databricks & The Future of Software (Ali Ghodsi) | 39:52–44:58 | | Meta’s Legal Trials (Julia Boorstin) | 44:58–[end] |
Tone & Language
- Direct, analytical, pragmatic: Experts debate without sugarcoating corporate realities or market uncertainties.
- Occasional humor and candid self-reflection: Eg. “Markets are funny. First they ignore this problem for two, three years, now they wake up and think it's going to happen in a week.” (Ali Ghodsi, 41:00)
- Balanced optimism and caution: There’s a clear message—tech disruption is not “all or nothing”; winners and losers will emerge based on moats, data, and adaptability.
Conclusion
This packed episode of The Exchange captures tech’s “AI moment” in real time: Microsoft finds itself in a tricky transition, Oracle and OpenAI ride new momentum, and Databricks explains what the coming age of AI agents means for who wins and loses the software wars. Meanwhile, caution reigns in crypto and old assumptions about where “safety” lies in stocks or metals are up for debate. A must-listen for investors and tech-watchers seeking nuance, context, and practical insight.
