# The 10 Podcasts TMT Analysts Should Be Listening To Right Now

> We scanned 538 podcast episodes and pulled the 10 highest-signal conversations for TMT analysts covering AI infrastructure, semis, hyperscalers, and enterprise software. Here are the episodes worth your time and the 20 stocks driving the debate.


> **TL;DR** — The best TMT signal is showing up in long-form audio before it hits sell-side notes. We scanned 538 podcast episodes and ranked the 10 highest-signal conversations for analysts covering AI infrastructure, semiconductors, and enterprise software — plus bull and bear cases on the 20 stocks being debated most.

## TMT is built different

If you cover general equities, your research workflow has an established structure to it. Trailing cash flows fit into a DCF and standard multiples like EV/EBITDA give you a starting point. GAAP profitability tells you most of what you need to know about bottom-line trends.

If you are a TMT analyst, you don't have that luxury.

You are in a sector where traditional accounting punishes the best business models. A generalist might reject a company for negative EPS, but your job is to figure out whether massive capex and heavy R&D spend are building a 130% net revenue retention machine or simply burning capital. You are constantly translating unit economics, LTV:CAC ratios, and intangible assets into future operating leverage.

Your biggest challenge is not building the model but getting the right inputs before everybody else does.

TMT analysts value better signal, and there is a new source they should be scanning.

## Why podcasts

[Podcasts](https://www.matterfact.com/blog/why-podcasts-why-now) are quietly becoming one of the highest-signal research inputs in TMT. The best episodes can surface things that matter before they appear in sell-side research.

You hear CEOs talk through spending plans before they file the 10-Q and investors explain what they are buying and avoiding. You will also hear operators describe bottlenecks that are not in a model yet — power, memory, data centers, custom chips, software pricing, agent adoption, cloud share, free cash flow.

Our AI agent (MatterFact Podcast) ran a scan across 538 recent podcast episodes and narrowed it down to the 10 highest-signal conversations for analysts focused on AI infrastructure, semiconductors, hyperscalers, and enterprise software. The goal is simple: help you generate ideas, improve your existing thesis on a current portfolio position, and get a better read on where the market may be too excited or too dismissive.

## 1. [Catalyst with Shayle Kann: Inside Google's Massive AI CapEx](https://app.matterfact.com/podcasts/84f89a1a0dba29a991cdec738837e0baaee9fda472c856a865431ec0efda39b7)

Start here. Shayle Kann's conversation with Amin Vahdat, Google's chief technologist for AI infrastructure, is one of the cleanest windows into how a hyperscaler actually thinks about the AI buildout. This is not generic AI talk. It goes into chips, power, labor, reliability, and the real tradeoffs behind data center scale.

If you cover Alphabet, Nvidia, Broadcom, Marvell, Vertiv, utilities, or power-exposed industrials, this is required listening. The takeaway: the AI race is an infrastructure race. The winner may not be the company with the best model. It may be the company that can source power, optimize silicon, manage reliability, and deploy capacity faster than anyone else.

## 2. [RiskReversal: Peter Boockvar on the AI Semi Trade](https://app.matterfact.com/podcasts/f09d615b7b1063fb83693844396bf6ae18d00fec25230c897d4ef3d387cbb65e)

Every long book needs a good bear case. Peter Boockvar makes one here. His argument: the semiconductor rally has become stretched, driven by front-loaded orders rather than durable demand. He walks through the AI beneficiary trade, data center hardware, hyperscaler free cash flow pressure, and the risk of a reversal if capital spending guidance changes.

Most TMT teams already know the bull case for Nvidia and semis. The harder work is understanding where it breaks. The key question: what happens if hyperscalers keep spending, but investors stop rewarding the spend?

## 3. [Cheeky Pint: Sundar Pichai on the Future of AI at Google](https://app.matterfact.com/podcasts/1af2b88eab93aa4cf36cba9709cf2eb25695e53744b14edc6773ad73bf7e60fa)

This is the CEO-level version of the Google thesis. Pichai talks through Google's AI strategy, the capex budget, memory and power constraints, and the cultural shift inside the company. 2026, he argues, is a supply-constrained year.

For Alphabet analysts, the market still tends to reduce Google to one question: will AI hurt Search? That framing is too narrow. The better question is whether Google's full-stack advantage matters more as AI shifts from experiments to high-volume production. Google has chips, models, products, distribution, and cloud. That does not mean the stock is easy. It means the debate deserves more than a chatbot share chart.

## 4. [No Priors: How Capital Is Powering the AI Infrastructure Buildout](https://app.matterfact.com/podcasts/7454d8a9cb087330bf500b0156f74df4783b61f3ef1aebeede7dc161fb4ecd3e)

Sarah Guo's conversation with Neil Tiwari of Magnetar is one of the better episodes on the money behind AI infrastructure. GPUs, data centers, long-term contracts, private credit, sovereign buyers, take-or-pay structures — these are now part of the core TMT research process.

The AI buildout is no longer just a technology story. It is a financing story. Analysts who do not understand how AI infrastructure gets financed are missing a large part of the picture.

## 5. [Morgan Stanley: AI's $3 Trillion Question](https://app.matterfact.com/podcasts/2307af61795855a2dc8207029bb0f6a48b6081c986667d5c5f05077af762874d)

Morgan Stanley's *Thoughts on the Market* puts the whole AI infrastructure cycle into a capital markets frame. The point: AI spending is so large that balance sheets matter again. That has implications for hyperscalers, software, REITs, private credit, utilities, chip suppliers, and industrial infrastructure.

If AI capex becomes a multi-year financing cycle, the winners may include companies far outside the obvious Nvidia trade.

```request-access
variant: inline
heading: Want this kind of synthesis on your own coverage?
buttonText: Request access
```

## 6. [20VC: Anthropic, the Pentagon, Stock Picks, and the Data Center Arms Race](https://app.matterfact.com/podcasts/94a8ebd7a5cdb51ab1448a98f8768e4ef29b6ad8eea7c5c46e483bf1930aacd7)

This one is opinionated and better for it. Jason Lemkin, Rory O'Driscoll, and Harry Stebbings debate stock picks, public software, data centers, and where AI is actually changing the market.

For software analysts, this is one of the strongest episodes in the set. The core question: can software companies reaccelerate with AI, or does AI compress seats, margins, and product value? That debate matters for CrowdStrike, Salesforce, Cloudflare, ServiceNow, Atlassian, Monday.com, Wix, Klaviyo, and many others.

## 7. [The Circuit: CAPEXXXXXX and the Trillion Dollar Datacenter Race](https://app.matterfact.com/podcasts/418e6476f09b386822b9e5406fb919f2515d2d0ad6f331fcf89583f6b632ffd1)

Ben Bajarin and Jay Goldberg are very strong on semis and infrastructure. This episode connects cloud share, chip allocation, custom silicon, and capex plans in a way that most single-topic pods do not.

Amazon's AWS position matters. Google's TPU story matters. Nvidia's Vera Rubin allocation matters. Broadcom and Marvell matter because custom silicon is not a side story anymore — it is one of the main ways hyperscalers try to control cost.

## 8. [Motley Fool Money: AI Capex Is Off the Charts](https://app.matterfact.com/podcasts/ca76243458a499f78ca16470e87f3957f61db581cc121bc7b7bd9e7a48fe9323)

Less institutional than some of the others, but it states the simple risks clearly. The valuable part is the skepticism around companies using heavy debt to chase AI infrastructure demand. CoreWeave and Oracle are the obvious debates.

The question is not whether demand exists. The question is who carries the risk if demand is slower, financing costs stay high, or one major customer becomes too important.

## 9. [Hedgeye: AI Capex, Back to Infrastructure](https://app.matterfact.com/podcasts/c7e3268f0ad247237f3b2745b4ac2d1221ebb3af03adc2274bb02d0fca9f2b80)

Hedgeye's *Protect the Pile* episode puts AI capex into a macro and portfolio risk frame. It widens the lens in a way most TMT pods do not: power, grid equipment, commodities, construction, cooling, data center real estate, industrial capacity.

Analysts who only look at chips will miss second-order winners. Analysts who only look at software will miss the stress building under some software business models.

## 10. [Chip Stock Investor: Beyond AI Data Centers](https://app.matterfact.com/podcasts/5ac0f18b305d5090c0e09406dd4ee55ff2da6cc95a372baeb15d0943bd5230f6)

This episode is about the second-order semiconductor names. Every TMT pod already covers Nvidia. The harder work is finding the suppliers that benefit from higher power density, more complex racks, timing, analog, power management, networking, and industrial recovery.

Names like Monolithic Power, NXP, Texas Instruments, Microchip, and Littelfuse may not have headline power, but they may tell us more about how broad the AI cycle is actually becoming.

## The 20 stocks driving the debate

Across these ten episodes, the same names keep showing up — but the bull and bear cases sit on top of each other. Below is the large-cap version of the debate sheet, distilled. The SMID-cap counterpart is available as a download.

### Large-cap: where the debate is loudest

| Ticker | Bull case | Bear case |
| --- | --- | --- |
| NVDA | $1T+ Blackwell/Rubin visibility through 2027; "inflection of inference" | Boockvar: rally is front-loaded orders, not durable demand |
| GOOGL | Pichai: full-stack TPU + Gemini 3, supply-constrained 2026 | Search disruption from chat-native AI; capex weighs on FCF |
| MSFT | Azure AI backlog + neocloud option value on 5-yr leases | $50–60B/yr Nvidia bill is forcing the custom-silicon roadmap |
| AMZN | AWS + Trainium scale; Anthropic positions itself as the second model | Hyperscaler capex discipline pressure (Oracle precedent) |
| META | Custom AI chip + "millions" of Nvidia processors, Reels/Ads ROI | $30–40B/yr capex with no external cloud to monetize it |
| ORCL | Rallied 10% on a non-raise of capex — discipline gets rewarded now | Heavy AI-tenant concentration; debt funding the buildout |
| AVGO | Custom silicon for hyperscalers is the central trade, not a sidebar | Customer concentration; lumpy program timing |
| AMD | MI300/MI400 closing the price/perf gap; sovereign and neocloud demand | Rasgon: "way too early to worry" — Nvidia moat is intact |
| CRM | Agentforce is the cleanest agentic-AI revenue story in software | 20VC: AI compresses seats and pricing power across SaaS |
| NOW | Enterprise AI workflow consolidation; Now Assist attach rates | AI-native rivals undercutting platform moat |

```request-access
heading: Download the SMID-cap bull/bear matrix.
description: Ten second-order names — Vertiv, Marvell, CoreWeave, CrowdStrike, Cloudflare, Atlassian, Monday.com, Monolithic Power, Littelfuse, Klaviyo — with the bull and bear cases pulled from the corpus. Request access to download.
buttonText: Download the SMID matrix
```

## Read the corpus, not the episode

Ten episodes, twenty stocks, one debate that is moving faster than sell-side notes can keep up with. Each episode is a useful hour. Read together, they do what no single hour does: they let you weight the views — Vahdat against Boockvar, Pichai against the chatbot-share narrative, Goldberg against Jensen — against each other, and against the actions of the people writing the largest cheques.

That kind of synthesis is what TMT analysts keep asking Matterfact to do for them. Not "summarise this episode" — "tell me what 538 episodes of serious AI infrastructure coverage add up to, with every source one click away."

```request-access
heading: Run your own corpus.
description: Matterfact is deployed with select institutional partners. Request access to run it on your own coverage.
buttonText: Request access
```