Feb 12, 2026 · 1h 30m · cheeky-pint
Ben Thompson from Stratechery on AI ads, the end of SaaS, and the future of media
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Stratechery founder Ben Thompson discusses the strategic and economic dynamics of the AI revolution, analyzing digital advertising architecture, agentic commerce, independent media business models, and semiconductor infrastructure constraints.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. John holds 19.9% of the talking time here. How this is scored →
John as informed peer 5.4 Guest teaching 4.8 Guest disagreement 3.4 John pushing back 3.3
speaking balance: gold is John, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 11:23 Vehement rejection of OpenAI ad strategy
Ben emphatically rejects OpenAI's conversational ad execution as terrible and the bare minimum easiest solution.
Hardest push from John ▶ 17:03 Pushback on profiling-based AI adsJohn challenges Ben's preference for broad user profiling by highlighting real consumer anxiety and creepy targeting backlash.
Biggest teaching moment ▶ 1:07:01 Economics of semiconductor fab depreciationBen provides a masterclass on capital expenditure, depreciation math, and logic monopoly dynamics explaining why TSMC refuses to overbuild for AI surges.
John holds their own ▶ 50:59 Anthropic Workday counterexampleJohn counters the narrative that enterprise SaaS is being replaced by AI code generation by pointing out that even leading AI labs like Anthropic buy Workday.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | John as informed peer | Guest teaching | Guest disagreement | John pushing back | Why |
|---|---|---|---|---|---|---|
| Aggregation Theory in AI and the Defense of Advertising | 5 | 3 | 2 | 2 | John opens by framing Aggregation Theory and notes Booking.com's outsized shareholder return relative to Google. Ben discusses how OpenAI failed to build an ad engine around ChatGPT, contrasting Valley anti-ad bias with the economic efficiency of ads. | |
| AI Advertising Architecture: Profiling vs. Search Context | 6 | 5 | 6 | 6 | Ben aggressively dismisses OpenAI's conversational ads as the bare minimum easiest solution, advocating for Meta-style user profiling. John pushes back by defending Google's relevance-ranked ads and highlighting consumer privacy backlash when targeting feels too invasive. | |
| The Power of the Feed and Algorithmic Media | 5 | 4 | 2 | 2 | John puts forward his thesis that tech giants expand primarily because their core innovation scales by orders of magnitude beyond founder expectations. Ben agrees and illustrates how Meta's algorithmic feed evolved to beat the social graph constraint. | |
| ByteDance, Geopolitics, and Algorithmic Governance | 4 | 6 | 4 | 3 | Ben explains why he supported the TikTok divestiture based on algorithmic curation rather than user data hysteria, recounting his Houston Rockets VPN experiment. John asks about the spin-out details and is surprised to learn ByteDance retained algorithm control. | |
| Agentic Commerce and the Danger of Over-Optimization | 6 | 4 | 4 | 5 | Ben cautions that agentic optimization risks eliminating unmeasurable human qualities, drawing parallels to Moreyball in basketball. John pushes back using Amazon Basics, arguing that functional commodity goods do not need a soul to create consumer surplus. | |
| The Four Levels of Agentic Commerce | 7 | 5 | 4 | 4 | John articulates Stripe's 3-level taxonomy of agentic commerce from form-filling up to preference embeddings. Ben counters that Meta's ad auction already acts as the world's largest autonomous agent anticipating user intent. | |
| Sponsor: Stripe Managed Payments for Mobile Apps | 7 | 4 | 3 | 5 | John dissects the software sell-off, arguing against the death of SaaS by pointing out that Anthropic recently deployed Workday. Ben clarifies that while systems of record survive, multiple compression occurs when high-multiple growth shifts into stable EPS value. | |
| The Stratechery Playbook and the Economics of Bundling | 6 | 6 | 5 | 5 | John suggests Substack could create a network-effect bundle similar to ShopPay. Ben strongly disagrees, explaining that top-earning creators face adverse selection incentives and will always defect from broad revenue-sharing bundles. | |
| Ben Thompson's AI-Assisted Writing Workflow | 4 | 3 | 2 | 1 | Ben details his use of LLMs for targeted background research and draft critiquing, avoiding automated prose generation while bypassing Google recency bias. | |
| The TSMC Bottleneck and AI Semiconductor Supply | 5 | 7 | 4 | 4 | Ben breaks down TSMC's fixed depreciation economics, explaining why the lack of logic foundry competition leads TSMC to rationally underbuild capacity. John probes whether upfront prepayments from hyperscalers could de-risk expansion. | |
| Live Shared Experiences, Provenance, and Global Innovation | 5 | 4 | 2 | 2 | The conversation covers educational testing in the LLM era, crypto digital provenance, the trade-offs of podcast retention vs poor shareability, and American innovation vs Chinese industrial scaling. | |
| Stripe Feedback and AI Training Data Markets | 5 | 5 | 3 | 1 | Ben delivers direct product feedback regarding a Stripe ACH edge case where a failed add-on cancels an entire plan, then discusses the need for scalable data generation markets for AI labs. | |
| Assessing Execution Across Big Tech Giants | 5 | 6 | 3 | 3 | Ben evaluates execution across Apple, Google, Microsoft, Meta, and Amazon, analyzing why Amazon's commodity cloud cost-reduction playbook faces headwinds during rapid generational AI leaps. |
Statements from this episode (34)
Ben Thompson: Taiwan is the best and most convenient place to live
“The best place to live is undoubtedly Taiwan. The one word everyone says for Taiwan sounds not that impressive, but the word is convenient. Like, it is the most convenient place to live.”
Ben Thompson: Taiwan is arguably the greatest Uber Eats market ever
“Taiwan is arguably the greatest Uber Eats market ever, because there, there's amazing options. It's all delivered by scooter, so it's always like 10 minutes to get dinner and things along those lines.”
Stripe host: Booking.com stock outperformed Google over the past 20 years
“If you invested a dollar in Booking.com and a dollar in Google 20 years ago, you made much more money as a Booking.com shareholder, and I think people don't appreciate that fact.”
Ben Thompson: OpenAI could stop model training and remain wildly valuable
“Actually, OpenAI could stop making models and be one of the most valuable companies in the world just because of ChatGPT. That's their most valuable asset.”
Ben Thompson: Contextual chat ads limit inventory and provoke bias suspicions
“If you're using the easiest possible way to target ads, which is based on the context of the conversation, we're going to show you a roughly relevant ad. Number one, your market's way smaller because you have to hope someone starts a conversation that matches …”
Thompson: Text is an inferior advertising medium compared to visuals
“Text isn't the best interface for ads. Obviously visuals are generally better.”
Ben Thompson: Google should monetize Gemini data through YouTube and Search
“I think the ideal outcome for Google is they never put ads in Gemini, but they understand so much about you because of what you do in Gemini, that they can then manifest that through ads on YouTube, through ads on Google, through ads on their other properties.”
Ben Thompson: Advertising business models are fundamentally incompatible with being platforms
“And the problem is that I think being an advertising based model is generally incompatible with being a platform. The whole point of a platform is you're letting something else shine, something else bring to the surface. You're just on your, you're the support…”
Ben Thompson: Digital ads are a societal good driving tech's consumer surplus
“Actually ads are good. They are societal good. They get, they are the driver of all the consumer surplus that tech throws off.”
Thompson: In algorithmic media, absolute volume of good content beats hit rate
“What actually matters, and this is a key thing, people get hung up on relative numbers, and what matters is absolute numbers. So it is better to have .one percent of your content is good if your content is like in the billions or trillions, as opposed to, oh, …”
Ben Thompson: TikTok data concerns are overblown compared to algorithmic control
“The problem with TikTok, and back then everyone was talking about user data. Who cares? Like the whole user data thing, people have this view of like the East German Stasi and like folders, like going through people's data. Like these are like vector databases…”
Ben Thompson: US TikTok divestiture is a disaster because China retains control
“The way that it shows up messily is we somehow do pass the law banning TikTok and it didn't get banned. And now it is sold, but China still controls the algorithm. So I think it's a big disaster.”
Thompson: Technology often strengthens legacy business models before killing them
“Technology often makes old business models even more powerful before it kills them.”
Ben Thompson: OpenAI's strategy resembles AOL's closed aggregator portal model
“OpenAI wants to be the place you do everything. They want to be like the aggregator. I think a critic would say people compare them to Netscape. I think the better analogy if you're an OpenAI skeptic would be AOL, where they want to be sort of like the interfa…”
Ben Thompson: Meta's Facebook advertising is the world's most successful agent
“Actually the largest and most successful agent in the world today is Facebook advertising, which is exactly right.”
Ben Thompson: Stratechery routine is three articles, one interview, three podcasts weekly
“Actually, my life is very, very optimized around. I write three pieces a week. I do an interview and I do three podcasts and all my focus and energy needs to be on that.”
Ben Thompson: SaaS growth slowdowns will force severe valuation haircuts
“People saying they're going to zero are wrong. But if the assumption is you're fine, but you're not going to be growing indefinitely, like that shift from thought of as being a growth company to being a stable. That's a haircut, and again, it's combined with t…”
Ben Thompson: Stratechery hit a $100k run rate in six months
“And at six months actually hit my one year goal, which was a thousand subscribers, thousand true fans, you know, it was a 100,000 dollar run rate.”
Ben Thompson: Independent subscription models cannot support long-form investigative journalism
“Like your traditional classic investigative journalism, six month sort of piece. It's not well supported by this. What did support that was the bundle. Having lots of different writers in one publication altogether.”
Ben Thompson: Top creators have no incentive to join Substack's bundle
“Yeah, the problem is the merchants who will drive, make that bundle valuable, have no incentive to join the bundle because they could make more just monetizing users directly.”
Ben Thompson: Spotify's bundle succeeded because it only negotiated with four entities
“Spotify is arguably the best bundle on the internet, but the reason why they were able to assemble the bundle is because they only needed to negotiate with like four, For entities.”
Ben Thompson: Google search has gotten worse due to recency bias
“And so I think I can say pretty authoritatively, Google has gotten worse. And I don't think it's Google's fault. I just think there's just, it's harder. What is Google's fault is they got so biased towards recency.”
Ben Thompson: AI is used for research and critique, not writing Stratechery
“So I do not use it to actually Generate any exact content, but targeted research and then critique. Those would be the two biggest use cases.”
Ben Thompson: The semiconductor market will have excess memory capacity soon
“We're going to believe it or not. We are going to have too much memory capacity in a few years because we have such a shortage right now.”
Ben Thompson: TSMC cut capex for two years following ChatGPT's release
“ChatGPT comes out. Every hyperscaler starts investing like crazy. What does TSMC do? They actually decreased their capex year over year, two years in a row. They did, they, there was no market response from TSMC to the ChatGPT moment.”
Ben Thompson: The tech industry faces a massive chip shortage around 2029
“I think we're looking at a massive short, shortness in chips in 20, 29 or so.”
Ben Thompson: Shared experiences and live events will appreciate under AI
“Things that are common experiences are going to be more and more valuable. Common content, common classroom time, live events, like shared experiences, because anything that's individualized is just going to be completely swallowed.”
Ben Thompson: Crypto will become increasingly important for digital authenticity
“And in that world, I think crypto as an identifier of Authenticity is going to be more and more important. Like at the end of the day, I want the original. I don't want a reproduction.”
Ben Thompson: Audio content suffers because listeners do not share it
“The problem is they don't share. Audio content is not shared.”
Ben Thompson: Media subscribers pay for ongoing production promises, not individual articles
“What you're paying for when you pay for me is you're paying for my ongoing production. I'm making a promise to you. I'm going to write something every day. And you're paying for that promise. You're not paying for the actual content. The content is a byproduct…”
Ben Thompson: Journalistic enterprises are doomed and need replacement by data markets
“What we clearly need is some sort of market mechanism For data generation that in the long run will replace what we're getting from journalistic enterprises, which are even more doomed than ever before.”
Ben Thompson: Apple builds great products but is a terrible platform steward
“The reality is when it comes to platforms, you have to build, the price of becoming a platform is making a great product. So Apple gets platforms so they make great products, and they're terrible stewards of the platforms. Microsoft is a great platform steward…”