The New York Times has spent the better part of a decade reinventing itself as something more than a newspaper. In a recent interview published by Stratechery, CEO Meredith Kopit Levien laid out the logic behind that transformation: a multi-vertical subscription bundle that pairs the flagship news product with Games, Sports, Cooking, and other standalone offerings. The strategy, Levien argues, creates a consumer relationship that is broader, stickier, and harder for competitors — including AI-generated content — to replicate.
The interview arrives at a moment when the Times occupies an unusual position in media. While much of the digital publishing industry has contracted under pressure from declining advertising revenue and platform dependency, the Times has grown its subscriber base and diversified its revenue streams. Its bundle approach borrows from the playbook of consumer technology companies: offer enough distinct products under one roof that cancellation becomes costly in terms of utility lost.
The Bundle as Business Model
The logic of bundling is well established in media economics. Cable television thrived for decades on the principle that aggregating channels of varying appeal into a single package maximized willingness to pay across a broad subscriber base. The Times appears to be applying a similar framework to digital subscriptions, but with an important distinction: each vertical — Games, Cooking, Sports — must function as a product worth paying for on its own, while also reinforcing the value of the whole.
This is a non-trivial operational challenge. Games, most notably the acquisition of Wordle and the expansion of the puzzle portfolio, requires product development talent more commonly found in gaming studios than newsrooms. Sports coverage, particularly after the Times acquired The Athletic, demands a different editorial voice and cost structure than traditional reporting. Cooking competes in a space crowded with free recipes and video content. Each vertical pulls the organization in a different direction, and the bundle only works if each piece delivers enough value to justify the composite price.
The advertising side of the business adds another dimension. Levien has pointed to growth in the Times' advertising revenue, a notable countertrend in an industry where programmatic commoditization has eroded publisher pricing power. A large, engaged, and identifiable subscriber base gives the Times something most digital publishers lack: first-party data and a premium context that brand advertisers are willing to pay for. The experimentation with vertical video suggests the company is also watching where attention is migrating and attempting to meet it without abandoning its editorial identity.
Human Expertise as an AI Defense
The most strategically consequential thread in the interview concerns generative AI. The Times has taken an aggressive legal posture against OpenAI, filing suit over the use of its content in training large language models. But beyond litigation, Levien frames the company's broader defense in terms of what AI cannot easily replicate: human expertise, editorial judgment, and institutional credibility.
This is a bet with significant implications. If generative AI reaches a point where it can produce content that is functionally indistinguishable from expert journalism — accurate, well-sourced, contextually rich — the moat narrows considerably. If, on the other hand, readers continue to value the provenance of information and the accountability that comes with a named institution and named reporters, the Times' investment in human capital becomes a durable competitive advantage.
The company's internal use of AI adds nuance to the picture. Like most large organizations, the Times appears to be adopting AI tools for operational efficiency while simultaneously fighting to prevent its content from becoming free training data for competitors. This is not a contradiction so much as a reflection of the current landscape: AI is both a tool and a threat, and the strategic question is where to draw the line.
What remains open is whether the bundle model and the human-expertise thesis reinforce each other or exist in tension. A bundle optimized for breadth and engagement could drift toward the kind of content AI generates cheaply. A commitment to deep human expertise could limit the scalability that bundles require. How the Times navigates that tension — and whether the legal battle over training data reshapes the economics of AI content — will determine whether the strategy holds.
With reporting from Stratechery.
Source · Stratechery
