Uber’s product strategy has always been a balancing act: expand the surface area of what the platform can do, while keeping the experience coherent enough that riders and drivers don’t feel like they’re being dragged through a maze of experiments. In a wide-ranging conversation with TechCrunch, Uber Chief Product Officer Sachin Kansal framed the company’s next phase as less about chasing every opportunity at once and more about building a set of “adjacent” products that reinforce the core promise—reliable movement of people and goods—while quietly laying the groundwork for much bigger bets like autonomous vehicles and financial services.
What makes this interview notable isn’t just the list of initiatives—hotels, robotaxis, AI upgrades, fintech ambitions—but the way Kansal described the underlying philosophy. Uber, he suggested, is trying to avoid becoming “everything for everyone,” not as a slogan, but as an operational constraint. The company wants to grow, but it also wants to reduce the cognitive load of complexity: complexity in partnerships, complexity in data pipelines, complexity in product surfaces, and complexity in how AI is introduced into real-world workflows.
That theme shows up repeatedly across the conversation, from Uber’s push into hotels to its increasingly intricate relationship with Waymo, and even in how Uber is organizing its autonomous-vehicle data work through AV Labs. And it culminates in a practical question: where does AI actually land in the product, and what does it change for the people using Uber day to day?
A platform that starts to look like travel infrastructure
Uber’s move into hotels may sound like a natural extension—after all, if you can get someone from point A to point B, why not help them find a place to stay? But Kansal’s framing was broader than a simple add-on. He described hotels as part of a “next experiences” push: a shift from being only a transportation marketplace to becoming a more complete travel utility.
The key nuance is that Uber isn’t positioning hotels as a standalone brand that competes with established hotel booking platforms on their own terms. Instead, the pitch is that Uber can bring context. Riders already use Uber when they’re planning or executing trips; drivers already operate within local geographies and time windows. Hotels, in this view, are another layer that can be surfaced at moments when users are already thinking about travel logistics.
This matters because it changes the product problem. Transportation is immediate and transactional: request, match, ride, pay. Hotels are more planning-oriented and often involve longer decision cycles. That means Uber has to think differently about discovery, pricing presentation, cancellation policies, and trust signals. It also means Uber has to integrate hotel-related flows without turning the app into a cluttered marketplace of unrelated categories.
Kansal’s emphasis on not being “everything for everyone” reads like a warning against exactly that outcome. If Uber adds too many categories without a clear logic for when and why they appear, the platform risks losing the simplicity that made it successful in the first place. Hotels, then, are not just a new vertical; they’re a test of whether Uber can extend its travel utility while preserving the clarity of the core experience.
Robotaxis aren’t a single timeline—they’re a relationship
If hotels represent expansion into a familiar consumer domain, robotaxis represent expansion into a domain where timelines, safety requirements, and partnership structures can shift dramatically. Kansal addressed Uber’s autonomous vehicle ambitions and its relationship with Waymo in a way that underscored how complicated these efforts become once you move from prototypes to real operations.
The most important takeaway wasn’t a promise of a specific launch date. It was the idea that Uber’s AV work involves an evolving relationship with Waymo—one that becomes more nuanced as programs mature. That nuance is often missing from public discussions about autonomy, which tend to treat partnerships as either “on” or “off.” Kansal’s description suggests something more realistic: autonomy programs evolve through stages, and those stages affect everything from operational constraints to data sharing to how the service is scaled.
In other words, robotaxis aren’t just a technology milestone. They’re an ecosystem milestone. They require coordination across mapping, fleet operations, safety validation, customer experience design, and regulatory compliance. Even when two companies share a common goal, the path to autonomy can diverge based on what each party is optimizing for—coverage, reliability, safety metrics, or speed of iteration.
Uber’s role in this ecosystem, as Kansal described it, is not simply to “wait for autonomy” but to build the product and data capabilities that make autonomy usable. That’s where AV Labs enters the story.
AV Labs: turning experimentation into a data operation
Autonomous vehicles are often discussed as if they’re primarily a hardware and software engineering challenge. But anyone who has worked in autonomy knows that the real bottleneck is frequently data: collecting it, labeling it, validating it, and using it to improve models in a disciplined loop.
Kansal described AV Labs as evolving into a data operation rather than just an experimentation arm. That distinction is subtle but significant. Experimentation implies novelty and exploration. A data operation implies repeatability and measurement. It implies that Uber is building systems designed to answer questions over time: What kinds of scenarios cause failures? How do model updates change performance? Which data sources are most valuable? How do you ensure that improvements generalize beyond the lab?
This is also where Uber’s “not everything for everyone” philosophy becomes operational. Data operations can balloon quickly. If you try to support every scenario, every sensor configuration, every model variant, and every partner requirement simultaneously, you end up with a sprawling pipeline that’s hard to manage and even harder to evaluate. By treating AV Labs as a structured data operation, Uber is effectively choosing a discipline: focus on the data that improves outcomes, and build the tooling to learn efficiently.
There’s another angle here: autonomy work is not only about improving driving behavior. It’s also about improving the product experience around autonomy. Riders need to understand what to expect. Drivers and support teams need tools to handle edge cases. Regulators need evidence. All of that depends on data quality and traceability.
So AV Labs isn’t just a technical unit; it’s a bridge between engineering progress and operational readiness.
Financial services: the platform’s next layer of trust and utility
Uber’s financial-services ambitions have been discussed before, but Kansal’s explanation tied them more directly to the platform’s role in supporting both riders and drivers. Financial services, in his framing, aren’t a separate business that happens to use Uber’s user base. They’re a way to deepen the platform’s ability to serve people through the moments that matter—payments, earnings, and potentially other services that reduce friction.
For riders, the obvious value proposition is convenience: fewer steps, smoother checkout, and more predictable costs. For drivers, the value proposition is different: financial services can address cash flow timing, payout reliability, and the broader economics of operating as a driver. When you think about it, transportation marketplaces are inherently financial marketplaces. They coordinate money, time, and risk. Uber’s move into financial services can be seen as an attempt to internalize more of that coordination rather than outsourcing it to generic payment rails.
But there’s a strategic reason Uber would want to do this now. As Uber expands into hotels and other “next experiences,” the platform needs a consistent way to handle transactions across categories. If you’re going to sell travel-related services, you need payment and trust infrastructure that works seamlessly. Financial services can become the connective tissue that makes the platform feel unified rather than stitched together.
And again, the “everything for everyone” message matters. Financial services can easily become a sprawling set of features if the company tries to replicate the entire fintech landscape. Kansal’s approach, as implied by the interview, is to focus on what Uber can do well: leverage the platform’s transaction patterns and operational knowledge to deliver services that are directly tied to rider and driver needs.
AI: not a science project, but a product upgrade
Perhaps the most grounded part of the conversation was Kansal’s discussion of AI showing up in ways riders and drivers will actually notice. This is where many companies stumble. AI initiatives can become invisible—internal optimizations that don’t change the user experience—or they can become flashy but shallow, adding features that don’t meaningfully improve outcomes.
Kansal’s emphasis was on noticeable improvement. That suggests AI is being applied to problems where small gains compound: matching quality, ETA accuracy, route optimization, fraud detection, customer support automation, and driver assistance tools. These are areas where AI can reduce wait times, improve reliability, and make the platform feel smarter without requiring users to understand the underlying models.
It also suggests Uber is thinking carefully about where AI should be introduced. If AI changes too many parts of the experience at once, it becomes difficult to debug issues and difficult to build user trust. If AI is introduced only in back-end systems, users may not perceive value. The sweet spot is AI that improves the experience in measurable ways while remaining stable and explainable enough for operational teams.
There’s also a cultural element. Uber’s product organization has to coordinate AI with operations, safety, and partner ecosystems. That’s especially true when you consider autonomy and AV Labs. The same data discipline that supports autonomy can also support AI improvements in other parts of the platform. But the product rollout strategy still has to be careful: autonomy is high-stakes, while many rider-facing AI improvements are lower-stakes but still require reliability.
So the AI story is not “AI everywhere.” It’s AI where it matters, and where it can be validated.
The hidden thread: managing complexity without killing momentum
If you step back, the interview reads like a blueprint for how Uber intends to scale without losing control. Hotels expand the app’s scope. Robotaxis expand the company’s operational footprint into a regulated, safety-critical domain. Financial services expand the company’s responsibility for money movement and trust. AI expands the company’s reliance on data-driven decision-making.
Each of these expansions introduces complexity. Partnerships with Waymo introduce complexity in coordination and timelines