Meta is preparing four generations of AI chips, while Uber will launch Zoox robotaxis in Las Vegas

The AI market has sent three strong signals at once: corporate demand for computing power remains high, Meta is accelerating its transition to proprietary hardware, and robotaxis are beginning to enter familiar consumer applications. Oracle, Meta, Uber, and Zoox have come into focus — companies from different segments, but with one common theme: AI and autonomous services are rapidly transforming from expensive experiments into basic infrastructure.

Three Market Signals

Oracle reported strong third-quarter sales and confirmed a confident forecast for the entire fiscal year. The key takeaway from this episode is not the report itself, but the reason behind such dynamics: demand for AI computing remains high, and businesses continue to pay for the infrastructure needed to train, deploy, and scale models. This is an important market indicator because, after a wave of investments, many expected a slowdown. So far, we don't see one.

• Oracle shows that corporate demand for AI infrastructure remains resilient.
• Meta plans to deploy four generations of proprietary AI chips by the end of 2027.
• Uber will add Zoox robotaxis in Las Vegas in summer 2026.
• Zoox reaches a mass audience through the Uber app, not just through a separate pilot.

Together, these news stories show how the entire AI product creation chain is changing. On one hand, cloud and computing providers are earning money from growing demand. On the other — platforms with billion-strong audiences are trying to reduce their dependence on external chip and service suppliers. And at the consumer level, autonomous technologies are beginning to get proper distribution through already-familiar applications.

Meta Goes into Hardware

Meta's plan to deploy four new generations of proprietary AI chips by the end of 2027 looks like an attempt to accelerate vertical integration. When model workloads grow rapidly, purchasing third-party accelerators becomes not only expensive but also strategically risky: delivery timelines, capacity shortages, and limited flexibility can slow down the launch of new features. Proprietary chips give the company greater control over where and how its AI tasks are executed.

For Meta, it's also a matter of economics. The more actively the company integrates generative capabilities into advertising, recommendations, search, and assistants, the higher the constant need for computing power. If part of this load can be shifted to proprietary hardware, there's a chance to better manage costs and optimize infrastructure for specific use cases.

Four generations in a relatively short timeframe — an ambitious pace that shows Meta views AI chips not as a side R&D project, but as one of the key assets of the coming years.

Uber Launches Zoox

The Uber-Zoox partnership also looks more significant than just another autonomous transportation announcement. The Zoox service, which is owned by Amazon, will be available to Uber customers in Las Vegas as early as summer 2026. This means robotaxis will arrive not through a niche app for early tests, but through a platform that people already use for regular rides. For the market, this is a completely different level of distribution.

This format is convenient for both sides. Zoox gains access to ready-made demand and a familiar user interface, while Uber expands its offering without needing to build a full autonomous transportation stack from scratch. If the launch goes smoothly and the service proves intuitive for passengers, the "autonomous fleet plus major booking platform" model could become the standard scheme for robotaxis entering the urban market. Las Vegas, in this sense, becomes an important testing ground for the next stage of commercialization.

What This Means

These news items fit well into one picture: money for AI infrastructure continues to flow, large platforms are building their own hardware, and autonomous transportation is gaining a mass sales channel. For the market, this is a sign that competition is shifting from loud demos to control over computing, cost management, and distribution.