Oracle Expands NVIDIA AI Deal, Adds Google TPUs to Cloud

By Shefali Rana | BizGossips.in

NVIDIA remains a major force in the artificial intelligence industry, but the battle for AI computing power is becoming more complicated. Oracle is expanding its partnership with NVIDIA to a multi-gigawatt scale while preparing to bring Google’s AI chips to its cloud platform and exploring offerings from other chipmakers.

Announced on October 27, 2026, the developments point to a broader shift in the AI infrastructure market: technology companies want more computing power, but they also want more flexibility in choosing who supplies it.

This does not mean NVIDIA is losing its leadership position. However, Oracle’s strategy raises an important question: could the growing availability of alternative AI chips gradually reduce NVIDIA’s influence over the market?

NVIDIA and Oracle Are Betting Bigger on AI

Oracle and NVIDIA plan to expand Oracle Cloud Infrastructure (OCI) Superclusters using NVIDIA’s next-generation Vera Rubin NVL72 systems. NVIDIA Vera CPUs are also generally available on OCI, supporting the computing tasks AI agents require, including executing tools, testing code, and retrieving information.

The companies also plan to introduce NVIDIA HGX Rubin NVL8 systems, featuring eight Rubin GPUs in an x86-based server configuration.

These systems are designed to support demanding AI workloads, including model training, inference, and agentic applications. As businesses deploy increasingly sophisticated AI tools, demand for computing infrastructure capable of handling complex workloads is expected to remain an important industry driver.

However, the companies have not disclosed the financial value of the expanded agreement.

That leaves a critical business question unanswered: how much will this infrastructure cost, and how quickly can the investment translate into sustainable revenue and profits?

Google’s AI Chips Enter Oracle’s Cloud

Here is where the story gets particularly interesting.

Oracle and Google have announced plans to make Google’s Tensor Processing Units (TPUs) available through OCI. The service is expected to begin in calendar year 2027, with capacity expanding over multiple years.

TPUs are specialised processors developed for AI workloads, including model training and inference. Their availability alongside NVIDIA GPUs could give Oracle customers more options when selecting computing infrastructure for different applications.

Oracle is also planning to offer chips from Cerebras Systems and Intel-backed SambaNova Systems, expanding its options for AI inference workloads.

This matters because different AI applications can have different computing requirements. Some customers may prioritise performance, while others may focus on cost, energy efficiency, availability or compatibility with existing systems.

By offering multiple chip architectures, Oracle could give customers greater flexibility instead of requiring them to depend on a single hardware ecosystem.

Is NVIDIA’s Dominance Finally Facing Pressure?

The headline may sound like a direct challenge to NVIDIA, but the reality is more complicated.

NVIDIA benefits from a broad ecosystem of GPUs, networking technologies, software and developer tools. Competing chips must offer not just attractive hardware, but also the performance, reliability and software support needed to operate AI systems at scale.

Google’s TPUs and other specialised processors provide alternatives for certain workloads, but their availability does not automatically make them interchangeable with NVIDIA GPUs.

Still, Oracle’s approach highlights an important commercial shift.

NVIDIA can remain a leading supplier while customers become less dependent on a single chip provider.

For cloud companies, access to multiple suppliers may help manage capacity constraints, compare costs, and match hardware to specific workloads. Over time, this could increase competitive pressure on pricing and performance.

The key question is whether alternative chips can deliver sufficient advantages to encourage widespread adoption beyond selected applications.

The Real Battle Is About Cost, Not Just Computing Power

Building advanced AI infrastructure requires far more than purchasing processors. Data centres need electricity, cooling, networking equipment and substantial capital investment.

That makes the economics of AI infrastructure increasingly important.

Reuters reported that Oracle maintained its fiscal 2027 capital spending target at approximately $90 billion to $95 billion. Such spending reflects the scale of investment required to expand computing capacity, but it also raises questions about financing, customer commitments, and the returns these projects must generate.

If demand for AI computing continues to grow, Oracle could benefit from increased cloud usage and infrastructure revenue. If customer demand or returns fall short of expectations, the financial burden of expanding capacity could become more difficult to manage.

More chip options could help customers optimise their spending, but they will not eliminate the underlying cost of building and operating large-scale AI infrastructure.

What Happens Next?

Oracle’s strategy suggests that the future of AI computing may involve several chip architectures serving different needs.

For NVIDIA, the expanded Oracle partnership represents an opportunity to supply advanced infrastructure at scale. At the same time, Oracle’s plans to introduce Google TPUs and other alternatives could give customers more bargaining power and greater flexibility.

For Google, expanding TPU availability through Oracle could extend the reach of its specialised AI hardware beyond its own cloud environment.

For businesses using AI, the potential benefit is greater choice—but the actual savings and performance improvements will depend on how the different systems perform in real-world applications.

The BizGossips Take

Oracle’s strategy is not necessarily a rejection of NVIDIA. It is a sign that cloud providers want to expand computing capacity while keeping their infrastructure options open.

That distinction matters. Competition in AI chips is not simply about replacing one dominant supplier overnight. It is about creating credible alternatives, improving efficiency and giving customers more control over their computing costs.

The bottom line: NVIDIA remains central to the AI infrastructure race, but Oracle’s expanding chip portfolio shows why its dominance should not be confused with a lack of competition. The next phase of the AI boom may be decided not just by who builds the most powerful chips, but by who delivers the best combination of performance, availability and cost.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. Product availability, infrastructure plans, and capital expenditure targets are subject to company announcements and may change.