OpenAI’s reported plan to buy Google TPUs matters less as an immediate revenue threat to Nvidia than as evidence that its biggest AI customers are building credible alternatives.
Nvidia has a reason to pay attention. OpenAI is reportedly planning to buy Google’s Tensor Processing Units, or TPUs, according to The Information, giving one of Nvidia’s largest customers another major accelerator architecture. The financial terms, chip volume and timing have not been disclosed, and neither OpenAI nor Google has confirmed the purchase. That makes the headline less dramatic than it first appears, but the strategic signal is hard to miss.
The Diversification Was Already Underway
OpenAI began renting Google’s TPUs in 2025. It then unveiled its own custom inference accelerator, Jalapeño, with Broadcom in June 2026, while Google has moved beyond keeping TPUs primarily inside its own cloud. On October 27, Oracle and Google announced that Google TPUs will be offered through Oracle Cloud Infrastructure from 2027, making OCI the first hyperscaler outside Google to offer them. The OpenAI report therefore arrives in the middle of a broader change in how AI compute is being supplied.
The important question is not whether OpenAI will stop buying Nvidia chips. There is no evidence of that. The question is whether large AI companies can split workloads among Nvidia GPUs, Google TPUs and custom accelerators according to price, availability, latency and energy efficiency. If they can, Nvidia still benefits from enormous industry growth, but its customers gain bargaining power.
Nvidia’s Current Numbers Are Still Hard To Ignore
Nvidia’s latest reported figures show why it would be premature to describe the company as being displaced. Data Center revenue reached $89.0 billion in fiscal Q2 2027, up 117% year over year and 18% sequentially, while total revenue reached $96.2 billion. Nvidia guided to approximately $108 billion of total revenue for the following quarter and said supply was expected to remain a bottleneck at least through fiscal 2028.
Figure 1. Nvidia Data Center revenue by fiscal quarter. Source: NVIDIA investor releases.
Figure 2. Year-over-year Data Center revenue growth reported by NVIDIA.
The Real Risk Is Pricing Power
The more meaningful threat is pricing power. Every credible alternative gives a buyer more leverage in negotiations, even if Nvidia remains the dominant supplier. If OpenAI can place some inference workloads on Google TPUs and others on its own accelerators while reserving Nvidia systems for workloads where they remain superior, the company does not need to replace Nvidia outright to change the economics of the relationship.
That is why Nvidia’s full-stack strategy matters. The company is selling much more than a processor: GPUs, CPUs, networking, storage, software and complete rack-scale systems. Its latest disclosures show Vera Rubin systems being deployed by major cloud partners, while Nvidia is also expanding its networking and inference products. A competing accelerator therefore has to compete with an ecosystem, not simply with the silicon on a single board.
Google Is Finally Commercializing The TPU Advantage
Google has spent more than a decade developing TPUs for its own workloads, but the commercial opportunity is now becoming much more visible. Oracle’s agreement creates an independent cloud distribution channel, while Google already has experience serving major customers such as Anthropic through its cloud infrastructure. An OpenAI purchase would be another step toward proving that TPU economics can work outside Google’s own applications.
That matters because a chip alternative becomes much more useful when customers can obtain it without rebuilding their entire infrastructure strategy. Oracle’s role is therefore strategically important: it lets customers compare accelerator architectures inside a familiar cloud environment rather than treating TPU adoption as a wholesale migration to Google.
OpenAI Is Building Leverage From Both Directions
OpenAI’s own silicon strategy makes the Google relationship even more revealing. In June, OpenAI and Broadcom unveiled Jalapeño, an inference accelerator designed around OpenAI’s models and serving requirements. OpenAI said the chip was developed as part of a multi-generation platform and would be deployed at gigawatt scale with data-center partners. Later performance results showed higher throughput per unit of power and lower latency on selected workloads, although the commercial scale of deployment remains the more important question for investors.
The logic is straightforward. OpenAI wants compute to become more abundant and cheaper, but it also wants control over the infrastructure economics behind its models. Owning some of the stack gives it leverage with outside suppliers; having multiple outside suppliers gives it leverage with each other. Nvidia can remain the largest supplier and still face a customer that is increasingly unwilling to depend on one architecture.
What Would Actually Worry Nvidia Investors
The next datapoints should be specific. Investors should look for the number of Google chips involved, the workloads they will handle, whether OpenAI is buying hardware directly or accessing capacity through a cloud provider, and whether the arrangement expands from inference into large-scale training. A relatively small TPU deployment would be strategically interesting but financially minor for Nvidia. A sustained, multi-gigawatt deployment across important OpenAI workloads would be much harder to dismiss.
Google’s disclosures matter just as much. Alphabet needs to demonstrate that TPU demand can become a meaningful external business, while Nvidia needs to show that its integrated systems continue to command a premium even as customers gain alternatives. The competitive question is moving away from ‘Who has the fastest AI chip?’ and toward ‘Who can deliver the lowest useful cost per token at scale?’ That is a much tougher market to dominate.
The Bottom Line
So, should Nvidia be worried? Yes, but not because one OpenAI order suddenly threatens its revenue. Nvidia’s latest numbers show a company still operating at extraordinary scale, with Data Center revenue growing more than 100% year over year and supply constraints extending well into the future.
The concern is structural. OpenAI, Google, Oracle and other large technology companies are building a market where compute can be sourced from several architectures, while custom silicon increasingly sits alongside merchant GPUs. Nvidia’s moat is still substantial because of CUDA, systems integration, networking and software, but the moat is being tested from several directions at once. The OpenAI-Google development is another sign that the AI chip market is becoming less dependent on a single architecture — and that is the part Nvidia investors should watch.
Original Sources
NVIDIA Q2 FY2027 results — https://investor.nvidia.com/news/press-release-details/2026/NVIDIA-Announces-Financial-Results-for-Second-Quarter-Fiscal-2027/default.aspx
OpenAI and Broadcom: Jalapeño — https://openai.com/index/openai-broadcom-jalapeno-inference-chip/
OpenAI full-stack compute strategy — https://openai.com/index/the-full-stack-behind-abundant-intelligence/
Oracle and Google TPUs on OCI — https://www.oracle.com/news/announcement/ai-world-oracle-brings-google-tpus-to-oracle-cloud-infrastructure-2026-10-27/
Reuters report on OpenAI and Google chips — https://www.reuters.com/business/openai-plans-buy-googles-ai-chips-information-reports-2026-10-28/
Johan Shamshad is a financial markets writer at Dave Finances covering cryptocurrencies, trading platforms, brokers, fintech, financial regulation, and developments across global markets. He previously worked at Gulf News, adding newsroom experience to his coverage of fast-moving financial and digital-asset markets.
His work focuses on identifying market-moving events, company developments, regulatory changes, product launches, and shifts in trading and financial infrastructure.
Johan contributes news and analysis designed to help readers understand not only what happened, but why a development matters and how it may affect the wider financial landscape. You can reach out to him via his social media accounts:
Linkedin: https://www.linkedin.com/in/johan-shamshad-742851262/
X: https://x.com/Yasmine_FX
Investing: https://www.investing.com/members/contributors/279781574