Enterprises Favor Non-Nvidia Chips Over Next-Gen GPUs in Evaluations

Enterprise buyers are more inclined to place non-Nvidia chips ahead of Nvidia's next-generation GPUs on their evaluation lists for the upcoming cycle, according to VentureBeat's July VB Pulse survey. Out of 170 AI infrastructure respondents, 39.4% are likely to evaluate alternative accelerators such as AWS Trainium, Google TPU, AMD Instinct, or Intel Gaudi within the next 12 months. This compares to 25.3% who plan to evaluate Nvidia's Blackwell or other next-gen Nvidia GPUs, representing a 14-point gap.

While Nvidia remains the default choice for most production environments, organizations are actively building optionality into their accelerator strategies rather than treating Nvidia as the sole option. This trend reflects a broader pattern where enterprises are optimizing their existing AI infrastructure before making major platform changes.

The survey data also suggests a decrease in urgency for immediate platform changes. Despite increased AI infrastructure utilization and exploration, only 28.8% of respondents expect a platform change within three months, down from 38.3% in June. Microsoft Azure saw the largest growth in production adoption, while Google Gemini and OpenAI maintained strong positions.

Utilization of enterprises' own GPU infrastructure also significantly improved between June and July, with the share operating at half capacity or less decreasing from 83% to 69%. Furthermore, uptime and reliability emerged as key effectiveness measures. Companies are setting higher bars for their infrastructure, increasingly focusing on practical performance and cost-effectiveness per workload.

Interest in Nvidia alternatives is concentrated among key decision-makers, with that share increasing since June. This indicates that accelerator diversity is becoming a strategic infrastructure consideration. Enterprises are seeking greater control and flexibility over their AI computing resources.