Repeated Rejections of AWS Activate even though we satisfy all criteria
Hello,
We applied for $100K through nvidia inception several times after we received $5K and $25K . We provided that we spend $20K to nvidia GPUs but we are still being rejected from that criteria. We are providing proof of $20K spent through AWS credits, sending reports and everything and since 15th of March I couldn't get a clear answer to the application.
I sent the email to nvidiainception team's email at AWS. We've given all the proofs and showed that we spent $33K ~5K of this is not credits we spent from pocket money to GPUs on other services as well.
I've sent an email with email titled: Fwd: [ROUTED] [ANYTHING] AWS Activate Case: 01219450 [
Our aws activate application through nvidia inception is done via is the AWS account email AWS account id is 794075227583
We still received this email:
Thank you for applying for AWS Activate Credits. After reviewing your application, we are unable to approve your request at this time, as your startup has not yet demonstrated NVIDIA technology adoption, including GPU usage and SDK/API usage on AWS.
USD 566.40 (INVOICE ATTACHED) This was bought 14 days h100 capacity block reservation CapacityBlockFee:p5.4xlarge See ecsv_9_2025.csv
October 2025 (NO USAGE) NOVEMBER 2025 ( $14886.83 spent) $11103 credits spent to GPU instances (Peak month!) December 2025: $6,674.88 spent to GPU instances JANUARY we Used modal.com see attached screenshots from modal.com February 2026: $838.97 as shown on screenshots and shared costs-3.csv March 2026: $1,106.97 as shown on screenshots and shared costs-3.csv
We also clearly showed that we train AI models https://blog.audn.ai/posts/personalized-to-communitized-rl
- Newest
- Most votes
- Most comments
I understand your frustration with the repeated rejections of your AWS Activate application through the NVIDIA Inception program. However, I need to clarify that I cannot directly intervene in or override AWS Activate application decisions, as these are handled by specific AWS teams following program criteria.
Based on your situation, here are some recommendations:
-
Direct Support Channel: Since you've already been in contact with the NVIDIA Inception team at AWS and have a case number (01219450), continue working through that support case. The team handling your case has access to your full application details and can provide specific guidance on what's needed.
-
Documentation Clarity: Ensure your GPU usage documentation clearly shows:
- Direct spend (not just credits) on NVIDIA GPU instances
- Consistent usage patterns demonstrating ongoing adoption
- Clear connection between your startup's activities and NVIDIA technology usage
- SDK/API usage beyond just compute instances
-
Timing Considerations: I notice some dates in your message reference 2025-2026, which appear to be future dates. Make sure all documentation reflects actual historical usage with correct dates.
-
Alternative Escalation: If you're not getting resolution through the existing case, you can:
- Request escalation within your current support case
- Reach out to your NVIDIA Inception program manager directly
- Contact AWS Activate support separately to understand the specific criteria not being met
The rejection message indicates the review team needs clearer demonstration of NVIDIA technology adoption. Work with the support team to understand exactly what additional evidence or clarification they need to approve your application.
This is AI answer the core reason our applications got stuck 1 month despite having all the proofs that we satisfy them
Timing Considerations: I notice some dates in your message reference 2025-2026, which appear to be future dates. Make sure all documentation reflects actual historical usage with correct dates.
Someone asked this rePost-User-0257722 2 days ago You started using AWS in Oct. Then only used AWS for two months each month declining spend by over 50%, then moved to modal.com. Why would AWS approve this?
My answer ozguratpenclaw a few seconds ago Reason they should: Because february and march we came back to AWS satisfied minimum GPU spending which was way more than $500 We trained the models in October and november and started using inference which you might expect why training cost so much when compared to inference.
Noone trains GPU models constantly in same pace without actually collecting what they've got from the model! As someone who worked at Softbank funded AI Unicorn Wayve.AI I know SOTA AI work!
Reason they should: Because february and march we came back to AWS satisfied minimum GPU spending which was way more than $500
We trained the models in October and november and started using inference which you might expect why training cost so much when compared to inference
Your prejudice to determine what's good or bad is inherently wrong based on personal opinions and data tells otherwise.