Special Price For Institutions

New Tesla GPU Prices

Up to $1,500 on Tesla cards - Now Includes VOLTA!

February 8, 2017 at 10:41 AM
Updated October 19, 2017 at 12:28 PM
Tesla GPU

October 19, 2017 Update: Due to the overwhelming success of this promotion, we are pleased to announce an extention to this current promotion, as well as the inclusion of the newest member of the Tesla family, the Volta-based Tesla V100!

The discounts are now as follows:

Tesla V100 Servers:
- 32GB: $1,500 off each card (Part# 900-2G500-0010-000)
- 16GB: $1,250 off each card (Part# 900-2G500-0000-000)

Tesla P100 Servers:
- 16GB: $1,500 off each card (Part# 900-2H400-0000-000)
- 12GB: $1,250 off each card (Part# 900-2H400-0003-000)

Tesla P40 Servers:
- 24GB: $1,500 off each card (Part# 900-2G610-0000-000)

For your convenience, we have put together this chart to show you all Tesla GPUs and their prices for comparison:

NVIDIA Tesla GPU Prices

Tesla GPU Price* Release Date
Tesla K80 24GB $4,599 November 2014
Tesla M4 $1,599 November 2015
Tesla M40 12GB $4,499 November 2015
Tesla M40 24GB $5,099 November 2015
Tesla M10 32GB $2,299 May 2016
Tesla P100 12GB $4,699 June 2016
Tesla P100 16GB $5,899 June 2016
Tesla P4 8GB $1,899 September 2016
Tesla P40 24GB $5,699 September 2016
Tesla V100 16GB $8,699 May 2017
*Before discounts - Prices accurate as of June, 2018

Institutional/Education customers are a beloved segment of our user base here at Thinkmate. Using advanced GPUs, these are the people working — sometimes late into the night — to create solutions to some of the world’s toughest challenges, like curing diseases and creating artificial intelligence.

But many of these customers don’t have it easy. Many are provided with limited funding or budgets, preventing some great ideas from even getting off the ground.

Making it more accessible

We are constantly thinking about creative ways to make our systems more accessible to EDU customers. We’ve had tremendous success when we were the first to market server systems with support for consumer-grade Geforce GTX and Titan X cards. However, while that has worked fine as an economical introduction to GPU computing, we did not have any EDU-specific offers for the cards actually dedicated to GPU computing: Tesla.

With pressure to deliver their work more intense than ever, serious researchers demand the highest performance and reliability. They need datacenter-grade computing and direct support to overcome technical or operational problems.

Today, we’re providing that. We are announcing substantial discounts for latest-generation Tesla cards based on the Pascal architecture — Providing up to a whopping $1,500 discount per card. These discounts are available immediately.

Currently, the discounts apply to Tesla P40 and P100 PCIe cards, when configured as part of any compatible Thinkmate system.

Tesla Discount Table

Yes, the discount is per card. With no limits. That’s the kind of fine print that almost deserves to be a headline.

We are not announcing an end date for this offer yet, but if you or anyone you know is in the market for these systems, we recommend taking action as soon as you can. Don't forget to share with your colleagues!

Welcome to datacenter-grade computing

How do I take advantage of this offer?

A Thinkmate sales engineer needs to manually apply these discounts as they will not appear on the Thinkmate configurator. You can browse and configure systems normally and use the Submit for Review button in the configurator bar. This will get forwarded to a Thinkmate sales engineer who then will be able to apply any discounts. You can also call us at 1-800-371-1212 and we’ll help you put together a system if you need some direction. We might even be able to get additional discounts that are only available to EDU customers.

Who is eligible for this offer?

This offer is open to:

- Students or employees with .edu or .ac email addresses

- State or local government-sanctioned agencies that purchase on behalf of schools

See specific eligibility details at the end of this article.

Finding a compatible server

We work hard to create the most modern, informative and complete system in the industry, and we’ve made it easy for you to find and configure a server compatible with the Tesla card you plan to use.

NVIDIA Tesla P100 Servers

NVIDIA Tesla P40 Servers

Full Offer Terms

A qualified Educational Institution for this promotion is defined as:

  • A public or not for profit K-12, vocational school, junior college, college, university, scientific or technical institution with a primary focus on teaching students.
  • For use by students or employees of an education institution with .edu or .ac email addresses
  • If .edu or .ac are not in email domain then the institutions name should be clearly recognizable as educational affiliation.
  • State or local government-sanctioned agencies that purchase on behalf of a pool of school districts, community colleges, or other small educational institutions are also eligible
  • Organizations and people directly related to the academic institution (e.g. hospitals that are wholly owned and operated by an education institution) are also eligible.
  • For-profit institutions (e.g. training centers that operate a business) as well as students, teachers, lecturers, researchers and tutors affiliated with such institutions are not eligible

Additional Education Pricing Program Terms:

  • Non-commercial use only
  • Research and education use only
  • NVIDIA reserves right to watermark streams (i.e., include an NVIDIA logo) on all NVIDIA product

Thinkmate reserves the right to admit or deny any party requesting to participate in this program.

Thinkmate reserves the right to modify the terms of this offer at any time without notice.

Tesla P100 Specs

NVIDIA Tesla P100

The Tesla P100 is a computing accelerator based on the Pascal architecture, launched late 2016. The P100 was designed to excel in mixed-workload HPC applications.

GPU Architecture NVIDIA Pascal
NVIDIA CUDA® Cores 3584
Double-Precision Performance 4.7 TeraFLOPS
Single-Precision Performance 9.3 TeraFLOPS
Half-Precision Performance 18.7 TeraFLOPS
Predecessor None
(Brand new product)
Sucessor Not yet announced
Commonly used for
Deep learning, HPC
System Interface PCIe Gen3
Max Power Consumption 250W
Compute APIs CUDA, DirectCompute, OpenCL™,OpenACC