Storj has acquired Valdi

Announcements
1

Hello Storj Community,

As you may have seen, today Storj announced its acquisition of Valdi, the leading distributed AI compute (GPU) provider. Valdi is a groundbreaking provider of on-demand GPUs widely used for AI training and inference in data-intensive industries including technology, research and life-sciences. Like Storj, Valdi enables providers to monetize spare capacity from underutilized assets. Storj partnered with Valdi starting in 2023 and integrated our products soon after. We recognized the value and impact of readily available compute and storage resources to meet the needs of AI innovators to test and prove concepts without upfront technology overhead. As such, we’ve decided to combine the two companies to offer a more complete cloud solution.

Here are FAQs that we hope are helpful to our community as we integrate our offerings.

Q. Why did the two organizations decide to join forces?

A. We’ve been working together and saw we can bring value to the market faster together, taking advantage of each organization’s strengths. Together we have decades of cloud experience.

Storj has been poised to expand strategically and both companies prioritize a sustainable approach to provide for our customers’ needs

This transaction is a step in Storj’s strategic roadmap towards building a new distributed cloud, aligned with the demands of tomorrow’s innovation landscape. This also empowers Valdi with Storj’s robust infrastructure and resources - coupled with the fast, secure, global and sustainable cloud storage that most of Valdi’s target customers also need.

Q. Does this mean I will be able to lease out my CPU or GPU to Storj, like I currently do with my hard drive? How can one become a GPU provider to the new Storj/Valdi compute network?

A. Currently Valdi only accepts GPUs in data centers. Storj is exploring the potential to expand on this. In the meantime, if you have access to data center GPUs, we would love to talk to you about becoming part of our provider network. Please reach out to join@storj.io.

Q: What can I expect as a storage node provider?

Valdi’s product offering already utilizes Storj’s cloud object storage as its main storage framework. As Valdi’s customer base grows, you can expect to see additional load coming from a wide variety of data centers worldwide, in usage patterns and use cases where GPUs are popular.

Q. What kind of GPUs are available on VALDI?

  1. Everything from H100s, A100s, L40S, A6000, 4090s and 3090s.

Q. Will Valdi and Storj operate as one company or separate entities?

A. Valdi is now a division of Storj and will operate as a business unit supported by Storj marketing, sales, and support resources.However, we are anticipating a great deal of common customers and suppliers/community members, and are actively working to ensure that they can interoperate smoothly.

Q. What future goals do the combined entities now have?

A. Storj is building a new global cloud for the future of innovation. This is a big step in that direction. We have a long-term vision for the distributed cloud - just as Amazon started with object storage and moved to compute, we started with distributed object storage and have now added distributed GPU.

Ultimately, by enabling underutilized assets around the world to leverage in a new era of cloud cloud computing that is high-performance, sustainable and secure, we:

  • democratize access to cloud computing, both for customers and providers
  • enable customers to access cloud storage and cloud compute that is globally faster, more secure, and more durable at a fraction of the cost and a fraction of the carbon footprint

Q. What is the role of the STORJ Token?

A. A portion of the transaction was funded by STORJ token and a portion was in cash. We are planning to enable purchase of GPUs using STORJ token in the future, as well as an option for GPU operators to be compensated with STORJ token.

25 Likes
2

Wow, that’s massive.
It’s also a great change in the direction of StorJ, it’s exiting times and I’d love to see where this goes : )

7 Likes
3

Congrats Storjlings: this is great news! The STORJ token was always supposed to be used to expand the business: so using it as part of an acquisition sounds ideal! :star_struck:

(Thinking-out-loud: All the recent Chia refugees coming to Storj are also often dragging along a bunch of GPUs they used for plot-compression. But I understand not wanting to take on vast.ai just yet :wink: )

5 Likes
4

Nice grab! Incredible how only two other people have replied to this news on the forums. Storj is increasingly one of my best assets, looking forward to the road ahead, keep up the great work!

3 Likes
5

Yeah, and did people miss this:

“We are planning to enable purchase of GPUs using STORJ token in the future, as well as an option for GPU operators to be compensated with STORJ token.”

That’s a whole new use for STORJ tokens! It’s still a utility token: but this is more utility!

3 Likes
6

How does this fit the ‘utility token’ tale? :thinking:

7

Easily. It’s a payment method. Not securities, not shares, not a commodities, etc., it’s an Utility Token. That’s mean, that you can buy the service with them in the Ecosystem. Also, you will be paid for the provided resources by the STORJ tokens as well.
So, what is bother you?
We also want to use STORJ tokens as a payment method for Valdi load as well.
I hope that the extending of the Storj Ecosystem is a good move, because we have so much in common.

3 Likes
8

I am most pleased with this sentence, as it suggests that further products are planned. I am sure that if STORJ one day offers multiple decentralized products like the big hyperscalers do (GPU, CPU, Storage etc.), there will be many more potential customers willing to use one or more of the services.

I think one big issue right now are the lock-in effects of the existing platforms. E.g. when I am currently keeping my data in AWS and want to use it to train some ML-models, the easiest/cheapest way would be to use AWS resources for that as well. If I want to keep the data in AWS and use compute resources from Azure, I always have to pay high egress fees that could be avoided.
I guess that is currently also one reason why potential customers are not moving to STORJ, because they would always have to pay the egress fees when they want to use their data somehow.

So for me it’s great news to hear. Once there will be more products and customers can make use of their stored data directly in the decentralized STORJ ecosystem, I am sure that utilization will increase a lot and we will all benefit from it.

2 Likes
9

Yes, that’s the plan. But we have a way before us to achieve that.
Now we have a GPU processing, which is on top right now (actually you may use a CPU too, by the way, however, GPU usage would be better right now).

2 Likes
10

I am sceptical when I see such announcements, given that in many discussions Storj says it is a company with limited resources to make certain things happen, specially technical ones progress in short term.
I suppose the new company has its own source of resources but still makes the whole thing more complex to manage and it is natural to think that this somehow diverts the attention and energy Storj should be putting in improving its product in different ways.

Interesting also to see that part of the transaction was funded by STORJ which currently has a lower value back what it was in Oct 2023, so it is certainly a bet for Valdi stockholders.

11

When you join forces it also means join resources. So Valdi is a division of Storj and we want only to help each other, not breaking processes. I always thinking that join forces is almost always a good move, we can use each other’s strengths to achieve better results (revenue and profit in particular).

4 Likes
12

Hi @Bryanm ,

Congratulations on the acquisition of Valdi! This is a significant move for Storj, and I can see how integrating Valdi’s GPU resources could open up new avenues for enhancing backup solutions, especially in data-intensive applications.

Given that Valdi’s product already utilizes Storj’s cloud object storage, I’m curious about how this integration might specifically impact the efficiency and scalability of backup solutions. For instance, could the distributed GPU network be leveraged to optimize deduplication or compression processes in backup workflows? Additionally, how might this acquisition influence the development of automated redundancy checks and recovery mechanisms within Storj’s ecosystem?

Looking forward to seeing how this evolves and the potential innovations that could arise from this strategic combination.

Best regards,
George

13

The backup is usually performed on the server, the GPU computing mostly used in the AI training/running industry/researches and the game industry. So it will not impact the efficiency of backup solutions anyhow at least now (I do not know any backup solution, which could use a GPU).

Do you want to implement a backup solution which could use a distributed GPU processing?

14

2 posts were split to a new topic: What’s the backstory on the Chia refugees?