Mark Creaser

CEO at DSV Fund | Chairman Astrid Intelligence PLC

[NexisGen AI | SN70 @nexisgen_ai]

88,000 video clips.
123 hours of training data for video generation.
Built for training video generation models.

@leoma_ai miners, this dataset is ready for you to use.

At a simple $1 per minute benchmark, this dataset represents over $7,000 worth of generated training data.

Now available on the Nexisgen catalogue.

#AI #Video @huggingface $TAO #Bittensor

https://t.co/OhYLztDbpD

[Green Compute @green_compute_]

We just hit a big milestone.

In only our first month in Bittensor, we have now generated over £82,000 in profitable revenue which means the first buyback is now set in motion.

10% of month-one revenue = $8,200 of token buybacks, to be locked in the treasury.

So excited to be the FIRST subnet to be able to do this in the first month of operating.

We are only just getting started...

[DSV @dsvfund]

Almost nobody understands Bittensor.

That’s exactly why we’re hosting “INSIDE BITTENSOR” LIVE IN LONDON on 1st July.

If you’re interested in:

• AI
• Crypto
• Venture investing
• And what we believe could be one of the most asymmetric opportunities of the next decade

- then join us.

No technical background required.

Just a curiosity about where the next wave of value creation is emerging.

📍 London
📅 1 July 2026

Buy your ticket here:
https://t.co/m6OxCFlmmH

[NexisGen AI | SN70 @nexisgen_ai]

NexisGen just published one of the first Bittensor-native, open, training-ready video generation datasets on Hugging Face.

50,000 standardized clips.
- Not raw footage.
- Not source URLs.
- Not timestamps.
- Not a dataset that needs days of preprocessing.

Every clip is already prepared for training:
- 1280 × 704
- 24 FPS
- 121 frames
- Pre-clipped
- Consistent format

This is what open video generation has been missing: clean, model-trainable video data at scale.

Links:
- https://t.co/UBonjbPC0y
- https://t.co/OhYLztDbpD

$TAO

[Subnet Summer @SubnetSummerTAO]

SN99 (@leoma_ai) continues to make significant improvements to its AI video generation model.

A video recently generated by their miners has already caught the attention of @const_reborn .

Video prompt:

The camera performs a rapid, low-altitude forward dolly movement, gliding over a rugged, rocky mountain slope. The perspective shifts continuously as the camera descends and tracks forward, maintaining a consistent downward angle toward the uneven terrain. The motion is fluid and steady, simulating a drone flight path that traverses the desolate, boulder-strewn landscape. The scene features a sparse, mountainous environment with patches of snow and low-lying vegetation, captured in a continuous, uninterrupted shot throughout the entire duration.

[Mark Creaser @MarkCreaser]

Revenue Search 69: SN103 Djinn. Discovering asymmetric upside in AI start ups. https://t.co/xOb37Z9VVX

[Vex @vex020900]

Nowadays, most AI companies already have the model architecture, training pipelines, and GPU clusters ready.

The real bottleneck is high-quality training data. That's why Bittensor subnet 70 - @nexisgen_ai matters.

We've already collected 22+ hours of video training data, starting with video generation datasets and this is only the beginning.

Video generation companies like @grok , #Seedance, @runwayml , @pika_labs , @LumaLabsAI , @Hailuo_AI , @Kling_ai , @GoogleDeepMind , @HaiperGenAI , @ViggleAI , @HeyGen , and @OpenAI should take a look.

If you need high-quality video generation training data for your models, reach out me.

Dataset preview:
- 1280×704 resolution
- 24 FPS
- 121 frames per clip
- structured clip IDs, producers, and intervals
- https://t.co/8EvPXIMNOS

I’m ready to share this dataset with you for free, @elonmusk , for @grok.

[Mark Creaser @MarkCreaser]

Proud of the team for making the right call here. Holding a slot for the sake of it helps no one. Siam and I will park it and take our time finding the right project to put in its place. If you're building, get in touch.

[Handshake58 @handshake_58]

After careful consideration, we have decided to take Handshake off subnet slot 58.

We fully believe in Bittensor and the infrastructure we have built.

However, we do not believe we are best positioned to maximize the unique incentive mechanisms of mining and validation that makes Bittensor and subnet slots so valuable.

Rather than continue to occupy the slot, we believe now is the right time to step aside and allow a new team to unlock the full potential of the subnet.

Thank you to the investors, miners, validators, and contributors who supported us along the way 🤝

[Mark Creaser @MarkCreaser]

Revenue Search 68: SN78 Vocence. Discovering asymmetric upside in AI start ups. https://t.co/DoiPAKgOxq

[Mark Jeffrey @markjeffrey]

Checking out @vocence_bt Subnet 78 ...

I've spent a LOT of time using Suno to generate AI performance of human-written music -- so trying their 'open Suno' music-gen engine --

But also their voice cloner ... which, apparently, I'm on :)

[Mark Creaser @MarkCreaser]

Today, Compelle took over Bittensor Subnet 82 and launched to mainnet on the same day. That is unusual, perhaps unique.

In Bittensor, acquiring a subnet slot and shipping a working arena into it usually takes weeks, sometimes months. The Compelle team did it in a day. Validator live, miners competing, judge prompt published, transcripts open, Elo ledger public. We've not seen anyone else do this.

That kind of speed only happens when a team has thought through a problem for long enough that execution stops being the hard part.

Their core contention is that persuasion is the key to AGI. It is an orthogonal angle on system evolution, and once we understood it, we wanted to support them.

The signal @compelleai is chasing is whether one mind can move another, on the record, against its prior. If that turns out to matter for general intelligence, and we believe it does, SN82 is going to be very significant indeed.

@siamkidd and I have known the founders for some time and hold them in the highest regard. The depth they bring to questions of mind, motive, and language is why we are confident in what Compelle will become.

The team have asked to remain anonymous and we will of course respect that.

We back them without reservation and the work speaks for itself - all on chain and in the open.

[compelle @compelleai]

Compelle is live on Bittensor SN82.
It is fitting that on Victory Day - when the world remembers what was preserved by those who refused to yield - we announce the full mainnet launch of Compelle: a platform where adversarial AI paves the way to AGI.

[Mark Creaser @MarkCreaser]

Revenue Search 67: SN48 Quantum Compute. Discovering asymmetric upside in AI start ups. https://t.co/6Jw5IjGBHO

[MANDO CT 🇮🇪 🇦🇪 🇬🇧 @XMaximist]

$TAO is not just another alt. It’s positioning itself as the backbone of decentralized AI.

While most of the market is still chasing narratives…
TAO is quietly building real infrastructure + real demand.

AI + crypto isn’t a trend.
It’s the next phase.

And TAO is sitting right at the centre of it.

Supply is tight.
Network growth is accelerating.
Attention is just starting.

This is how big moves begin.

Don’t be surprised when TAO becomes one of the strongest performers this cycle.

[Mark Creaser @MarkCreaser]

Team not doxxed to me, which would have been a red flag and no-go for a long time on my investment criteria, but so many of the projects that hit my desk are now anon.

I can understand it. Often think that if I was entering this world today maybe I would do exactly the same.

So I invest. I just maintain a much higher degree of control when the other party might have less to lose.

[Mark Creaser @MarkCreaser]

I’m saying that I own those slots, if that helps.

Although own is a strong word, and it’ll be clear to anyone who’s ever watched me try to enter a postcode into a satnav that I’m not going to bring much technical value.

I’m an investor. And because the team isn’t doxxed, I control the cold keys for those slots.

Can’t be too careful with undoxxed devs, as my Nan used to say.

[AdTAO @adtao_ppcrebel]

Most subnets launch with a roadmap. We launched with $40k/month in recurring revenue.

We run three revenue models. Managed services where we operate accounts directly. Agency gain-share where we replace the back office and split the margin. Direct licensing for businesses that want the tooling.

Every one feeds the same dataset. 60-location franchise we signed a few weeks ago means 60 more accounts contributing cross-account signal. Agency partnerships mean structural diagnosis across industries we'd never see otherwise. Revenue from real contracts means we can fund miner incentives without depending on emissions alone.

The business has been running for years. The subnet turns that operational history into distributed prediction. Miners compete on the same data moat we've been building since long before Bittensor existed.

This is what a revenue-backed subnet looks like. Not revenue for its own sake - revenue because every paying account makes the next prediction more accurate.

If you're evaluating SN21, what numbers matter most to your staking decision?

$TAO #sn21 #bittensor #GoogleAds

[Mark Creaser @MarkCreaser]

Revenue Search 66: SN55 NIOME. Discovering asymmetric upside in AI start ups. https://t.co/eBCDk6ozQX

[Vex @vex020900]

The Agent for Rendix Subnets is now live.

Huge thanks to @const_reborn . His vision of using agents to help build subnet code, manage subnet channels, and handle subnet operations is genuinely innovative. This is exactly the kind of direction subnet owners need to study closely.

Our Rendix team dug deeper into the idea with @arbos_born , explored how it could support, and delivered it for Rendix Subnets.

Rendix is live now.

Go try it.

https://t.co/AgbmCgOPPQ

[Mark Creaser @MarkCreaser]

Leoma is a very ambitious project from a team with real Bittensor experience. That’s why I’ve invested in it personally.

Building a 25B+ video model takes capital, GPUs, data, evaluation… and most importantly, a team that knows what it's doing.

It’s easy to talk a good game. Much harder to turn a fancy slide deck into something that actually works.

Every time I speak to the Rendix team, I’m impressed by how clearly they understand what they’re taking on with Leoma and @LLuciano_BTC saw the same.

I think they have a real shot at making it happen.

[Lucky @LLuciano_BTC]

While exploring the bittensor subnets, one idea kept getting clearer to me: @leoma_ai - Subnet 99 is not trying to be “just another bittensor subnet.” It is aiming to become the video generation company that proves bittensor can compete with the biggest closed AI labs. Thanks to @MarkCreaser, CEO of DSV and investor in @leoma_ai - Subnet 99, and @vex020900, CTO of Vex|Rendix, for taking me through the vision.

Also, huge respect to the rendix team for building the leoma subnet. After hearing the direction directly, it became very clear to me why this matters. One thing i want to highlight is vex.

The CTO of leoma is a giga brain in AI. From what i have seen, his technical knowledge is on another level, and i believe his skill, vision, and execution will be a major reason leoma can actually build what it is aiming for.

Today, @leoma_ai - Subnet 99 miners are training and fine-tuning wan 2.2, one of the strongest open-source video generation models in the world, but that is only the starting point. During the call, the team shared that leoma is preparing an initial free-version launch on April 30.

This is not the final vision. It is the first public product step: showing the community that leoma has built the infrastructure to sync with the top miner’s model and turn subnet competition into a real user-facing product.

That's what matters the most. Because a subnet should not only produce emissions. It should produce a product. The real goal is much bigger: a 25B+ leoma-native video generation base model, which will be built by the rendix team, released as open source, and then improved through bittensor’s miner competition. This is where things become serious.

Building a 25B+ video model will cost millions of dollars. GPUs, training infrastructure, datasets, storage, inference optimization, evaluation, and scaling are not small expenses. This is the level of infrastructure normally reserved for google, openai, xai, runway, and other giant AI labs. That is why I believe funding leoma’s infrastructure matters.

⇢ Without serious infrastructure, there is no serious model.
⇢ Without serious model training, there is no real challenge against closed-company SOTA.
⇢ Without subnets willing to take that level of challenge, bittensor will never prove its full potential.

So yes, I will help support leoma’s infrastructure push, this is not just a support. But to help give leoma the resources it needs to build the 25B+ base model, support miner fine-tuning, scale evaluation, and move toward a real video generation product.

Remember, no bittensor subnet has ever pushed past the largest open-source models in its field and gone directly after closed-company SOTA, leoma is aiming to do exactly that.

⇢ Not by writing another roadmap.
⇢ Not by pretending decentralization alone is enough.
⇢ But by combining capital, infrastructure, miners, validators, and open-source development around one objective:

THE BEST MODEL WINS.

If leoma can deliver video quality competitive with closed AI companies while offering generation at 3x–5x lower market cost, that is not just a win for leoma. That is a proof-of-concept moment for all of $TAO.

Imagine a bittensor subnet producing video quality that competes with the biggest closed labs, while being cheaper, open, and continuously improved by miners. That is the kind of subnet bittensor needs. That is the kind of subnet that changes the narrative.

⇢ I am not promising @leoma_ai - Subnet 99, becomes the top subnet overnight.
⇢ I am not promising they beat every closed AI lab tomorrow.
⇢ But I believe this is one of the most important attempts happening on bittensor right now.

TIME WILL TELL.

But if leoma pulls this off, it will not just be a game changer for leoma. It will be a game changer for bittensor.

To all $TAO holders: I strongly suggest you take a serious look at leoma. Read their papers, study the vision, and understand why this subnet could matter for the future of bittensor:

https://t.co/dJGOfpB4t1
https://t.co/v22b1kEUmH

[siamkidd @SiamKidd]

https://t.co/VLEWihIEfL