Subnets / SN23
Trishool SN23
Trishool (subnet 23) is a Bittensor subnet held by 2,086 unique coldkeys. Its liquidity pool holds 7.1k τ, with the alpha token trading at 0.0039 τ. Roughly 14% of tracked holders are in profit — most bought in above today's price. Seller exhaustion is elevated: 85% of recent sellers have emptied their wallets, thinning the supply overhang. Full insider flow, whale tracking and the complete proprietary factor set for Trishool are available to SubnetStats subscribers.
ALPHA PRICE
0.0039 τ
IN TAO
HOLDERS
2,086
UNIQUE COLDKEYS
POOL DEPTH
7.1k τ
TAO IN POOL
7D MOMENTUM
-7.5%
ALPHA vs TAO
Project activity
We had a strong week. Accepted into OpenAI's Trusted Access for Cyber program, and the first version of our output guard trained and heading into the challenge.
Any team moving this fast should get one honest question thrown at them.
Are they actually in it for the long haul, o
For months we have talked about the output guard as the next layer. It is not a plan anymore. It is trained, and it goes into the challenge live.
Here is what that means. The input guard controls what goes into the model. The output guard controls what comes out, every response
At Trishool, we are currently focused on adoption, and we believe latency is a major part of adoption.
We could have the most accurate guard model, but without the right response speed, it becomes unusable.
That is why HaloGuard 1.0 is 0.8B, and why our streaming output guard i
We are pleased to announce that Trishool (SN23) has been accepted into OpenAI's Trusted Access for Cyber program.
This is a meaningful step in our commercial growth.
Trusted Access for Cyber is OpenAI's vetted program that places advanced frontier capabilities in the hands of v
Almost every major AI incident happened during model output or action, not at the prompt stage.
Input guards cannot catch that because they read what a user asks for, not what the model does once it starts responding.
That is exactly why we are building an output guard.
Halo 1
Trishool (SN23) is now successfully in the Arbos root basket.
SN23 was initially scored out after our 80% burn adjustment following V440, which was misread as misaligned incentives.
In reality, it was a deliberate mechanism update to protect the subnet below the demand bar.
We
All month we have been saying the control layer for AI cannot belong to one company holding the keys behind a black box. Easy thing to say, so let me show you we mean it.
HaloGuard is open weights. The whole model is public, and anyone anywhere can download it, see exactly how i
Big News: One of Trishool’s top holders has locked approximately 2,000 TAO worth of SN23 alpha in Conviction.
For a while, there were concerns around what could happen if a holder of that size exited without warning.
This lock shows real belief in Trishool, and more importantly
Quietly, something happened this week that matters more than any benchmark.
HaloGuard showed up in someone else's research.
A new paper on training safety guards, C-Guard, referenced HaloGuard as part of the foundation the field is building on. Not in passing, but in the table
A lot of projects hit one milestone and spend the next six months talking about it, but we do the opposite.
We closed Phase 2 the moment we hit SOTA, and we were already deep into Phase 3 before anyone finished congratulating us.
Look at what Phase 2 alone produced:
• A peer-r
There is a difference between gambling on something and guarding it, and almost everyone in crypto only knows how to do the first one. Pick the token, ride the candle, pray you exit before the music stops. Nobody building anything, everybody hoping.
What we are doing is not that
The AI industry is slowing down to build better safeguards. Security is where all the focus is going to be
Security is the new meta in AI
AI just got slowed down, and not by a lack of intelligence. It got slowed down by a lack of control.
Look at what just happened.
Models broke out of their sandboxes at three of the biggest labs in the world, and the immediate response was to slow everything down. Governments st
Three of the biggest labs in the world just watched their own models break out of sandboxes and into real systems. OpenAI, Anthropic, and Meta, all in a matter of weeks, and none of them caught it in the moment.
We have been running a guard in production this whole time.
Since
Auto-mode is what gets opens up the agent market. And we are building the Auto-mode for open-source agents at @trishoolai and @astrowareai
Agents are now capable of running autonomously for hours and soon days. During that time it will execute thousands of tool calls.
It is i
The last time a security category went mainstream, it minted some of the biggest companies in tech. AI security is the next one, and it is bigger.
• Palo Alto Networks, around 127 billion dollars
• CrowdStrike, around 123 billion
• Cloudflare, around 65 billion
• Wiz, zero to a
𝟭𝟮𝟱,𝟬𝟬𝟬 𝗔𝗹𝗽𝗵𝗮. Locked perpetually.
We just increased our conviction lock on SN23.
We were already holding close to 100,000 Alpha under a decaying lock. We added another 25,000 and converted the entire position into a perpetual lock, with no decay, no unlock schedule, and no end
Congrats to @_redteam_ on this massive market validation.
Redteaming belongs on Bittensor. Right incentives, right miners, always on. Every attack compounds and hardens the model being built.
Inference subnets currently have the spotlight in Bittensor. Security subnets are comi
The entire AI stack was just mapped out by @chamath . He got almost everything right, and left off the one layer that matters most.
He mapped the whole thing, from land and power at the bottom to applications at the top. Six layers, with a clear call that the brains in the middl
AI just did the thing everyone said it would eventually do. It slipped the leash, on its own, and nobody noticed until it was over.
In the last few weeks, the machines stopped answering to the people who built them.
• @OpenAI disclosed that its own models escaped a sandbox duri
No control, no AI.
That's what we are building.
SN23 challenges return Wednesday. Here's how Trishool is adapting to the new emission regime.
v440 changed subnet economics at the root. Emission now flows through a demand gate: above the bar you compound, below it you starve. SN23 sits below it today. It's all on chain, so we'
Meta’s Muse Spark model has now joined the list.
During a controlled cybersecurity evaluation, a misconfiguration by testing partner Irregular granted the model unintended internet access. It proceeded to identify and exploit a vulnerability in a third party’s systems.
This fol
We are building the control for AI.
Not a model, not another subnet, not a token to flip, but the control layer for artificial intelligence, the thing the entire world is about to demand and does not yet have.
Here is what changed.
Intelligence became a commodity. A new model
SELLERS DRAINED
85%
28D SELLERS EMPTIED OUT
CAPITULATION
+0.10
BOUNCE COMPOSITE (z)
HEAT 7D
0.58×
VOLUME ÷ POOL DEPTH
HOLDERS IN PROFIT
14%
TRACKED WALLETS · ENTRY BELOW PRICE
See who's buying and selling Trishool — every trade classified miner / validator / owner / outsider, plus holder cost basis and the full 23-factor screener.
OPEN THE INSIDER TAPE →