Nuance

Been watching the numbers closely, and I believe the halving is about to expose this inside Bittensor’s $TAO’s economics. On paper, it’s simple: emission gets cut by 50%. But in reality, that one change doubles the weight of every mistake happening in the system right now. If a subnet is mispriced today, the halving makes that mistake twice as expensive. If a subnet is basically dead low volume, low activity, no real users the halving makes that dead weight twice as heavy. If emission is flowing to the wrong places, that misalignment becomes twice as visible. And every validator decision suddenly carries twice the consequence. People talk about halving like it’s some kind of abstract monetary event. It’s not just that. It directly exposes the economic structure underneath the network. And right now that structure has cracks. Think about new subnets for a second… they’re already fighting uphill battle. They get less emission to start with, they have to compete against incumbents with massive legacy weight, and they’re doing this while network-wide volume is collapsing. After the halving, they have even less emission to bootstrap, more competition for stake, and fewer trades to signal their true value. It’s a tough environment. This is why I keep saying the next phase is going to force the network to evolve. You can’t hide behind high emission once supply is cut. You can’t coast as a zombie subnet once every $TAO becomes more valuable. You can’t ignore mispricing once it stands out. The halving is a filter. It rewards real innovation. It rewards high-quality new subnets. It rewards the alpha markets that actually reflect demand. And it rewards any newcomer willing to prove they deserve a spot. Everything else? It’s going to get squeezed. This isn’t negative. This is needed. It’s just the moment where the network has to decide if it wants to become sharper, leaner, and more competitive, or if it wants to drag dead weight deeper into the next era. Either way, the halving won’t let anyone hide. That’s why I think the next few months will be the most revealing we’ve ever seen. This is where the winners will really shine.

🚨 $TAO Emissions Got an Upgrade 🚨 Bittensor just changed how new TAO is distributed. No longer based on Alpha EMA prices, now driven by net TAO inflows. 🧮 Old model → Price-based Subnets with higher Alpha prices got more TAO. This caused issues: • Bias toward big liquidity pools • Subnets could still earn even as capital left • TAO/Alpha injections reinforced price loops 🌊 New model → Flow-based TAO emissions now follow real network inflows. Each subnet tracks its Net TAO Inflow EMA, the average of TAO staked minus unstaked. • EMA smooths over ~87 days (≈ 86.8d half-life) • Positive inflows → earn emissions • Negative inflows → zero emissions until positive again ➡ Anchors emissions to genuine demand, not speculation. 💡 Why it matters • Rewards sustained inflow, not short-term price pumps • Stabilises emissions over time • Makes growth measurable on-chain Long-term builders win. Liquidity farmers don’t. 💎 Root Claim Upgrade Root stakers can now choose how to claim rewards: • 🔄 Swap – auto-convert Alpha → TAO (default) • 💎 Keep – hold Alpha to back the subnets you earn from “Keep” reduces Alpha sell pressure & supports subnet ecosystems. (Partial-keep coming soon.)

I get why people are skeptical. Most crypto projects chase hype. But Bittensor $TAO is different it’s aiming at a problem that’s about to define the decade. We’re heading toward an AI monopoly that makes Big Tech look tame. Right now: Only 3–4 companies can train frontier models: OpenAI, Google DeepMind, Anthropic, Meta. Training costs are exploding. GPT-4: $100M. GPT-5: $1B. By 2030, $10B+ runs. That means only 2–3 entities. Will control all meaningful AI. And that’s where the real danger starts: 1. Corporate Alignment Problem Google decides what’s “true.” No competition. No alternative viewpoints. 2. Innovation Gatekeeping You want AI for research? Pay their price, follow their rules, accept their limits. 3. Data Slavery We train their models for free, they own the value. 4. Single Points of Failure One AWS outage. One regulator. And half the world’s AI goes dark. Bittensor exists as infrastructure insurance. It doesn’t need to be perfectly decentralized just decentralized enough to: • Prevent single-entity control • Enable permissionless innovation • Distribute value fairly Think Linux vs Windows. Linux didn’t win the market, but it forced competition, powered critical infrastructure, and gave the world a real alternative. Bittensor is the Linux of AI. Real use cases are already emerging: – Sovereign AI for nations too small to build their own. – Specialized models Big Tech ignores. – Uncensorable research. – Economic participation for contributors GPUs, data, validation work paid in $TAO. Without Bittensor by 2030: – 2–3 companies own intelligence itself – Prices rise – Innovation is permissioned – Nations get locked out – “AI safety” becomes corporate policy With Bittensor: – Constant competition – Open innovation – Shared infrastructure – Value flows to contributors – Multiple safety approaches $TAO it’s the coordination mechanism. No incentives, no network. No network, back to monopolies. Bittensor might fail. It’s hard tech. But the alternative total corporate control of intelligence is worse. The question isn’t “Will Bittensor beat Google?” It’s “Do we really want to live in a world where Google has no competition?” We’ve seen this movie before. Oil. Banks. Big Tech. This time, it’s intelligence itself. And that’s why I think $TAO matters.

2/ Subnet registration and deregistration have returned to Bittensor. The subnet limit will remain at 128 meaning that when a new subnet registers, it will take the place of the lowest performing non-immune subnet. A 4-month immunity for new subnets will allow the time for subnets to get their projects up and running.