Browse all takes

Bull thesis on Bittensor / TAO at ~$3B mcap. Frame: "TAO 2026 = ETH 2016 = BTC 2013."

Core mechanic: Bitcoin paid miners to produce hashes that secure the network but are otherwise worthless. Bittensor pays miners — data scientists, ML engineers, AI researchers — to produce useful AI work. Validators score outputs via Yuma Consensus; TAO flows to whoever produces the most valuable work. Network is organized into 128+ subnets, each focused on a specific task (trading signals, LLM training, computer vision, code generation, financial forecasting). Some subnets generating millions in revenue, with Intel and PwC partnerships.

Tokenomics mirror Bitcoin: 21M fixed supply, no pre-mine, no VC allocation. First halving Dec 14 2025. BTC price went 83x in the year after its first halving in 2012.

Smart-money signals: Barry Silbert / DCG launched Yuma Group dedicated to accelerating Bittensor. Grayscale filed Form S-1 to convert GTAO Trust into a spot ETF. Stillcore Capital (Mark Jeffrey, Jason Calacanis, Rob Greer) targeting $1T mcap by 2030, aiming to own 1% of all TAO. Unsupervised Capital projects $4,800 by Dec 2027 (19x), bull case $10,800 — and that's before Covenant-72B, Jensen mentioning Bittensor, and PwC's formal alliance.

Subnet-level conviction picks:

• Targon (SN4) — decentralized AWS for AI; Targon VM gives encryption + hardware-backed protection so hardware operators can't access data, weights, or workloads. Co-authored a paper with Intel in March 2026. Built by ex-OpenTensor founders (Robert Myers — among first 3 people ever in the Bittensor Discord; James Woodman ex-GSR).
• Vanta (SN8) — disrupts the $20B prop firm industry. Single eval, 100% profit split, fully on-chain verification. Already net profitable on revenue vs miner emissions. Hyperscaled is the Hyperliquid version.
• Chutes (SN64) — #1 open-source provider on OpenRouter, 9.1T tokens processed. Decentralized AWS with no CEO.
• RESI (SN46) — institutional-grade real estate intelligence. 98% accuracy remote appraisals on a $600T asset class running on broken legacy MLS systems. 1000+ appraisals + nationwide lender partnership in week one. Strategic investment from Stillcore.
• Affine (SN120) — built by Const himself (Bittensor co-founder, wrote the Yuma Consensus + subnet architecture). Continuous evaluations on open-source reasoning models, leverages Chutes for hosting.
• Score (SN44) — first subnet ever to partner with a Big Four firm. Manako product distributed by PwC France to retail, manufacturing, logistics, energy enterprise clients.
• Oro (SN15) — autonomous AI shopping agents. 45 Oro agents have outperformed GPT-5.4 on hard online shopping evals.

Frame: Bitcoin = money. Ethereum = apps. TAO = intelligence. The gap between what TAO has built and how it's currently priced is one of the most asymmetric opportunities in crypto.

Arrakis follow-up to its earlier "Who's trading on HIP-3?" piece, this time using deterministic Hyperliquid order-metadata tags (TIF, builder code, fill flag, hold time) to mechanically classify every wallet across the four Trade.xyz markets (xyz:CL, SILVER, TSLA, XYZ100) over March 10–31, 2026: 79,622 wallets, $51.95B total volume.

Key finding: the sybil layer inflated wallet count, not dollar throughput. The "Airdrop Farmer" bucket holds 35,091 wallets (44% of users) but generated only $0.40B (0.77% of volume). 99.9% of those farmer wallets trace back to a single Polymarket operator ("Themino") running 70 chains of 34,553 wallets through a baton-pass farm — using HL's $1 internalTransfer primitive, each wallet runs a 5-step sequence in ~26 seconds. Total fees Themino paid: $34,510.

Real volume comes from identifiable books. Market makers: 363 wallets (0.46%) carried 63% of volume ($32.75B). The #2 MM ("Powell") is a Polymarket user running multi-market quoting. Jump Crypto ($3.15B), Selini Capital ($1.03B across 3 wallets — two MM, one HFT), Wintermute ($230M) all visible. Builders split into algorithmic (Tread.fi, Origami — replaced wash-trading with retail market-making, now populate top-of-book on nights/weekends when traditional MMs aren't quoting), wallet-integrated (Phantom, MetaMask, Rabby — $1–3K median per wallet), and apps (Insilico, hypurrdash, etc — fewer wallets, higher per-wallet volume). Retail: 22% of top-400 retail volume ($1.63B) is verifiable Polymarket users. Total Polymarket footprint across MM+SAT+retail on Trade.xyz: ~$6B. Kraken dominates CEX-funded retail; Hyperunit + deBridge dominate bridge-funded.

Conclusion: layered answer to the sybil debate. Yes there's a sybil layer (predictable pre-TGE). No evidence of separate high-volume wash-trading. Real volume runs through identifiable professional desks + a Polymarket-overlapping retail base.