AI stocks outperformed every major index in 2025 Model development consolidated behind a few closed labs controlling the most capable models and the most valuable training data A generation of crypto-native AI entrepreneurs chose an alternative path Bittensor launched in January 2021 and is now the most developed public market for open-source AI intelligence 128 subnets with an aggregate market cap grew from $4M at launch to a peak of $1.5B The market decides which subnets get fed with emissions flowing to subnets showing traction Our latest report maps the three investment eras of Bittensor: 1) Era 1 - Genesis (2021-2024) 2) Era 2 - dTAO (2024-2025) 3) Era 3 - Institutional Era (2025 ) We also share a few subnets we are keeping a close eye on along with index options for those who want a hands off approach Full report in the next post below

In the U.S., 2.1M manufacturing jobs will go unfilled by 2030, at a cost of $1T annually That’s the demand side. The supply side: 16,000 humanoid units deployed globally in 2025, up from 2,000 the year before Private robotics equity has already created tens of billions in value Embodied AI changes labour cost structure the way software changed information cost structure BMW has run Figure AI robots on daily 10-hour shifts for over 11 months across 30,000 vehicles and 1,250 runtime hours Leading companies carry combined valuations exceeding $85B, but most of that value will accrue to private investors No retail-accessible instrument offers specific exposure to named pre-IPO humanoid companies RoboStrategy is the exception that proves the rule: a listed fund trading 160–700% above NAV because scarcity drives premium XMAQUINA’s RCM Protocol converts verified private robotics equity into onchain subDAO tokens, bringing liquidity to illiquid markets Full Khala Research report in the next post below

Ask a frontier AI model to check if a smart contract is safe It will return a confident, structured response, which is usually wrong Even if it's a known phishing contract, the model reaches for web search instead of a block explorer, using the wrong tool for the task The CAIA benchmark shows the best general frontier model can only reach 67% on crypto tasks vs 80% human analyst because models choose web search most often, even when specialist tools are available AI models need better routing for crypto data Surf is built to solve that failure Full Khala Research report below:

In 1997, HTTP status code 402 was reserved for 'Payment Required'; a native payment layer built into the web itself For almost 30 years, nobody could make it work There was no digital cash and no settlement layer that could handle sub-cent transactions. The web defaulted to advertising and credit card checkout flows instead Stablecoins and Layer 2s changed that. x402 finishes what HTTP 402 started; payment as native to the web as loading an image Most people looking at agentic commerce see one maturity cycle, but there are two: 1) The retail narrative is approaching its peak; CT is buzzing, ecosystem maps are proliferating, every protocol is announcing x402 compatibility. That will correct 2) The institutional adoption curve is on a completely different trajectory. For this layer, the trough already happened in 2025 Stripe, Visa, AWS, Google, Coinbase, Circle and Cloudflare are shipping production infrastructure because their own internal analysis on projected agent transaction volume justify it In our x402 report, we map the protocol architecture, the agentic stack, the institutional landscape, and assess where the value could accrue The full report is in the next post below:

AI Agents are becoming perpetual revenue generating machines The ability to ship code more effectively than humans in a fraction of the timeframe provides an enormous opportunity to create an eternal revenue flywheel This is the "Dawn of the Autonomous Software Factory" Agents are already generating six figures in value and shipping 100s of viable apps per week Billions are being wiped from traditional SaaS company valuations as their future cashflows evaporate, being replaced with an uptick in Frontier AI model subscriptions In this report, we cover five OpenClaw projects that are experimenting with onchain value accrual mechanisms The full report is in the next post below

Training a frontier AI model cost $5M in 2020. Today it costs $1B The gap keeps compounding, in favour of the same five companies A decade ago, nobody believed open-source could compete with enterprise software. But then it did Bittensor (TAO) is a global open market where subnets compete to produce AI services, getting paid based on quality output This is what crypto is good at - coordinating users with shared incentives Over the past 12 months, significant progress has been made in the ecosystem; recurring fiat revenue, enterprise contracts, decentralized research and more In our TAO report, we: 1) Profile 5 subnets generating revenue across inference, computer vision, compliance, and drug discovery 2) Show how token incentive mechanics create advantages centralized companies cannot replicate 3) Breakdown catalysts and risk factors of subnets and Bittensor The link to the full report is in the next post below:

70% of global GDP is still tied to physical labor and location-dependent industries AI capabilities are capped until it takes physical form; This is the promise of Embodied AI Once we cross the "utility tipping point" for general-purpose robotics, the fundamental cost of labor drops toward zero We aren't just looking at a workforce shift; we’re looking at a transition to a post-scarcity economy where a high standard of living is decoupled from traditional labor This sounds like science fiction, until you realize the infrastructure required to scale it Crypto becomes this missing link Decentralized networks solve the massive capital formation and coordination challenges required to deploy millions of autonomous units globally In our inaugural @KhalaResearch report, we: 1) Identify the pioneers leading the charge in Decentralized Robotics 2) Explain the synergy between blockchain incentives 3) Share exclusive insights from subject matter experts on the future of this vertical The link to the full DeRobotics report is on the next post: