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:

Full report available here: khala.io/walrus-solving-the-…

It’s now obvious that memory is one of the biggest bottlenecks in AI But the next question is whether that memory can be trusted “Context is the new bottleneck” - Jensen Huang at CES 2026 He’s right, but only half right Every AI agent deployed today still has the memory of a goldfish because context windows are stateless by design, starting each session from zero That is unfortunately how LLMs work But storage capacity is only half the battle, the other half is trust Current memory providers can write memories to docs, databases or internal logs, then ask enterprises to trust that record later The problem is obvious: one altered memory corrupts every decision that relied on it As agents move into regulated workflows, this becomes a real issue The EU AI Act requires high-risk AI systems to maintain automatic event logging and traceability, meaning enterprises will need proof of what data an agent accessed, what it remembered and whether that record was tampered with Our latest report covers AI memory and how Walrus tackles this problem with portable, verifiable and programmable memory for agents The full report is in the next post below

Pretty good high level of what we're building. > an agent is only as good as the data it can reach and the API it can call > acquisition is the most direct bet on the agent thesis > as trading goes agentic, whoever holds the cleanest, most complete, most programmatically accessible record of onchain assets becomes the layer agents are forced to route through > blockworks will become the single system of record for all onchain assets > combined biz = issuers publish in, consumers query out, and the platform sits in the middle as the record both depend on

Full report: khala.io/bittensor-an-invest…

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

Full report available here: khala.io/xmaquina-onchain-ma…

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

Full report: khala.io/surf-data-platform-…

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:

Full report available at: khala.io/x402-completing-the…

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:

Agentic Commerce is hitting escape velocity x402 provides the permissionless rails for agents to embrace self sovereignty This standard has been missing for 30 years: HTTP 402 "payment required" now has a solution, and it's supported by a flourishing ecosystem Our full x402 report will be published soon; drop a follow & turn notifications on

Full report available at: khala.io/openclaw-ecosystem-…

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

Link to the full report below: khala.io/bittensor-the-intel…

The signal-to-noise ratio is deteriorating We intend to flip the switch and provide you with high quality research Our next report on Bittensor will be dropping soon Follow @KhalaResearch and turn notifications on to remain ahead of the curve

The Bittensor (TAO) Ecosystem Map: A map of 100 (live) @opentensor subnets; Use this as your initial compass to navigate the TAO ecosystem Turn notifications on; our full report will be published in the next few weeks Let us know in the comments which segment you find most interesting:

Read the full report here: khala.io/decentralized-robot…

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: