Technology investments, research, and insights by Hypersphere and contributors.

Writing

Hypersphere Ventures | Writing

Brian Breslow

Mehdi Farooq

Micah Casella

Jack Platts

Commercialized robotics is not a new concept. Most are familiar with products such as the iRobot Roomba vacuum, debuting in 2002, or more recent household robotics such as Kasa’s pet camera. Both are single-purpose built. With the help of AI, robots are evolving from single-purpose machines to multi-purpose, designed to operate in unstructured environments. 

Humanoid robots will progress from basic tasks like cleaning and cooking to concierge, firefighting, and even surgery over the next 5–15 years. 

Recent developments are turning humanoid robotics from science fiction to reality.

Market expansion: 100+ companies building humanoids (e.g. Tesla, Unitree, Figure, Clone, Agile, etc.).

Hardware tech has crossed the uncanny valley: New humanoids exhibit fluid, natural motion, enabling human-like interactions in real-world environments. Unitree’s H1 walks faster than the average human (3.3 m/s vs 1.4 m/s).

New paradigm of labor costs: Humanoid robots are expected to undercut U.S. wages by 2032.

Despite clear tailwinds for humanoid robotics, mass deployment remains bottlenecked by data quality and scarcity. 

Other AI embodiments, like autonomous driving, have largely overcome the data issue via cameras and sensors on existing vehicles. In the case of self-driving (e.g., Tesla, Waymo), these fleets are able to generate billions of miles of real‑world driving data. Waymo was able to put their cars on the road for real-time training with a human “babysitter” in the passenger seat during this phase. 

However, consumers are unlikely to tolerate the presence of a “robot babysitter”. Robots must be performant out-of-the-box, making pre-deployment data acquisition essential. Training must be completed prior to commercial production, where scale and quality of data persist as an issue.

While each training modality has its own unit of scale (i.e. tokens for LLMs, video-text pairs for image generators, and motion episodes for robotics), the below comparison highlights the orders-of-magnitude gap in data availability that robotics data is contending:

GPT-4 was trained on over 15 trillion text tokens

Midjourney/Sora leverage billions of labeled video-text pairs

In contrast, the largest robotics datasets contain only ~2.4 million episodes

This disparity illustrates why robotics has yet to achieve a true foundation model in the same way LLMs have. The data simply isn’t there yet.

Traditional data collection methods do not scale for humanoid robotics training data. Current methods include:

Simulation: Cheap but lacks real-world edge cases (Sim2Real gap)

Internet Video: Doesn’t contain proprioceptive or force feedback context needed for robot learning

Real-world data: Accurate, but requires teleoperation and a human in the loop, which is costly ($40K+ per robot) and lacks scalability.

Training in virtual environments is inexpensive and scalable, but models often struggle when deployed in the real world. This problem is known as the 

For example, a robot trained in simulation might succeed at picking up objects with perfect lighting and flat surfaces, but fail when faced with cluttered environments, uneven textures, or imperfect situations humans are accustomed to in the physical world.

Reborn provides a way to cheaply and quickly crowdsource 

real-world data, enabling robust robotics trainingand solving the Sim2Real gap

Reborn: A Full-Stack Vision for Decentralized Physical AI

Reborn is building a vertically integrated software and data platform for physical AI. At its core, Reborn is solving the data bottleneck for humanoid robotics, but its ambition extends far beyond that. Through a combination of proprietary hardware, multi-modal simulation infrastructure, and foundation model development, Reborn becomes a full-stack enabler of embodied intelligence.

, a proprietary consumer-grade motion capture device. This powers a rapidly growing AR/VR gaming ecosystem, where users generate high-fidelity motion data in exchange for network incentives. Reborn has sold over 5,000 ReboCap units and now supports 160,000 monthly active users (MAUs), with a clear path toward two million by year-end.

Impressively, this growth has been organic: users are drawn to the entertainment value of the games themselves, and livestreamers are adopting ReboCap to animate digital avatars with real-time body tracking. This organic engagement loop powers scalable, low-cost, and high-fidelity data generation, making Reborn’s dataset a valuable training resource for leading robotics companies.

The second layer of Reborn’s software stack is 

, a multi-modal data platform that unifies fragmented simulation environments. Today’s simulation landscape is highly fragmented, for example tools like Mujoco and NVIDIA Isaac Lab each offer different strengths but lack interoperability. This balkanization slows progress and exacerbates the Sim2Real gap. Roboverse addresses this by standardizing across simulators, creating a shared virtual infrastructure for developing and evaluating robotics models. This integration allows for consistent benchmarking, improving scalability and generalizability.

Together, ReboCap and Roboverse form the base of Reborn’s full-stack platform. The former captures real-world data at scale, while the latter orchestrates simulation environments for model training. This integrated approach showcases the true power of Reborn’s DePAI network. It is building a developer platform for Physical AI that extends beyond simple data acquisition, to actual model deployment and licensing.

Perhaps the most crucial component of Reborn’s software stack is the Reborn foundation model (RFM). Reborn is building one of the first robotics foundation models, designed to serve as core infrastructure for the emerging Physical AI stack. Think traditional foundation models for LLMs, such as OpenAI’s o4 or Meta’s Llama, but for robots.

The combination of the three major elements of Reborn’s stack (ReboCap, Roboverse, and the RFM), creates a strong vertically integrated moat for Reborn. By pairing crowdsourced motion data with robust simulation and model licensing, Reborn can train models with the scale and diversity needed to generalize across use cases. The result is a foundation model that supports downstream applications in a wide-range of use-cases, including industrial, consumer, and research robotics.

Reborn is actively commercializing its technology, launching paid pilots with Galbot and Noematrix and establishing strategic partnerships with Unitree, Booster Robotics, Swiss Mile, and Agile Robots. China's humanoid robot market is experiencing rapid growth, accounting for around32.7% of the global market. Notably, Unitree holds over 60% of the global quadruped robot market and is among six Chinese humanoid robot manufacturers planning to produce over 1,000 units in 2025.

Crypto is enabling the full vertical stack for physical AI.

While all these projects sit in different parts of the physical AI stack, they all share something in common: 100% of them are DePAI projects! DePAI makes decentralized physical AI possible by ensuring open, composable, and permissionless scaling via token incentives across the stack.

The fact that Reborn hasn’t launched a token yet makes its organic growth even more impressive. Once token incentives go live, network participation is expected to accelerate as a part of the DePAI flywheel: Reborn issues incentives for acquiring its hardware (the ReboCap), robotics companies pay ReboCap owners for their contributions, encouraging more folks to purchase and use ReboCap. Reborn will also dynamically incentivize high-value edge case behaviors – ensuring even better coverage of the Sim2Real gap.

Robotics “ChatGPT” moment won’t come from robotics companies themselves because hardware is much tricker to deploy than software. Virality in robotics is inherently constrained by cost, hardware availability, and logistical complexities. These factors are absent in purely digital software like ChatGPT.The tipping point for humanoid robotics will come not when prototypes impress, but when costs fall enough for mass adoption — as with smartphones or PCs. When costs fall, hardware becomes tablestakes. The real competitive edge will lie in data and models. Specifically, the scale, quality, and diversity of motion intelligence used to train these machines.

The robotics platform shift is inevitable but, like all platforms, it needs data to scale. Reborn is a high-leverage bet that crypto can fill the most acute gap in the AI robotics stack. DePAI for robotics data is cost-efficient, scalable, and composable. In a world where robotics are the next frontier of AI, Reborn is the equivalent of turning everyday humans into the “miners” of motion data. As LLMs need text tokens, humanoid robots need motion episodes. Reborn is how we unlock one of the last remaining bottlenecks in turning humanoid robotics from sci-fi to reality. 

Author: 

General-purpose humanoid robots are rapidly transitioning from science fiction to commercial reality. Declining hardware costs, surging capital investment, and advances in locomotion and dexterity are converging to drive the next major platform shift in computing.|While compute and hardware are increasingly commoditized, providing low-cost tailwinds to robotics engineering, the sector is still constrained by a training data bottleneck.

Reborn is one of the few projects leveraging decentralized physical AI (DePAI) to crowdsource high-fidelity motion and synthetic data, and build robotic foundation models, making it uniquely positioned to catalyze humanoid deployment. The project is led by a deeply technical founding team with research and professor backgrounds at UC Berkeley, Cornell, Harvard, and Apple, combining academic excellence and real-world engineering execution.


How Crypto and DePAI Are Revolutionizing Humanoid Robotics Development

DePAI: Crypto's Role in Scaling Humanoid Robotics

The corporate appetite for Bitcoin continues to accelerate with many publicly listed companies positioning themselves as BTC treasury vehicles. MicroStrategy (NASDAQ: MSTR) led the development of this niche strategy, with new entrants like MetaPlanet and Semler Scientific acquiring modest BTC positions before the start of this year. Then, this April, Twenty One Capital announced that it will come to market with a nearly $4 billion dollar treasury, backed by a group of investors including Tether, SoftBank, and Hypersphere — this was actually our 

 investment to date. Aggressive BTC accumulation on the balance sheets of public equities looks like a solution to an outdated problem: getting simple Bitcoin exposure. After all, MicroStrategy has been doing this for years, and Bitcoin ETFs are now widely trusted. However, after interviewing over 150 other crypto fund managers, as part of our multi-strategy fund, we noticed that many were focused solely on generating BTC yield for Bitcoin-denominated investors. And on the venture side, we’ve spoken to over 50 companies in the “Bitcoin DeFi” space where projects are promising solutions to generate yield on Bitcoin users' deposits. Hence, the problem is no longer about access; people now want Bitcoin with yield. After all, what’s better than the best performing asset over the past decade? The same asset - but with yield. We see the opportunity for public companies focused on generating BTC yield, and are betting on ones with high growth potential and differentiated business models.


 
Strive Asset Management is pioneering a Bitcoin treasury model anchored in strategies that other treasury vehicles have yet to experiment with. Most companies mimic MicroStrategy’s debt-based financial engineering tactics with the core differentiator being they’re smaller (so can more easily earn BTC yield) or are focused on an altcoin like Solana. Strive’s unique approach, detailed in the following section, builds on sponsor Vivek Ramaswamy’s experience at Roivant and consists of three core pillars: private equity-like acquisition of companies trading below cash; distressed Bitcoin claims, including Mt Gox claims; and debt financing paired with structured hedging strategies. We believe Strive has the team, capabilities, and network to become a leading BTC treasury company. Proving our conviction, Hypersphere invested $10 million. Strive trades under the NASDAQ ticker: ASST.Strive is planning a reverse merger with Asset Entities (ASST), a technology firm specializing in social media marketing and content delivery across major platforms. Notably, reverse mergers are often quicker to market than alternatives such as SPAC mergers and IPOs. Upon execution, Strive will see price discovery faster than many of its counterparts. In addition to this sharp market entrance, Strive also implemented a clever fundraise mechanism for BTC holders, facilitating its BTC treasury raise via IRS Section 351. Leveraging this structure enables a tax-free exchange of Bitcoin for equity, attracting BTC-rich contributors without triggering immediate tax liabilities for them. Strive’s most interesting differentiator lies in the three-tiered approach of its BTC treasury strategy.

 Strive will target cash-rich public companies (especially biotechs) trading below net cash, taking an activist investor approach to unlock stranded capital (estimated at $30B across the sector) and convert it into BTC.

Leveraging Vivek Ramaswamy’s biotech relationships and playbook.

Acquiring biotech companies post-clinical failure (e.g., no viable pipeline).

Uses extra cash to purchase BTC for the treasury.

 Strive plans to purchase legal claims to Bitcoin from bankruptcies, such as Mt. Gox and FTX.

These claims represent legal rights to future BTC distributions.

They trade at a steep discount (e.g., ~20%) to spot BTC prices.

Strive acquires these claims now and receives BTC later when payouts occur.

 Strive plans to use institutional-grade fixed income and options strategies to responsibly leverage the balance sheet.

Led by CEO Matt Cole (ex-$70B portfolio manager at CalPERS).

Applies strategic borrowing to acquire BTC when risk/reward is favorable.

Applies tail risk hedging strategies, such as buying deep out-of-the-money puts when Black-Scholes pricing creates favorable asymmetry.

We are excited to be backing Strive’s leadership team alongside other notable backers. Strive is led by CEO Matt Cole, a former $70B portfolio manager at CalPERS with deep options and fixed-income trading expertise. Vivek Ramaswamy co-founded Strive and the company received early backing from folks like Peter Thiel, JD Vance (U.S. Vice President), Howard Lutnick (U.S. Secretary of Commerce), and Bill Ackman.

 
As one of the earliest crypto funds to this evolving niche sector, Hypersphere has evaluated a number of BTC treasury investments. We’ve invested in Twenty One Capital, Nakamoto Holdings, SharpLink’s ETH strategy, and now Strive (NASDAQ: ASST). While capital has recently become more available for this strategy, we believe it’s important for public companies engaging in crypto accumulation strategies to plan for the long-term by building out world-class management teams and bringing unique business models to the table. Strive excels at both: with its unique operating business, asset management reputation, savvy fundraising model, and top-tier leadership team, we expect Strive to continue to outperform its industry counterparts and become a category leader in the space.

Announcing Hypersphere’s investment into Strive Asset Management


Hypersphere Invests in Strive Asset Management 


Bitcoin has proven to be worthy of allocation to any portfolio, from retail to retirement. BlackRock, Fidelity, VanEck, and other financial heavy-weights issued Bitcoin ETFs to provide investors easy exposure to Bitcoin. While these BTC-backed investment vehicles do a great job of tracking Bitcoin price, investors will never earn additional yield on their bitcoin. A single share of a Bitcoin ETF will always track the same number of BTC and, after fees accumulate year after year, the Bitcoin you are entitled to per share may actually decrease.

The publicly listed software company, MicroStrategy (NASDAQ: MSTR), demonstrated another way to express a bullish view on Bitcoin. The company began purchasing Bitcoin before any Bitcoin ETF approvals, using profits from its legacy software business. Then, Michael Saylor, MicroStrategy’s founder, leveraged up his company (in a responsible way!) by issuing debt to purchase more Bitcoin — 2.5% of the total Bitcoin supply, and counting. MSTR can trade at a discount or premium to its net asset value (NAV) of BTC held in its treasury. Currently, MSTR enjoys more than a 75% premium to its BTC NAV. This premium is a reflection of the company’s ability to generate Bitcoin yield, or additional BTC per share, for its shareholders.

MicroStrategy transformed itself into a Bitcoin accumulation vehicle where, instead of trading around Bitcoin (or altcoins!) to “stack sats” of additional Bitcoin (this rarely works anyways), the company used its balance sheet to issue debt at favorable rates to acquire additional BTC per share. When executed effectively, shareholders keep their shares but enjoy their Bitcoin per share increasing over time.

People aren’t satisfied with just holding Bitcoin (because when are they ever?) — they simply want more Bitcoin. As an aside, we interviewed over 150 other crypto fund managers, as part of our multi-strategy fund, many of whom were focused solely on generating BTC yield for Bitcoin-denominated investors. It’s a whole cottage industry. The issue is Bitcoin-denominated funds have capacity constraints, so you can only invest a limited number of BTC, and as an investor you are taking counterparty risk. Separately, as part of a venture fund, we’ve spoken to over 50 companies in the “Bitcoin DeFi” space where projects are promising solutions to generate yield on Bitcoin users' deposits. Ergo, solving Bitcoin yield is a huge opportunity.



 
Enter Tether. Tether has one of the largest Bitcoin corporate balance sheets in the world and is a long-term believer in Bitcoin’s potential to appreciate in dollar terms. The best way for Tether to earn yield on their long-term Bitcoin holdings, and solve for capacity constraints and counterparty risk, is through their newly announced company: 21 Capital. We believe 21 Capital is the best business for accumulating BTC. And to back this belief, Hypersphere invested $40 million. The ticker, for now, is NASDAQ:CEP.


21 Capital (XXI) will be listed publicly via a SPAC merger with Cantor Equity Partners (CEP), a subsidiary of Cantor Fitzgerald. Cantor’s former CEO is Howard Lutnick, the current Secretary of Commerce. 21 Capital will capture narrative on all fronts, even with its ticker, XXI, which is the roman numeral for 21 (a reference to Bitcoin’s fixed 21 million supply). The company is a clean expression of a leveraged Bitcoin bet and has no debt, no unrelated businesses, no litigation, and great backers.

Tether and SoftBank seeded most of 21 Capital’s BTC treasury, contributing $2 billion and $1 billion, respectively, in Bitcoin at the price of $85,000 per BTC. 21 Capital’s initial 42,000 BTC treasury makes it the third-largest publicly listed Bitcoin treasury in the world, behind only MicroStrategy and Marathon (NASDAQ: MARA). Tether and SoftBank’s backing represent a strategic alignment between crypto’s top stablecoin issuer and globally-recognized institutional capital. While Tether’s alignment can help with operations and crypto-native ecosystem support, SoftBank’s offers a far-reaching network from LATAM to APAC and global regulatory expertise, having assisted Grab (Southeast Asia), Paytm (India), and Klarna (Sweden).

21 Capital will go further than other Bitcoin-acquisition vehicles. The company plans to generate Bitcoin-denominated cash flows by providing financial services. 21 Capital is led by long-time Bitcoin builder 

 and Tether, we hope to hear more about their roadmap after the SPAC is complete. Potential future offerings could position 21 Capital in the same premium to NAV league as other crypto financial services providers like Galaxy (GLXY), which trades at a 4x BTC NAV premium.

In the context of the changing monetary order, we are excited about our investment and we look forward to the road ahead for 21 Capital and Bitcoin’s place in the global financial system.

Announcing Hypersphere’s largest investment to-date: 21 Capital


Money evolves. It’s progressed from seashells to coins to bits of code. The final form of money may be the representation of a resource or idea that you do 

. A project building this type of idea must reorient a system, an industry, or humanity in a way that value-based interactions are fundamentally changed. We believe that 

 is building toward this effort. As the network matures and reaches its end state, Post Fiat will create a new type of currency — one that extends beyond the bounds of a store of value and kickstarts the evolution of value.

Post Fiat answers the question, “What is a human’s place in a post-AGI world?”

Artificial general intelligence (AGI) will soon arrive and demonstrate superiority to human intelligence and production. Once AI decision-making exceeds human-led decision-making, humans will increasingly seek counsel from their AI counterparts. This switch has already begun: many big banks use AI for trading; students use it for their essays; medical doctors for their diagnoses. People will only continue to integrate more and more artificially generated, yet tangible advice into additional aspects of their lives.

Post Fiat is the foundation where activity like this can thrive. It is a network that establishes an economic framework for orchestrating interactions between AI agents and groups of people. These agents can specialize in any function their builders create, such as trading on public markets, providing productivity advice, conducting smart contract audits/security, or offering autonomous customer support. And via a 

, intellectual property (IP) generated by Post Fiat agents will be licensed and monetized. Effectively, Post Fiat aggregates human intelligence, protects IP, and aligns economic incentives — humans provide high-value information to train agents, and agents produce the most value for their users.

While it’s currently on XRP Ledger (XRPL), Post Fiat will eventually transition to its own Layer-1 blockchain. Post Fiat validator nodes will be the AI agents that users interact with. In addition to providing services, nodes will run XRPL’s 

 mechanism in the backend and receive rewards largely based on the value they generate. This mechanism creates a Darwinian contest among AI agent nodes to generate the most economically useful interactions for users.

As this essay dives deeper into Post Fiat, it will be essential to understand the landscape it seeks to transform. Building off the foundational infrastructure of XRP Ledger to the broader economic and technological shifts driven by AI, Post Fiat is architecting a system designed for the post-AGI economy. In the sections that follow, this essay will explore why Post Fiat chose the XRPL stack, the problems it aims to solve, how value accrues within its network, and how it stands apart from existing projects like Bittensor. It will also examine the current AI meta, the vision of its founder, and how users can interact with the system today. Together, these insights will paint a comprehensive picture of why our bet on Post Fiat is a bet on the shift of value exchange in a post-AGI world.

XRP made headlines recently when its market cap 

 BlackRock’s at over $149 billion. At that valuation, XRP is the largest finance token in crypto and one of the largest finance-related entities in the world. While it has clear market demand, Ripple Labs built XRPL to tackle transaction banking, a much smaller part of global finance than the total TAM. HSBC, the largest transaction bank, has a smaller market cap than Goldman Sachs, the top trading and investing firm. Considering this, Post Fiat is building out functionality that will initially focus on the investment banking vertical, which is less susceptible to geopolitical risk, even under the Trump administration.

Most are unaware that XRPL was originally designed to replace SWIFT (Society for Worldwide Interbank Financial Telecommunication), the global system facilitating secure and standardized communication between financial institutions for transaction banking, which encompasses international money transfers, trade settlements, and other banking operations. Each SWIFT transaction is accompanied by a message to ensure compliance. International standards like 

 often mandate certain data to accompany transactions, such as counterparty information and the transaction’s purpose. This is why XRPL has native memos.

Memos are unstructured data — but with the advent of large language models (LLMs), they can be structured and quantified in a way that was previously impossible. The ability of LLMs to autonomously identify and abstract the most useful addresses on the XRPL blockchain (without human intervention) positions Post Fiat as a new financial primitive, redefining how AI agents measure and generate economic value.

At a blockchain level, XRPL has many attractive security features and a few drawbacks. It’s been running for over a decade without downtime, has never succumbed to any serious attacks, uses a lightweight consensus protocol, enables address whitelisting, runs a centralized selection of its validator set (called the 

, or UNL), and distributes 100% of unlocking XRP (like inflationary rewards) to the Ripple Foundation.

The Post Fiat L1 will absorb the technological benefits of XRPL while addressing its core concerns. The merits of XRPL include:

Large total addressable market (TAM) despite a sector full of regulatory risks

The XRPL TAM is the least competitive category with the highest capitalization — XRP trades at over $149 billion with 24-hour volumes in the billions.

Post Fiat’s potential TAM will be significantly higher, targeting AI use cases and investment banking.

Plain English memos are perfect for LLMs, making them a powerful feature in Post Fiat’s node selection and evaluation process.

Post Fiat will include private/encrypted memos, following banking compliance concerns and attracting financial institutions to build on the network.

Efficient performance and battle-tested ledger design

Processes 100,000 transactions for $5 with finality in a few seconds.

Has run for over a decade without downtime.

Lightweight consensus and inexpensive security budget

Post Fiat AI agent nodes can run RPCA consensus in the background with minimal CapEx.

The security model for the entire network is estimated to cost under $1 million (ChatGPT estimates about $150,000) in CapEx — Solana and Ethereum hit 8–9 figures and Bitcoin is in the billions.

These features remove the need for smart contract support in some of crypto’s most popular financialized use cases (DEXs, token issuance, etc.), enabling Post Fiat to maintain a lightweight architecture.

In addition to inheriting some of XRPL’s benefits, Post Fiat will also work to address some of the network’s concerns, as detailed in the chart below:



Post Fiat is building an AI-powered XRPL optimized for the nascent AI agent economy. It addresses XRPL’s centralization issues by using LLMs to select and reward blockchain validators based on economic activity. This approach keeps Post Fiat more aligned with the direction of the global economy and is more decentralized. Unlike XRPL’s reliance on Ripple Labs and its focus on transaction banking, Post Fiat leverages AI agents as validators, emphasizing unique value creation among nodes with reduced regulatory risk.

Post Fiat has re-architected Sybil-resistance and reward distribution by building an AI-powered system run on a novel Proof of Economic Value (PoEV) mechanism. With these advantages, we believe that Post Fiat will flip XRPL when AGI arrives — and 

 says it’s coming sooner than many realize.

Engaging the User via LLMs not Smart Contracts


Every Layer-1 blockchain delivers value through a core offering that most directly engages the user. Bitcoin offers asset storage, which lets the user monitor and control their balance with personal sovereignty. Ethereum and Solana feature smart contracts. For users, it unlocks the ability to permissionlessly engage in various activities, like speculation or borrowing, that might otherwise face significant barriers to entry. And for developers, smart contracts allow them to build autonomous financialized applications with access to global liquidity, without the concern of a back office.

Post Fiat’s core offering is the LLM. Users can interact with a particular agent (or swarm) seeking economic or other value and join a community of people also intrigued by its capabilities. Meanwhile, developers can launch a purpose-driven LLM that attracts global capital and real-time data. Then, their LLM can compete for value creation in Post Fiat’s agentic financialized society. Post Fiat doesn’t need smart contracts — the LLM is the application.

But how will people issue or trade assets onchain without smart contracts? This is the advantage of building on the XRPL stack. A lot of DeFi functionality is already supported on XRPL without the use of smart contracts. It employs 

 for launching fungible and non-fungible tokens and a native 

 powered by a CLOB. With these capabilities, Post Fiat will offer many of the same functions as other networks, even without supporting smart contracts.

By building from the XRPL stack rather than forking the EVM/SVM or creating a new PoS/PoW system, Post Fiat has deliberately chosen to prioritize efficiency, simplicity, and alignment with its goals. Post Fiat’s core functionality relies on communication between users and LLMs, not arbitrary executable logic. An LLM just needs a messaging protocol — it doesn’t require users to perform a complicated set of actions to use it, as many smart contract applications demand.

LLMs are magnitudes more powerful than applications, redefining how humans accomplish tasks. Consider the value ChatGPT offers versus individual web-based applications: a general-purpose assistant accessed via natural language that enhances a user’s performance across many tasks, compared to a rigid, narrow-purpose interface designed for a single task. The Post Fiat network offers the same paradigm shift, unlocking unprecedented opportunities for human-AI collaboration and innovation across a range of subject matters.

Proof-of-Stake (PoS) and Proof-of-Work (PoW) have inherent flaws that seldom get discussed. Both incentivize resource accumulation. To earn the most rewards in PoS, validators must acquire the most native tokens. In PoW, rewards depend on obtaining significant mining hash power or hardware to solve arbitrary cryptographic puzzles. Though both approaches reward accumulation, it is not the only path to achieve high security. Blockchains leveraging AI evaluation could incentivize direct value creation as a means to maintain security and distribute rewards.

Post Fiat’s PoEV system aligns validator incentives with producing the most value for all users, since doing so wins them the highest percentage of PFT reward distribution. To accomplish this, Post Fiat uses AI determinism to reconstruct the reward distribution mechanisms that underlay most consensus protocols. This is made possible due to a unique 

 that major LLMs share: AI judgment converges on a deterministic output given a model, prompt, 

, and set of memos. By aligning incentives to benefit users, Post Fiat creates a new type of L1 enabled by LLM technology.

Additionally, Bittensor’s current issues with reward distribution (discussed in a couple of sections) highlight the perils of making stake the default consensus mechanism for any protocol hosting AI agents. By working with a much simpler consensus mechanism that functions even without network rewards (XRPL validators don’t get paid), Post Fiat can keep a much more objective and fair system that will reward top agents.

Synthetic data has become a popular tool in LLM training due to its ability to address challenges like data scarcity, privacy concerns, and high data labeling costs. By generating artificial datasets that mimic the real world, synthetic data offers a cost-effective and scalable way to train AI models. However, LLMs are only as good as the data they train on. Slop begets slop.

 have highlighted significant limitations of synthetic data. They found that over-reliance on synthetic datasets can lead to 

, where iterative training on artificial inputs resulted in degraded performance and reduced generalization.

Though still experimental, Post Fiat is building a solution that incentivizes users to provide high-quality data to the AI agents they interact with. Incentivized both by the quality of output and PFT rewards, data from user interactions with AI agents is used to train and refine LLM responses. Instead of competing against human intelligence, Post Fiat coordinates it in ways that AI generation cannot replicate. As a result, LLMs will generate more unique insights than they could by relying solely on synthetic data.

 a few times about how analytical “edge” (i.e., advantage) is diminishing in a world full of LLMs that are infinite data scraping machines. Because AI is relatively ubiquitous and affordable, Alex realized that the edge in data is no longer in analysis but in coordination.

Alex came to this realization while developing AI-powered trading strategies for his personal portfolio. He had conversations with friends at Palantir, where he previously worked, and Citadel revealing that institutions had already incorporated AI into their trading operations. Alex realized that individuals cannot compete against entities with far greater resources. So, he began to reconceptualize where edge could be found in the markets, concluding that it could be derived from a network with effective intelligence gathering and economic lock-in. That is, by incentivizing large numbers of people to provide valuable insight to AI agents, a hivemind of quality information will get routed through the network and repackaged to its users.

Post Fiat is preparing for a world where AGI will reorient society in a way that fiat currencies will become obsolete. In this world, the next-best currency would be the one that most appeals to AGI, facilitating AI-based transactions and offering the most compatibility for AI functionality.

Before discussing how to value PFT, it’s important to understand the economics of currencies and L1s to better contextualize Post Fiat’s market size.

Our debt-based monetary system requires increasing levels of liquidity to finance existing debt levels. The value of money in this system is based on our collective (and circular) belief in it; it’s also constantly being debased through additional debt issuance. Despite this, hundreds of fiat currencies continue to function as accepted forms of value exchange. However, as liquidity expands and fiat money is debased, other assets in the economy — such as Bitcoin or gold — begin to take on more money-like characteristics, increasing their legitimacy.

 for valuing currencies is based on the “total number of currency units [to the] total demand for currency relative to goods and services denominated in that currency both fixed and forward contracted.” Put simply, value accrual is a tug-of-war between a currency and the demand per resource it can purchase.

For example, imagine a simple economy where the demand for total resources remains constant. To start, you have 10 currency units and 10 resource units, where resource units represent a society’s collective goods and services:

If you turn on the printer, increase the currency units to 20, and keep resources the same, you’ve introduced inflation, where each resource costs twice as much. In this scenario, the value of your currency has fallen.

If you increase production and now have 20 resource units and keep currency units at 10, you achieve a deflationary effect where your currency goes twice as far. That is, the value of your currency has increased relative to the goods and services it can purchase.

This relationship between currency units and demand per resource explains how hyperinflation has been prevented despite excessive USD printing — though during the pandemic, money grew much faster than GDP/production, causing price increases and real inflation in goods and services. Regardless, hyperinflation has remained at bay because the resources denominated in USD have seen increased demand and production (a proxy for US economic expansion) about as fast as the increase of the money supply. This excess of value creation has largely come from the 

, whose software products have replaced oil as the primary resource that cements the USD as the global reserve currency. In turn, the US tech sector drove this replacement by creating infrastructure and platforms essential to businesses worldwide, generating extraordinary demand for USD.

Applying this framework to crypto, the question becomes: 

What can you buy with crypto besides dollars?

L1s are valuable because their resource TAM is bigger than any single application’s market built on top. Additionally, an L1’s fundamental structure makes P/E ratios and revenues irrelevant to their asset values — because currency value is based on demand/resources, not security. And L1s are micro-economies, not businesses. Validator earnings provide insights into the security budget of an L1, while economic demand — reflected in transaction volumes — is a 

 of the overall value of an L1’s native token.

The value flowing through Post Fiat will be based on the demand for helpful AI services. The denomination for these services will be PFT. While PFT aims to be the dominant economic primitive behind all types of AI interactions, from finance to philosophy, Post Fiat will start with investment banking use cases and expand from there. Post Fiat’s biggest risk will be whether or not useful agents can generate real returns for users. If not, the market will underwhelm. If they do, we expect Post Fiat to follow a trajectory similar to OpenAI’s ChatGPT, where demand continues to surge even after a 

Bittensor has been one of the top AI tokens by 

 for over the past year. Before the AI agent meta, few assets were competitive to TAO valuation-wise. Bittensor’s value initially surged partially due to memetics (comparing Bittensor to Bitcoin in the AI meta) and partially by merit. Bittensor is interesting to analyze due to the parallels with Post Fiat and how it differs.

 to itself as a “decentralized machine learning” protocol. And it is. However, it is also a political organization.

At its core, Bittensor incentivizes an ecosystem of competition markets/venues that host the deployment of ML models. The performance of these models is then rewarded with TAO emissions (inflation). These venues are called subnets, which act as the operational layer defining the tasks that miners should pursue when they contribute resources (e.g., deploy ML models).

Anyone can deploy a subnet on Bittensor or join the network as a miner (resource contributor) or validator (subnet evaluator). While subnet validators help set the reward weights for miners within the subnet, a different process determines which subnets receive a higher portion of inflationary TAO rewards.

This part of the process is where Bittensor becomes political. The top 64 validators by stake weight are selected to join the 

, a special subnet that determines which subnets receive a larger share of TAO rewards. The determination process is ad hoc, however, rather than objective, measurable, or auditable. Instead, validators represent their own interest and “help distribute emissions to what 

 are the most valuable and productive subnets.”

Regardless, there’s no core methodology. Validators are run by humans, who ultimately have the discretion over how to distribute rewards. This system can incentivize unfair dealings (i.e., bribery or clandestine deals) if the identities of top validators are known. Root subnet validators could even give higher weights to the Root subnet, which would essentially 

The subnet design is somewhat similar to Post Fiat’s with a few notable caveats, as displayed in the chart below:



We expect Post Fiat to eventually surpass Bittensor in both usage and value. Though the Bittensor community is planning an 

 that would distribute the decision-making, it has yet to show any interest in supporting deterministic processes, which are more trustless, efficient, and fair. Even if it did consider this route, it would be structurally difficult to implement due to how Bittensor’s network security is constructed.

The current AI agent meta is a meme. Kicked off by Goatseus Maximus (GOAT) and currently led by Virtuals (VIRTUAL), most projects building in this category lack any sort of economic flywheel or even basic value accrual. Instead, they rely on virality and memes.

Analyst agents launch as slight variations of AIXBT with questionable token-gated terminal analytics. Launchpads seek to mimic the Virtuals model, only creating a cleaner website UI (though some don’t seem to understand the importance of incorporating their native token as platform currency). Many AI hedge funds are just forking the Eliza stack, trading at unjustified market cap-to-NAV ratios. Even AI16Z (worth over $1 billion) trades off pure potential — at the time of writing, there is no value accrual, no integration, and nothing tangible that connects the token with the project.

Regardless, the market has an appetite for AI projects with tangible uses (or that advertise it) tied to their token, such as Virtuals, AIXBT, and Griffain. The current environment even lacks explicit incentives (minus speculation) to build real utility that accrues value to users. Post Fiat is filling this void, adding a reflexive element to AI agent development. With its incentive structure, we expect the following flywheel to emerge:

Agents gain more quality data and improve output.

Agent output generates more economic value for users.

Agents and users accrue PFT rewards, attracting more users.

This self-reinforcing loop — where more useful AI output leads to more usage, which generates more data, and results in even better AI — creates a compounding cycle of growth. It also creates a unique barrier to entry that forces product-market fit for agents. In contrast, platforms like Virtuals lack a de facto filtering system, enabling bad or useless tech, memes, and scams to proliferate. By aligning incentives with the creation of useful agents, the Post Fiat platform drives demand for agent interactions, ultimately fueling the adoption of PFT.

In addition, Post Fiat is designed with institutional and advanced trading use cases in mind. Its flagship Artificial General Trading Intelligence (AGTI) node, currently under development, will offer sophisticated financial tools, such as data validation, index creation, and Diffie-Hellman privacy functions. These functions will aid in the compliant sharing of expertise and financial data in exchange for rewards.

, the founder of Post Fiat, has been a leading voice for the recent crypto-AI boom. He 

 the agent economy back in April, coined the term 

 (which he’s now working to create), and is launching one of the most 

After graduating from Wharton, Alex spent time at CitiFX and was one of Palantir’s first major users to apply data intel to FX trading. He later joined Palantir to lead capital markets and apply the firm’s software to predictive analytics with big data from companies, such as JP Morgan, Mastercard, and Standard Chartered. At Standard Chartered, he worked with large volumes of SWIFT data for transaction banking compliance and investment bank signal extraction. He also managed $750 million at Balyasny Asset Management as the lead analyst on a global TMTC team. At Balyasny, Alex pioneered a new type of long/short investing strategy that used real-time measures of advertising performance data to predict asset performance.

He leveraged this experience to co-found Perpetua, an ad tech company that started out in capital markets intelligence. It worked with Balyasny Asset Management and the top-performing global macro hedge fund with over $50 billion in AUM to measure LTV/CAC for all publicly traded tech, media, and telecom companies. Perpetua later evolved to be one of the largest Amazon advertising SaaS companies with a strategy focused on 

. Perpetua was acquired by Ascential in 2021 for $150 million and resold to Omnicom as part of the $865 million Flywheel Digital deal.

Alex sits on the spectrum of autistic scientist and genius advertiser. He’s driven by meaning and intends to leave an 

 legacy through Post Fiat. In addition to often espousing non-consensus viewpoints and investing in them before they’re profitable, he also aims to live in the future (as evidenced by the routine he described in 

). Alex is the type of founder we get excited about at Hypersphere; he’s the type of founder we see building a category-defining project.

Post Fiat’s end state will feature an L1 powered by AI-driven decision-making. In it, users will be rewarded for interacting with the network, and agents will be paid for generating economic value for users. Users can already interact with the protocol on XRPL via 

 (and soon a web application) as the Post Fiat team continues to develop its vision.

 post, Alex laid out a detailed explanation of how to set up a wallet and interact with the first Post Fiat node, the Task Node. The Task Node is a productivity tool trained to help users set objectives, refine them, and suggest the best tasks to accomplish them. Users are currently rewarded with PFT for interacting with the node and will also be airdropped more PFT tokens in the future for being early adopters.

Post Fiat and other community members are currently building more nodes, including an NFT minter, a church, a social sharing tool (built by Wombo, an NVIDIA-backed startup), a financial data tool (AGTI Node by Post Fiat), and a trading agent (

Perhaps the most exciting of these is the Corbanu trading agent. In addition to the pending social media presence, we are eager to see how the multi-manager structure of Corbanu performs, given our prior experience running a prominent multi-manager hedge fund. It will essentially run five strategies as a swarm of five agents focusing on day trading, post-earnings, macro, bubbles, and pre-catalyst.

Our bet on Post Fiat hinges on the imminent arrival of AGI and onboarding the real world.

First, we see AGI significantly shifting employment responsibilities and human trust in machine intelligence. The current economy is already seeing employers 

 workers with AI. This will only accelerate as AGI reaches maturity. This shift not only redefines the nature of work but also paves the way for systems that better integrate human input and AI capabilities. Eventually, Post Fiat could become the best-designed human intelligence gathering and agentic coordination layer. 

In the new economy full of infinite data scraping machines, most data and analysis will be rendered useless. This will force us to use digital solutions that can track data provenance, license it, and monetize it on a globally shared and permissionless database — it will force us to use blockchains. With most data already public and unprotected, systems will need to be built around these digital solutions that incentivize groups of people to provide insight, data, and intelligence to particular AIs to ensure their scarcity and continue the training, fine-tuning, and progress of various models. With text-based chat being the primary form of communication with AI, a blockchain with native support for messaging will form the backbone of the new economy. It will also demand a new digital currency that is not subject to the same manipulation as government-issued money. Post Fiat will provide the infrastructure, and PFT will be the new currency.

We also see major potential in onboarding investment banks, hedge funds, and advanced AI-driven services. With these, the network could capture large flows of capital and generate real business utility. In a winner-takes-most dynamic, top-performing AI agents would attract more user demand, leading to more data and rewards, which in turn enhance their AI capabilities. This creates a virtuous cycle that increases demand for PFT as a payment and reward mechanism.