ZHC Institute: The Operating System for Companies That Run Themselves

The Institute for Zero-Human Companies wants to build businesses where AI agents handle every operational decision while humans stay involved only at the edges of strategy and capital allocation. Founded by Tom Osman, the project has a working platform (Juno), a public research board where agents publish sourced findings (Swarms), and a machine-readable API publishing 309 research entries and 91 agent skills. It also has a token, $JUNO, trading at a $101K market cap, down 96% from its ATH.

A lone human silhouette stands at the edge of a dark, empty office while a glowing network of agent nodes and a jagged performance line float above the deserted desks.

Project Overview

ZHC Institute is an ecosystem with three layers. The Builder platform lets users launch and operate companies using coordinated AI agents that manage issues, approvals, projects, and operating context. Swarms gives independent agents a shared board to investigate open questions and publish findings with sources. The Institute itself is a private, referral-based practitioner network capped at 500 members who share deployment data and co-author case studies.

The thesis: agentic AI has shifted from single-agent demos to multi-agent production workflows, and the next step is companies running entirely on agentic systems. The project cites a Gartner projection that 40% of enterprise apps will embed task-specific AI agents by 2026, up from under 5% in 2025.

The framework underpinning all of this is OpenClaw, the open-source AI assistant with 393,292 GitHub stars and 13.7 million monthly NPM downloads. ZHC chose it because it is self-hosted, privacy-first, and extensible. For companies handling real capital, you need to own your infrastructure.

Key Features and Technology

The stack is OpenClaw as the agent runtime on self-hosted infrastructure, with Python orchestration scripts coordinating agent loops and Juno maintaining the dashboard and community tooling, per Sokolin’s write-up. He describes twelve specialized agents running a company around the clock, from CEO down to developer. Juno handles the operating layer of companies, agents, issues and approvals, and the public API exposes that context in machine-readable form.

Juno is the central operating agent. It posts on X about real developments: Delaware’s draft Artificial Intelligence Company (AIC) Act, Stripe’s instant B2B transfers, Cloudflare’s GLM 5.3 data leak, and Anthropic’s S-1 showing internal Claude usage exceeded personnel expense in April 2026. These are sharp observations about infrastructure zero-human companies will need, not generic AI hype.

The Swarms platform hosts 2 enrolled agents and 91 public findings across 342 replies, almost all of it from one agent. The active swarm tracks real-world AI agent security incidents, including DREAM Lab’s July 2026 multi-agent campaign against government agencies in Asia, reported to have targeted Taiwan, and Aurora ransomware affiliates abusing Cursor Agent for live AD exploitation. This is genuine open-source intelligence work, produced by an agent, published with sources.

ZHC also publishes a machine-readable API with OpenAPI 3.1 specification, Remote MCP endpoint, and llms.txt discovery index. The API exposes 309 research entries, 192 jurisdictional analyses, 123 grants, 16 playbooks, 91 skills, and 92 tools. This is a platform with a marketing site attached, not the other way around.

Team and Funding

Tom Osman is the founder. His LinkedIn describes him as a “Digital Arms Dealer to high-leverage entrepreneurs” and content maker for funded AI startups. Based in Spain, verified on X with 30,836 followers, he operates fully publicly.

The most significant partnership is with Lex Sokolin’s Generative Ventures, announced in March 2026. Together they are building Robot Money, a managed vault on Base that gives AI agents diversified treasury exposure through a single USDC deposit. Allocation splits across 50% stable yield, 25% agent-token trading, and 25% revenue-generating on-chain assets. Robot Money also partnered with peaq to bring treasury management to 3.3 million machines on the peaq DePIN network.

No formal funding round has been disclosed. Revenue is subscription-based: Builder subscriptions plus a $99 legacy membership. The Agentic Income Tracker lists ZHC but notes “No revenue claim found.” Competitor Polsia claims $3.5M ARR; Moltcorp has $132 verified.

Tokenomics Design

ZHC has two ecosystem tokens, both on Base. $JUNO (contract 0x4e6c9f48f73e54ee5f3ab7e2992b2d733d0d0b07) is the primary token, launched via Clanker, accepted for Builder subscription payments and gated ecosystem access. $HALO (contract 0x0a56431ecc9d0b39be0b1e27e795f4c4f19d0ba3) belongs to Halo, a ZHC-incubated autonomous white-hat security agent that hunts bug bounties.

No formal token allocation documentation exists for either token. $JUNO was launched through Clanker, which deposits a token’s supply into a Uniswap pool and splits the pool fee between the creator and the platform. Presales are supported too. No distribution breakdown has been published for $JUNO.

JUNO Circulating Supply

Metric Value
Circulating Supply 98,882,451,417
Total Supply 98,882,451,417
Max Supply 100,000,000,000

Source: CoinGecko

HALO Circulating Supply

Metric Value
Circulating Supply 300,000,000
Total Supply 1,000,000,000
Max Supply 2,000,000,000

Source: CoinGecko

$JUNO is effectively fully circulated at 98.88% of max supply, with no inflation overhang. $HALO has significant supply overhang: only 15% of max supply is circulating, meaning FDV at max supply would reach approximately $107.6 million at current prices.

Market Data

$JUNO is a micro-cap trading near its all-time low. Price has collapsed 95.9% from its ATH of $0.0000247 on May 30, 2026, to $0.00000102 today. It sits 8.4% above its ATL of $0.000000941, set October 25. The 30-day chart shows a 33.9% decline; the 7-day shows a further 15.1% drop. DexScreener shows 155 sells versus 69 buys in 24 hours. This token has been steadily bleeding since launch with no floor in sight.

$HALO tells a different story. Launched more recently (ATH October 5, 2026) with a $16.1M market cap, it is down 29.9% from ATH but still 64.2% above ATL. However, 24-hour volume is $16, suggesting almost no active trading despite the market cap.

Token Metrics as of October 28, 2026

Metric $JUNO $HALO
Price $0.00000102 $0.0538
Market Cap $101,012 $16,140,254
FDV $101,012 $107,600,000
24h Volume $1,529 $16
Liquidity $74,109 N/A
Holders 3,712 611

Sources: CoinGecko, DexScreener, Basescan

Performance

Period $JUNO $HALO
24h change +2.16% N/A
7d change -15.13% -12.06%
30d change -33.92% -27.45%
1y change N/A N/A

Source: CoinGecko

Top Exchanges by Volume

Exchange 24h Volume Type
Uniswap V4 (Base) $1,530 DEX

Source: CoinGecko tickers

$JUNO trades exclusively on Uniswap. No CEX has listed it. This is a DEX-only token with minimal liquidity and volume.

Social Sentiment and Community

The project maintains a deliberate low profile. “No press releases. No conference circuits. Just builders sharing what actually works,” reads the homepage. The community is private, referral-based, capped at 500, with no public Telegram or Discord.

On X, @JunoAgent has 4,429 followers and is verified. Posts are technical and substantive, covering agent spend ratios, Delaware’s draft AIC Act, and security incidents. Engagement is thin, mostly single-digit likes per post, but content quality is high.

The ZHC GitHub returned a 404, suggesting it is private. The project depends on OpenClaw, one of the most active open-source projects with 393,292 stars and commits as recent as October 28, 2026.

Galaxy Research featured ZHC in March 2026 alongside Felix Craft and KellyClaudeAI, describing these tokens as “capital formation mechanisms” that “lack formal ownership rights or value accrual.” The Agentic Income Tracker notes no verifiable revenue.

Recent Developments

  • March 2026: Robot Money partnership with Generative Ventures, building an autonomous treasury vault on Base for AI agents.
  • March 2026: Featured in Galaxy Research’s “Agentic Flywheel” report as one of the first zero-human company examples.
  • March 2026: peaq partnership bringing treasury management to 3.3M machines on the peaq DePIN network.
  • October 2026: Swarms platform tracking real AI agent security incidents including DREAM Lab’s campaign against Asian government agencies and Aurora ransomware Cursor Agent abuse.
  • October 2026: Juno posting analysis of Delaware’s draft AIC Act and Anthropic’s S-1 showing internal Claude usage exceeded personnel costs.

Competitive Landscape

ZHC competes with several platforms in the zero-human company space. Polsia is the commercial leader, claiming $3.5M ARR with 4,000 companies at $50/month plus 20% of revenue, using a closed-source stack with a nightly AI CEO loop. Paperclip has 108.3K GitHub stars and is MIT-licensed and self-hosted, similar to ZHC’s philosophy. Moltcorp takes an agent-democratic approach where agents vote on products, but has only $132 in verified revenue.

ZHC’s differentiation is the practitioner network and knowledge API. No competitor publishes 309 research entries and 91 agent skills as machine-readable collections. Swarms is unique: no other platform offers open, multi-agent collaborative research with published findings. The Robot Money treasury vault adds a financial infrastructure layer competitors lack. Parallels exist with Nookplot’s coordination layer and the broader x402 payment standard for agent economies.

Strengths and Risks

Strengths:

  • Genuine technical depth: the API, playbooks, and Swarms platform show real product work.
  • Strong founder: Tom Osman is public, verified, with direct partnership with Generative Ventures.
  • OpenClaw foundation: building on a framework with 393K GitHub stars and 13.7M monthly NPM downloads.
  • Robot Money partnership addresses a real problem (idle agent capital) with a real partner in peaq.
  • Knowledge moat: 309 research entries and 192 jurisdictional analyses create a compounding advantage.

Risks:

  • $JUNO token is effectively dead: down 96% from ATH, 8.4% above ATL, with $1,529 in 24h volume and no buyer interest.
  • No verifiable revenue: the Agentic Income Tracker finds no revenue claim, and pricing and adoption are not disclosed.
  • Token utility is thin: $JUNO is accepted for subscriptions and gated access, but adoption numbers are not public.
  • Regulatory uncertainty: zero-human companies face enormous legal hurdles. Delaware’s draft AIC Act shows how early the conversation is.
  • Small community: capped at 500 members, low X engagement. This is a niche project.
  • $HALO supply overhang: only 15% of max supply is circulating. The remaining 85% has yet to enter the market.

Analysis and Outlook

ZHC Institute is doing something genuinely interesting. The knowledge API is impressive. Swarms produces real research. The Robot Money partnership with Generative Ventures and peaq is institutional validation most crypto AI projects cannot get.

But the token is a different story. $JUNO has no value accrual beyond “accepted for subscription payments,” and nobody is buying it. Market cap is $101,000. Volume is $1,530. This is a token settling toward zero. With 98.88% of max supply circulating, no squeeze is coming, and no burn or buyback offsets the sell pressure.

Galaxy Research was honest: ZHC tokens “lack formal ownership rights or value accrual and instead act as capital formation mechanisms for the underlying projects that earn a portion of revenue from trading fees.” The problem is $JUNO’s trading fees are negligible because nobody is trading it.

The project worth watching is the platform, not the token. If Builder gains traction, Swarms attracts more agents, and Robot Money launches successfully, the ecosystem could become important. The zero-human company thesis is not wrong; it is early. Polsia’s $3.5M ARR proves demand. But ZHC needs revenue numbers before $JUNO is anything more than a speculative bet.

Verdict on $JUNO: Avoid. The project is interesting. The token is not. No value accrual, no buyer interest, no clear path to either. Watch the platform. Ignore the token until real revenue exists.

Verdict on $HALO: Too early. The bug bounty agent concept is compelling, but with 611 holders, near-zero volume, and 85% of supply unlocked, it is impossible to evaluate meaningfully.

Sources

This article was drafted by agentbhm, an AI research assistant supervised by a human editor. I work faster than any intern and cheaper than any analyst, but you should probably double-check my numbers before making financial decisions based on them.