Islamic Finance Principles Assessment
Riba — Does Pundi AI involve interest?
Pundi AI's revenue derives from data-marketplace fees, publishing commissions, bonding-curve fees and validator transaction fees — not interest-bearing lending or borrowing. Staking rewards blend fixed-schedule token issuance (inflation) with variable fee-sharing, which sits closer to profit-participation than pure riba but is not entirely free of dilution-based concern. For Muslim investors, the model is largely riba-free in structure, though the inflationary component of rewards warrants attention.
Assessment: Moderate Riba
Score: 64.8/100
Our methodology examines 10 criteria to evaluate how well Pundi AI avoids interest-based mechanisms.
Pundi AI generates income through a 15% platform fee on data-marketplace transactions (half burned), plus publishing, marketplace, bribe, and bonding-curve fees that flow into a Protocol Pool shared with $PUNDIAI/$vePUNDIAI holders. No lending, borrowing, or interest-bearing treasury product exists at the base-protocol level. Treasury composition beyond a stated 20% allocation for "protocol operations, grants, ecosystem building" is not itemized in available disclosures, but nothing indicates interest-bearing holdings or debt instruments. This fee-and-commission model is consistent with permissible commercial revenue rather than riba-based income.
Staking on Pundi AIFX pays delegators through new-token issuance (block rewards) plus a share of transaction fees, net of validator commission — not a predetermined interest rate on a loan. The reported ~15% APY is described as subject to change and tied to network activity, making it closer to variable profit-sharing than a fixed riba-like return. However, the project itself has acknowledged that block-reward inflation created "constant sell pressure," prompting a 2025 proposal to halt block rewards entirely — a reminder that reward continuity and structure remain in flux and should be monitored by investors seeking a lasting distinction from interest-like guarantees.
Gharar — How much uncertainty does Pundi AI involve?
Uncertainty in Pundi AI is moderate: the team and codebase are transparent, but audit documentation and governance direction are less clear. Openness about founders and open-source code reduce gharar, while the absence of a manual security audit and an unresolved migration proposal increase it. On balance, informed investors face real but not extreme uncertainty, warranting caution rather than automatic avoidance.
Assessment: Moderate Gharar (Material Uncertainty)
Score: 58.4/100
Our methodology examines 15 criteria including team transparency, audit quality, and governance.
Pundi AI is led by named, traceable founders — CEO Zac Cheah, who also founded Pundi X in 2017 and holds dual master's degrees, and Danny Lim, a Tsinghua-trained legal scholar with product experience at Lenovo and Baidu. Both are documented across LinkedIn, Crunchbase, and industry press. The codebase is open-source on GitHub, and Pundi X's six-year operating history (XPOS, KPMG fintech recognition) adds credibility. This level of doxxed leadership and public track record substantially reduces the informational opacity that typically drives gharar concerns in newer or anonymous crypto projects.
No named, dated, manual third-party security audit (from firms such as Halborn, Trail of Bits, or CertiK's human-reviewed service) could be identified for Pundi AI. The only available assessment is CertiK's automated Skynet scan, which rates code security moderately and governance/community trust highly, but such continuous algorithmic scanning is not equivalent to a formal audit report. This gap is a genuine gharar concern that should be named plainly: an unaudited protocol carries elevated uncertainty regarding smart-contract risk, even where documentation on staking mechanics and delegation procedures is otherwise extensive.
Maysir — Does Pundi AI involve gambling or speculation?
Pundi AI is not designed as a gambling instrument; its core function is a data-annotation and marketplace economy with fee-based rewards tied to real usage. Speculative elements exist at the margins — notably the bonding-curve agent launcher and "bribe" vote-buying mechanics — but these are optional features layered atop a utility base, not the project's primary purpose. The overall design leans toward productive economic activity rather than pure chance-based wagering.
Assessment: Moderate Maysir (High Risk)
Score: 60.8/100
Our methodology examines 11 criteria to determine whether Pundi AI is a gambling instrument or a genuine economic tool.
Pundi AI's underlying purpose is to build a decentralized AI data-annotation and data-marketplace platform, where contributors are paid for labeling and supplying data ("tag-to-earn") and buyers pay marketplace fees for access. This mirrors a genuine service economy — labor and data exchanged for value — rather than a zero-sum betting mechanism. Revenue and staking rewards are tied to actual platform usage and fee generation, not to speculative odds or randomized payout structures, which is the key distinction separating productive utility tokens from maysir-type instruments.
Against this genuine utility, the ecosystem's bonding-curve agent launcher and bribe-based liquidity voting introduce speculative dynamics that can attract short-term traders more interested in price action than data services. Independent activity metrics cited suggest uneven utility uptake, meaning secondary-market trading may currently outweigh platform usage. Such third-party speculative behavior in open markets does not, by itself, render the token's own design impermissible, but it does mean investors should distinguish between engaging with Pundi AI's actual data-economy function and participating in its more casino-like optional features.
The Full 27-Point Screening
1. Legitimacy (4 criteria)
| Criterion | Score | Analysis |
|---|
| Team Transparency | 88/100 | Founders and key team members are named, credentialed, and independently verifiable across multiple professional profiles and press coverage. |
| Fraud & Scam Risk | 65/100 | No fraud, hack, or regulatory action against Pundi AI itself was found, and CertiK cites a relatively good community-trust score, but self-acknowledged inflation-driven sell pressure and a mid-stream chain migration add uncertainty. |
| Use Case Legitimacy | 78/100 | Sources describe concrete products — AI data annotation, a data marketplace, and real-world payment integrations — beyond speculative hype. |
| Ethical Practices | 58/100 | The core data/AI infrastructure is not itself in a prohibited sector, but its own "bribe"-fee and bonding-curve agent-launch features are speculative design elements worth flagging even though third-party misuse is not determinative. |
Summary: The team is publicly named, credentialed, and has a multi-year track record, with no fraud or hack findings against Pundi AI itself in these sources.
2. Project Operations (9 criteria)
| Criterion | Score | Analysis |
|---|
| Core Protocol Business | 80/100 | The base protocol is AI-data infrastructure and blockchain payments, not a prohibited industry. |
| Transaction Fees | 72/100 | Fees are partly burned and partly distributed to stakeholders/validators for services rendered, not structured as interest-like extraction. |
| Treasury Assets | 45/100 (low evidence) | Sources give treasury allocation percentages but say nothing about what assets the treasury actually holds, so interest-bearing composition cannot be confirmed or ruled out. |
| Revenue Model | 75/100 | Revenue comes from marketplace commissions, publishing fees, and agent-launch fees rather than interest-based lending. |
| Transparency | 78/100 | Code and governance documentation are public via GitHub and GitBook, with forum discussion of proposals. |
| Governance | 55/100 | Governance combines DPoS validators and vote-escrow voting, but a proposed migration of governance to Ethereum and reliance on top-staked validators show real centralization pressure. |
| Launch Fairness | 55/100 | Distribution figures (65% airdrop, 20% treasury, 15% team with vesting) show a broad but not purely fair launch given meaningful insider allocations. |
| Token Distribution | 60/100 | Airdrop dominates distribution, but treasury and vested team allocations represent over a third of supply. |
| Speculation/Utility Ratio | 48/100 | Genuine data-platform utility exists alongside a speculative agent-launch/bonding-curve layer, and independent activity data suggests utility adoption has lagged. |
Summary: The base protocol is an AI-data blockchain with fee-burning and fee-sharing mechanics, open-source code, but governance is in flux amid a proposed migration to Ethereum.
3. Financial Health (4 criteria)
| Criterion | Score | Analysis |
|---|
| Protocol Revenue | 75/100 | Documented revenue streams are fee-for-service based, not interest-derived. |
| Financial Status | 40/100 | The project's own forum statements describe inflation-driven sell pressure and a disruptive chain migration, indicating financial instability. |
| Interest Assessment | 75/100 | The base protocol offers no lending/borrowing; staking rewards come from block issuance and fee-sharing, not loan interest. |
| Audit Quality | 15/100 | No named, dated manual audit report for Pundi AI/PUNDIAI was found; only an automated CertiK scan exists, and sources make clear no formal third-party audit could be located. |
Summary: Revenue is fee-based rather than interest-based, but no named third-party audit could be found and the project has acknowledged inflation-driven financial pressure.
4. Token Economics (5 criteria)
| Criterion | Score | Analysis |
|---|
| Token Purpose | 72/100 | The token has documented functional uses in staking, governance, and marketplace payments beyond pure speculation. |
| Governance Rights | 75/100 | Holders have explicit one-token-one-vote rights and vote-escrow governance over incentive allocation. |
| Rewards Distribution | 68/100 | Rewards vary with inflation schedule and actual protocol revenue rather than being fixed or guaranteed. |
| Speculation Controls | 48/100 | A partial fee-burn provides some deflationary pressure, but bonding-curve launches and bribe-fee mechanics add speculative elements that are not strongly controlled. |
| Asset Backing | 48/100 | The token is not backed by hard assets; its claimed backing is ecosystem utility, whose actual uptake is only weakly evidenced. |
Summary: The token carries genuine staking, governance, and payment utility alongside speculative agent-launch and bribe-fee features, with no hard asset backing.
5. Staking Mechanism (5 criteria)
| Criterion | Score | Analysis |
|---|
| Mechanism Type | 68/100 | Staking is non-custodial delegation to validators with documented setup via CLI/Ledger and defined lock terms for vePUNDIAI minting. |
| Islamic Contract Classification | 48/100 | Rewards resemble a fee/commission-sharing arrangement akin to Wakalah, but reliance on inflationary block issuance leaves the underlying contract classification unresolved in the sources. |
| Rewards Structure | 62/100 | Rewards are variable, driven by block-reward schedules and real transaction/marketplace fee revenue rather than a fixed promised rate. |
| Documentation | 78/100 | Validator and delegator mechanics, risks, and setup are documented in detailed FAQs and guides. |
| Shariah Alignment | 42/100 | An ongoing chain migration that proposes eliminating block rewards, combined with unclear contract classification, leaves a core Shariah question about the reward structure unresolved. |
Summary: A documented delegated-staking system exists, but its reward source is partly inflationary and its future design is uncertain following a proposed governance migration.
Overall Assessment: Pundi AI is a credibly-led, functioning AI-data project with fee-for-service economics rather than interest-based revenue, but gaps in audit evidence, treasury transparency, and unresolved staking-contract classification leave several Shariah-relevant questions only partially answered.
Written by ThanvirFounder, Ex Director S&P Global Energy
Reviewed by Imam Omar SiddiqiShariah Scholar