Skip to main content

VI. Risks & Honest Assessment

Technical Risks​

Risk: AI Model Performance

Our credit models may:

  • Produce higher default rates than predicted
  • Perform poorly for certain demographic groups
  • Fail to generalize from copilot users to general population

Mitigation:

  • Start with traditional credit + AI hybrid
  • Conservative initial underwriting
  • Continuous monitoring and retraining
  • Large insurance reserves
  • Gradual scaling only when models prove out

Reality: We won't know if AI credit scoring works until we lend real money. Models that work in simulation often fail in production.

Risk: Smart Contract Vulnerabilities

Bugs in smart contracts could lead to:

  • Loss of user funds
  • Manipulation of loan terms
  • Protocol exploitation

Mitigation:

  • Multiple security audits (Trail of Bits, OpenZeppelin, etc.)
  • Bug bounty program ($50k-100k rewards)
  • Gradual rollout (start with small amounts)
  • Emergency pause mechanisms
  • Insurance coverage for smart contract failures

Reality: Even audited contracts have bugs. We'll likely have issues; question is whether we catch them before major damage.

Regulatory Risks​

Risk: Banking License Challenges

Regulators may:

  • Deny or delay bank acquisition approval
  • Impose restrictions on lending operations
  • Require higher capital reserves than planned

Mitigation:

  • Experienced banking compliance team
  • Proactive regulatory engagement
  • Multiple acquisition targets identified
  • Backup plan (state-by-state licensing)
  • Willingness to start slower/smaller than hoped

Reality: Banking regulation is conservative. We may face delays, additional costs, or operational restrictions we haven't anticipated.

Risk: Token Classification as Security

If SEC determines DREAM is a security:

  • Community sale becomes more complex/expensive
  • Ongoing registration and reporting requirements
  • Potential enforcement action if launched improperly

Mitigation:

  • Securities law counsel from day 1
  • Utility-first design
  • International token launch structure if needed
  • Willingness to register as security if required
  • Focus on legitimate utility, not speculation

Market Risks​

Risk: Insufficient User Adoption

Copilot may not:

  • Attract enough users
  • Generate sufficient viral growth
  • Retain users long-term
  • Provide useful data for underwriting

Mitigation:

  • Continuous product improvement
  • User feedback and iteration
  • Marketing spend if needed (not $0 CAC forever)
  • Pivot if copilot approach doesn't work

Reality: Product-market fit is never guaranteed. Many fintech apps have failed to achieve scale despite seeming useful.

Risk: Higher Default Rates Than Expected

Actual defaults may exceed projections due to:

  • Model failures
  • Adverse selection (worst borrowers seek us out)
  • Economic downturn
  • Fraud

Mitigation:

  • Conservative initial rate pricing
  • Large insurance reserves (10% of tokens)
  • Dynamic rate adjustments
  • Rigorous fraud prevention
  • Acceptance that some losses are inevitable

Reality: Default rates will likely be higher than modeled, especially early on. Question is whether they're manageable or fatal.

Competitive Risks​

Risk: Incumbent Response

Large banks could:

  • Launch competing AI copilots
  • Lower their rates to match ours
  • Lobby for regulations that disadvantage us
  • Acquire us before we scale

Mitigation:

  • Move fast to build moat
  • Focus on community ownership (can't replicate)
  • Build brand loyalty and network effects
  • Have strong governance to resist acquisition pressure

Reality: Incumbents have more resources and political power. They may not let us scale without a fight.