When a former Treasury Secretary tells public companies they can stop reporting every 90 days, he's really saying AI is moving faster than quarterly earnings cycles can capture.
The Summary
- Steven Mnuchin backed the SEC's proposal to let US companies shift from quarterly to semiannual earnings reporting, speaking at the Milken Institute Global Conference
- The timing matters: as AI capital expenditures balloon and deployment timelines stretch, quarterly windows create perverse incentives
- This isn't just accounting reform — it's acknowledgment that innovation cycles and reporting cycles have diverged
The Signal
Mnuchin's support for semiannual reporting lands while companies are pouring billions into AI infrastructure with multi-year payback horizons. The quarterly earnings treadmill forces CFOs to justify investments that won't show revenue impact for 18 months. That's manageable when you're optimizing ad spend. It's brutal when you're building agent platforms.
The proposal gives executives breathing room to make long-term bets without explaining every quarter why the spending hasn't turned into profit yet. Microsoft's $80 billion AI infrastructure commitment doesn't fit neatly into Q2 guidance. Neither does training foundational models or building agent orchestration layers that might not generate revenue until 2027.
"Quarterly reporting creates artificial urgency around investments that need years, not quarters, to mature."
Mnuchin also addressed AI's broader economic impact at Milken, though the specifics weren't detailed in the Bloomberg coverage. What matters is the framing: a former Treasury Secretary sees AI spending not as tech sector excess but as capital reallocation significant enough to discuss alongside Fed policy and federal deficits.
This connects to a larger shift. Companies building in Web4 — the agent economy, autonomous systems, AI that builds AI — operate on timelines incompatible with 90-day reporting cycles. You can't build a useful enterprise agent in one quarter. You can't tokenize supply chains or stand up decentralized compute networks on earnings call schedules.
Key dynamics at play:
- AI infrastructure investments front-load costs, back-load returns
- Agent development requires sustained R&D without immediate revenue signals
- Public market pressure for quarterly growth conflicts with building durable platforms
The semiannual shift removes one constraint. It lets builders build without performing quarterly justification theater. That's incrementally bullish for companies making genuine long-term AI bets and bearish for those using "AI investment" as an excuse for mediocre execution.
The Implication
If this proposal passes, watch which companies immediately switch to semiannual reporting. That's your signal for who's actually building for the long term versus who was using quarterly deadlines as forcing functions for accountability.
For investors, this means adjusting how you track AI companies. The old playbook — quarter-over-quarter revenue growth, margin expansion every 90 days — doesn't map to businesses building foundational agent infrastructure. You'll need to evaluate execution on longer horizons with different metrics: model improvement velocity, agent deployment breadth, infrastructure utilization rates.