Back to snapshots

OpenAI

Capital Gravity

Transitioning5.40 GPIOPAI.PVT2026-06-10

A company rarely gets heavy all at once. First the old win keeps getting a vote, the clean plan starts paying rent to yesterday's structure, or the best people work around the system to keep the day moving.

Use this snapshot to spot the pattern early: what still helps the company move, what slows the next move down, and where the pressure may show up before the market gives it a lazy name.

The Read

The habit under the headline.

Capital Gravity

OpenAI embodies the AI industry's central tension: exponential compute requirements colliding with linear revenue growth. The company ships fast, pivots strategically, and dominates consumer AI with Decision Latency (3.5), Error Correction (4.0), and Knowledge Velocity (4.0) all scoring in field-to-transitioning range. But Capital Intensity (8.5) is particle-state physics. Every capability leap requires 10x more compute. Training runs that cost millions in 2023 cost billions in 2026. The $1.4 trillion infrastructure commitment isn't excess. It's table stakes for maintaining leadership. But those commitments create Structural Lock-In (6.0) that constrains strategic options. Oracle needs OpenAI's IPO to service its bonds. CoreWeave's debt is collateralized by future compute demand. Microsoft

Scorecard + Read Checks

The number, then the pressure points.

GPI Score

5.40

State

Transitioning

Decision Latency4OpenAI still ships quickly, but large partner commitments and capital planning add drag around the edges.
Error Correction5The company can adjust product direction quickly, while governance, trust, and safety issues remain heavier correction loops.
Knowledge Location5Research and product judgment are still concentrated in a small technical and executive core, even as distribution broadens.
Structural Lock-In6.5Compute, cloud, chip, and data-center commitments create strategic lock-in around the frontier buildout.
Talent Flow5OpenAI attracts elite talent, but intense competition and scale add retention and coordination pressure.
Capital Intensity9Frontier AI scale requires massive data-center, GPU, cloud, and power commitments before demand fully matures.
Knowledge Velocity4.5Product learning remains fast, but infrastructure physics now shape how quickly capability reaches users.

Numbers Worth Holding

The filing pile gets smaller here.

$500B valuation (October 2025), world's most valuable private company
$20B annual revenue run rate (2025), projected $30B (2026), 236% YoY growth
$1.4T infrastructure liabilities through 2027+ ($300B Oracle, $250B Microsoft, $38B AWS)
$17B projected cash burn (2026), $9B (2025), profitability not expected until 2030
$20B shortfall in 2026 as supplier bills come due, $80B+ deferred commitments maturing
$96B debt carried by partners (Oracle, CoreWeave) financing OpenAI infrastructure
800M weekly active users, 1M+ business customers, 80% of generative AI traffic
3,000+ employees (2025), 20x growth since 2020 from 375

Still Working / Still Stuck

What still has legs. What still drags.

Still working
  • Market leadership
  • Rapid product velocity
  • Strategic pivots
  • Technical talent
  • Revenue growth
  • CEO stability
Still stuck
  • Extreme capital intensity
  • Profitability timeline
  • Supply chain dependency
  • Partner debt burden
  • Enterprise share loss
  • Knowledge politicization

The Line

"You can have fast decision-making or extreme capital efficiency, but not both at frontier AI scale. OpenAI chose frontier. Now physics applies."