OpenAI on Track to Double Revenue Ahead of IPO

OpenAI is reportedly on track for more than $40B in annualized revenue, about 2x its end-2025 run rate, driven by growth in paying ChatGPT users, enterprise adoption, and Codex usage. The key offset remains the very high cost of computing power required to sustain AI expansion.

Analysis

The market implication is not that a private AI lab is suddenly “profitable,” but that inference demand is still outrunning supply constraints. If revenue is scaling this fast while compute remains the bottleneck, the economic rent should continue to flow to the infrastructure layer first: GPUs, networking, packaging, and hyperscaler cloud consumption. That argues for continued multiple support in names that monetize training/inference intensity, even if end-user application monetization remains uneven.

The less obvious read-through is competitive pressure on horizontal software. Faster adoption of coding and enterprise assistants raises the probability of workflow substitution in developer tools, support automation, and general productivity software; the pain likely shows up first as slower net new ARR growth rather than immediate revenue declines. Public-market beneficiaries and casualties will lag the private company narrative by 1-3 quarters, so the next checkpoints are cloud capex commentary, AI gross margin disclosure, and whether enterprise pilots convert into multi-year seats or remain usage-heavy, low-visibility demand.

The key risk is that headline revenue growth can mask negative unit economics if compute costs scale faster than monetization. If pricing remains usage-based and competition compresses ARPU, the story reverses from “AI demand boom” to “subsidized land grab,” which would cap valuation upside for the whole AI complex. What would falsify the bullish infrastructure view: any sign of capex moderation from hyperscalers, GPU lead times normalizing sharply, or management commentary that AI revenue growth is decelerating while costs stay elevated over the next 1-2 quarters.

Key Decisions for Investors

• Stay long NVDA / AVGO / TSM into the next hyperscaler capex prints; this is a 1-3 month trade on sustained inference demand, with the main risk being a sudden check in cloud spending or evidence of inventory digestion.
• Use MSFT and GOOGL as higher-quality AI monetization proxies versus the private beneficiary set; long both on any post-earnings pullback if they continue to show AI adoption converting into cloud consumption and not just cost inflation.
• Pair trade: long AI infrastructure basket (NVDA, AVGO, ANET) / short a basket of horizontal software exposed to AI feature substitution (e.g., ADBE, SNOW, DDOG) for a 3-6 month relative-value move if enterprise buyers keep reallocating spend toward AI-native workflows.
• Watch for a reversal trigger in hyperscaler margins: if incremental AI investment stops lifting revenue growth or cloud gross margins compress for two consecutive quarters, reduce exposure to the semiconductor leg first.
• Avoid chasing the private-company narrative directly; there is no public equity claim here, so the cleaner expression is through suppliers and platform owners where the risk/reward is observable and the downside is more measurable.