Anthropic’s S-1 reveals two areas where OpenAI could have the upper hand, experts say

Anthropic reported first-half 2026 revenue of $16.3 billion, up 1,300% year over year, but its implied Q3 revenue of roughly $16.9 billion represents 47% sequential growth, down from 132% and 144% in the prior two quarters. Anthropic warns that its growth may not be sustainable, while analysts expect OpenAI’s eventual prospectus to show accelerating revenue and see its consumer brand, advertising and potential hardware as advantages. Price cuts and large compute commitments raise questions about margins and returns; OpenAI’s IPO timing remains speculative, with a filing viewed as likely sometime in 2027.

Analysis

The key underwriting question is whether rapid usage converts into durable contribution profit, not which company posts the faster headline growth. Price cuts may accelerate adoption while lowering revenue per unit of compute; large capacity commitments can then turn demand volatility into operating leverage in the wrong direction. For Anthropic, the combination of discounting and long-dated compute deals raises the risk that reported revenue growth masks weaker incremental economics. Verify gross margin after inference costs, discount-adjusted retention, customer concentration, and the terms and utilization requirements of compute commitments before treating the S-1 as evidence of a scalable model.

OpenAI’s consumer reach and potential advertising channel are not automatically a margin advantage: ad load could weaken trust or engagement, while consumer usage can be compute-intensive. Conversely, Anthropic’s enterprise focus may support deeper workflows and contracted demand, even without mass-market distribution. The competitive outcome may therefore be segmented rather than winner-take-all. A future OpenAI prospectus could reset relative valuations, but leaks and executive comments are not a substitute for audited unit economics.

Near term, IPO appetite and reported growth may dominate pricing; over 1–3 months, filing detail and allocation/valuation terms matter more; over 6–18 months, inference costs, utilization, and customer retention decide whether growth creates equity value. A consumer hardware launch could eventually pressure Apple’s assistant positioning, but product details and adoption are unproven, so this is a watch item, not an Apple earnings thesis. Contrarian risk: investors may overvalue consumer distribution while underpricing compute intensity—and may also overreact to slowing percentage growth while absolute demand continues to expand.

Key Decisions for Investors

• Do not chase ANT solely on revenue growth. If evaluating an IPO allocation, condition participation on valuation and disclosure of inference-adjusted gross margins, compute obligations, and customer concentration; these inputs are missing here.
• Use the next 1–3 months to monitor Anthropic’s discounting, enterprise retention, and capacity utilization alongside any audited OpenAI disclosure. A widening gap in incremental margins—not top-line growth alone—would support a relative-value position.
• No immediate AAPL trade: treat OpenAI hardware as a 6–18 month competitive watch. Reassess only after product specifications, distribution, user adoption, and evidence of substitution for smartphone-based assistant use emerge.
• Falsify the bearish ANT unit-economics thesis if subsequent filings show improving gross margin after inference costs, strong retention without rising discounts, and compute commitments matched by sustained utilization. Conversely, worsening margins or material guidance reductions would strengthen the risk case.