Internal data shows OpenAI's new-user growth under 1%
- OpenAI internal data reveals new-user growth under 1% for ChatGPT, with most signups from paid distribution deals failing to retain active users long-term. - Only 0.7-1% of platform accounts qualify as "real" active users after accounting for high distribution costs, per leaked metrics shared on X. - This exposes freemium model's retention flaws amid rising costs, sparking calls for paywalls and pricing shifts to hit $11B revenue target.
OpenAI's user growth is stalling hard. Internal metrics leaked this week show new-user acquisition under 1% — despite massive spending on partnerships and ads. The numbers expose a brutal truth: freemium signups don't stick, masking deeper retention problems. This could derail the company's path to profitability. ### What do the leaked numbers actually say? The data comes from OpenAI's platform dashboard, shared by X user @whateverthisisx. It tracks "new users" from distribution deals — think Apple, Samsung, Microsoft Copilot bundling ChatGPT access. Growth rate: under 1% month-over-month. Worse, just 0.7-1% of total accounts are "real" active users, defined as those engaging beyond the free tier without heavy subsidies. Distribution costs eat 80-90% of lifetime value from these users. ### Why is "under 1%" such a red flag? OpenAI hit 300 million weekly active users last year — explosive growth fueled by viral word-of-mouth. But that's old news. Now, with market saturation, they rely on paid channels: $100M+ deals with Apple for Siri integration, Samsung for Galaxy AI, enterprise tie-ins. The leak shows these deliver volume, not stickiness — 99% churn fast, leaving slim margins. Analysts peg true organic growth near zero. ### How does freemium play into this? ChatGPT's free tier hooked billions — but retention craters post-hype. Free users poke around, then ghost. Paid distribution amplifies this: bundled access spikes signups, but without habit-forming value, they lapse. Internal signals flag "platform accounts" as mostly bots, one-offs, or subsidized ghosts. Lifetime value per user? Under $1 after costs for 99% of them. This reignites the freemium debate — great for top-of-funnel, terrible for monetization at scale. ### What's the cost breakdown look like? Acquisition ain't cheap. Apple deal alone: $200M upfront plus rev-share. Samsung, similar. Total distribution spend topped $1B last year, per insiders. Yet revenue per free-ish user barely covers it — ChatGPT Plus at $20/month converts <5%. Enterprise APIs carry the load, but consumer side lags. Leaked chart shows user cohorts from Q4 2025 decaying 95% in 90 days. Brutal economics. ### Why has retention been ignored until now? Early days, growth covered sins — 1M users in 5 days, no need for paywalls. Sam Altman chased dominance over dollars, betting network effects. But 2026 realities hit: $11B revenue goal needs 100M+ paying users. Analysts like a16z's Andrew Chen call it "classic freemium trap" — leaky bucket drains cash. Insiders whisper pricing hikes or tiered free limits incoming. ### How does this hit partnerships? Partners like Apple balk at subsidizing churn. Siri + ChatGPT integration? Usage spiked, retention tanked — Apple pays per query now. Samsung faces similar. This squeezes negotiations: OpenAI wants more cash, partners demand better ROI. Could force exclusive paid models or stricter access. Complicates Altman’s “AI for everyone” pitch. ### What changes are coming next? Expect moves fast — Q2 earnings loom. Options: jack Plus to $30/month, cap free queries, push "Pro" tiers. Or double-down on verticals like education, where retention hits 20%. Bottom line — growth story shifts from users to dollars. If they don't fix acquisition math, $100B valuation wobbles. OpenAI knows: scale without profit is just expensive hype. Watch for pricing news by June. ```
