Doximity — the professional network roughly 80% of US physicians carry in their pocket — reported fiscal first-quarter revenue of $156.6 million, up 7% from a year ago. That beat the Street’s roughly $151.7 million and cleared the top of management’s own $151–152 million guide by about three percent. Net income was $24.3 million, a 16% margin. Adjusted EBITDA was $74.8 million, a 48% margin. Adjusted earnings of $0.29 a share missed the $0.30 consensus.
Read that list again and try to find a number worth doubling a company over. Seven percent growth. A penny miss. And the full-year adjusted EBITDA range didn’t rise with revenue — it moved to $309–329 million from the prior $323–335 million. Management raised the revenue outlook to $671–681 million from $664–676 million and guided the September quarter to $170–171 million. More top line, less profit.
The stock closed Thursday at $20.66. It hit $37.49 in after-hours trade, an 81% move, spiked further before the open, and was still up around 89% at midday Friday. CNBC’s headline was simply that the shares had doubled. Whatever the market repriced here, it wasn’t the quarter.
What actually moved it
Two things, both non-financial. Doximity said active users of its AI Scribe — the ambient tool that drafts a clinical note while a doctor talks to a patient — grew tenfold in July versus the same month last year. And the company said Doximity Ask, its HIPAA-compliant clinical assistant, ranked first among US-based models on the real-world clinical sample inside the NOHARM safety benchmark, the slice of that study designed to mirror how physicians actually query these tools at the bedside.
Neither of those is revenue. A tenfold jump in scribe users tells you nothing about what those users pay, and a benchmark result is a claim about quality, not a contract. But together they answer the question that has hung over this company for a year: is the AI spend buying adoption, or just buying compute?
The May mirror
This is the part worth keeping. In May, Doximity fell about 24% in a single session. The stated reason was soft guidance and rising AI compute costs pressuring margins — analysts cut price targets on exactly that. The company was spending heavily on AI and could not yet show what the spending bought.
Today the spending is still heavy. The profit guidance is still lower than it was. The only variable that changed is that management brought a usage number and a benchmark result to the call. Same balance sheet, opposite reaction, roughly a 24% drawdown swapped for something close to a double.
Our take: The market is no longer paying for AI investment. It is paying for AI evidence — and it will discount your entire P&L to get it. Doximity did not earn its way to this price; it produced one metric that could not be faked and one third-party result it did not grade itself. That is now the going rate for a re-rating. The flip side is that a stock that doubles on two proof points can halve on their absence, and a 10x usage number with a flat revenue line is a monetization problem waiting for its own headline.
What to watch
- Whether scribe usage shows up in the revenue line. Ten-fold users against 7% growth is a gap. The September quarter’s $170–171 million guide is the first place it can start closing.
- The EBITDA range. If revenue guidance rises again next quarter and the profit range keeps sliding, the story stops being “investing in AI” and starts being a cost structure.
- Who copies the playbook. Expect a wave of earnings calls this month leading with an adoption multiple and a third-party benchmark. The ones that work will have both; the ones that only have the multiple will get found out.
- Where the price settles. The stock ranged from up 34% to up well over 100% across a single session. Anything that volatile is a narrative being priced in real time, not a valuation.