Oklo: A $7.5 Billion Bet That Revenue Would Eventually Show Up (It Did. All $1.2 Million of It.)

Let's start with the number that should stop you cold: Oklo trades at roughly 6,170 times sales. Not price-to-earnings — there are no earnings — price-to-sales, and the sales in question are $1.2 million trailing twelve months. That's not a valuation multiple, that's a rounding error wearing a market cap.

I say this as someone who actually likes the story. Small modular reactors, AI data centers begging for baseload power, a balance sheet that could survive a nuclear winter (pun very much intended) — $2.47 billion in cash against $3.96 million in debt. This is a company that will not run out of money before it runs out of patience from its shareholders. The problem is those are two very different clocks.

The chart tells you what the story doesn't

Back in January, OKLO was being talked about north of $105 as the "AI + Nuclear" trade caught fire — Reddit was calling it validated, momentum was doing what momentum does. By August, the stock had round-tripped down to the low-$40s, and by late August it had actually printed a 52-week low of $36.61, an 80% retracement from its all-time high of $193.84. That is not a pullback. That is a parabola getting its lunch money taken.

There was a piece of chart-watcher optimism floating around in early August calling for an ascending-triangle breakout above $42.35–$43.20 on the way to $44. Cute setup, wrong outcome — the stock proceeded to break down through support instead, hitting that $36.61 low a few weeks later. It has since bounced back to around $40, which sounds encouraging until you note the stock is still sitting 7% below its 20-day average, 12% below its 50-day, and a brutal 40% below its 200-day. Lower highs, lower lows, no basing pattern. This is a knife that hasn't finished falling; it's just resting on the way down.

What the fundamentals actually say

Q2 did produce a milestone worth acknowledging: Oklo's first-ever revenue, $1.21 million at a healthy 40% gross margin. Genuinely nice. Also genuinely irrelevant against a $73.2 million operating loss and $48.5 million net loss for the same quarter. More concerning is the trajectory of the cash burn — free cash flow went from -$23 million in Q3 2025, to -$50.7 million in Q1 2026, to -$141.7 million in Q2 2026, as capex on the Aurora Powerhouse ramped from $5 million to $94 million a quarter. That's the price of building actual reactors instead of PowerPoint slides, and I respect it. But it means the runway, however long, is being consumed faster every quarter, not slower.

Then there's the stuff that doesn't show up in a DCF: roughly $5.6 million in insider sales from the COO and CFO in early August, and institutional ownership sliding 5.7 points down to 47.3%. Executives are allowed to diversify. Institutions are allowed to trim. But when both happen while the stock is already down 44% year-to-date, it's not exactly a vote of confidence from the people with the best seats in the house.

The bull case isn't wrong, it's just early

Full commercial operation is targeted for late 2027 or early 2028, and licensing — not technology — is the thing standing between Oklo and revenue that matters. Wall Street's own price targets span from $51 to $130, which tells you the Street has no idea what this thing is worth either, and that spread alone should make you size any position accordingly. The AI-power-demand thesis is real. The execution timeline is long. Both things are true simultaneously, and the stock price over the last eight months has been the market violently re-learning that distinction.

My take

Don't chase the bounce off $36.61. This is a stock where the fundamentals argue for patience and the chart argues for staying away, and when both agree, listen. A reclaim of $45–$47 on real volume would change my mind about a trend shift; a retest and hold of $36.61 would set up a legitimate tactical long. Until then, Oklo is a story stock trading on a story timeline, and November 10's earnings print is the next date the market will use to decide whether the story is still worth believing.