The stock peaked at $193.84 not long ago. It’s sitting in the mid-$50s right now. And somehow, both sides of this trade think they’re obviously right.

That’s the part worth slowing down on. Not the Meta deal — everyone’s covered that. Not the short interest number, which is high but not catastrophic. What’s interesting is the complete lack of consensus on a name that has more visible catalysts than almost anything else in the nuclear space. Usually when the facts are this clear, the positioning gets easier. With OKLO, it’s gone the other direction.

Here’s where I’m at on it.

On January 9, 2026, Meta Platforms signed a prepayment agreement with Oklo to support a 1.2 gigawatt power campus on 206 acres of former Department of Energy land in Pike County, Ohio. First phase — roughly 150 megawatts — is targeted by around 2030. Full buildout by 2034. The stock immediately ran 47% in the first week of January, briefly touched $96, then spent the next three months giving most of it back. By late March it was at $45.58. Then it rebounded to $79 by early May. Then Q1 earnings came in and it faded again. That’s four distinct moves in five months, each one violent, each one reversing. The 52-week range is $34.88 to $193.84. Both ends of that range are real prices that real people paid.

Slight tangent, but it matters: what the Meta deal actually represents isn’t revenue. Oklo hasn’t generated meaningful revenue yet and won’t at commercial scale until late 2027 at the earliest — the Aurora Powerhouse currently under construction at Idaho National Laboratory is the first. What the Meta deal represents is proof that the hardest problem in early-stage nuclear is solvable. CEO Jacob DeWitte said publicly that landing the first anchor customer is the biggest hurdle for any SMR developer. They cleared it. Everything after that gets marginally easier to negotiate, easier to finance, easier to permit.

The Q1 numbers are worth knowing. Net loss of $33.1 million, versus $9.8 million a year ago. Operating losses of $51.2 million, partially offset by $21.3 million in interest income from a very large cash pile. EPS of –$0.19, which matched consensus. The company ended the quarter with $2.5 billion in cash and marketable securities — including $1.18 billion raised via ATM during the quarter itself. Full-year operating cash burn guidance is $80 million to $100 million. Capital expenditure guidance is $350 million to $450 million across power, fuel, and isotopes.

Then, on May 13, they filed a fresh $1 billion ATM equity shelf.

That’s the dilution overhang that doesn’t get enough airtime. The company has the cash to operate for years. But every raise between now and first commercial power is a real cost to existing holders, and at the current pace — roughly $1.18 billion raised in a single quarter — that adds up fast. Bears are right to flag it. The question is whether they’re right on timing.

The short interest as of April 15 sat at 28.6 million shares, or roughly 20.32% of available float — up from 24.1 million shares the prior period. By shares outstanding, the figure is closer to 16.45%. Either way, OKLO is the most heavily shorted name in the utilities sector for companies above a $2 billion market cap. The days-to-cover ratio sits near 3, which isn’t extreme. But it doesn’t need to be extreme. With a beta near 2.70 and average daily moves that have routinely run 10% or more on catalyst days, what matters is the mismatch between commitment and patience. Most short positions in a name like this are not built to survive a 40% squeeze. A few of them will be wrong on timing and exit into strength, and that exit amplifies whatever the catalyst already started.

Two catalysts from the last few weeks that the market mostly underreacted to:

On April 23, Oklo announced a three-way collaboration with NVIDIA and Los Alamos National Laboratory. The focus is applying AI-driven simulation and materials science to accelerate nuclear fuel development — specifically validation work for Oklo’s Pluto reactor, which was selected under the DOE’s Reactor Pilot Program. This is not a commercial agreement. It won’t move the revenue line. But it ties Oklo’s core technology into two institutions that are exceptionally well-resourced, and it signals that the DOE and NVIDIA are both comfortable enough with the platform to put their names on it publicly.

Then on May 6, the NRC approved Oklo’s Principal Design Criteria topical report for the Aurora Powerhouse — on an accelerated schedule, completing the review in less than half the traditional timeline. The PDC approval establishes the core safety and performance standards that can be referenced across all future Aurora licensing applications. What that means operationally is that Oklo doesn’t have to re-litigate the same foundational questions with each new project. The licensing process becomes more like a checklist than a ground-up argument. That is a structural efficiency gain that compounds across a fleet-based deployment model.

Most of the coverage treated the May 6 approval as a stock-move story. Nobody spent much time on what it actually does to the timeline math for reactors two through ten.

On the options side, implied volatility is structurally elevated and for obvious reasons — NRC milestone approvals, customer announcements, military deployment decisions, and ATM equity raises all carry event-driven spikes. For traders who want long exposure with defined risk, a bull call spread in August or September expiry in the $70/$85 range captures the upside thesis while containing the premium cost in a high-IV environment. If the bear case is the focus — timeline slippage, escalating costs, or continued dilution pushing the stock back toward its March lows — a bear put spread in the $55/$45 zone reflects that directly. The iron condor case exists too: range-bound action between catalysts, elevated premium, income generation while waiting for the next move to clarify direction. All three are reasonable. None of them is obvious.

Here’s what I keep coming back to. The bear case is real and the numbers support it — pre-revenue, years from deployment, ongoing dilution, and a market cap near $9.5 billion that is pricing in a future that hasn’t happened yet. Every one of those things is true. And yet, Oklo now has a signed prepayment from one of the five largest companies in the world, an NRC licensing process that just got meaningfully faster, $2.5 billion in cash, and a collaboration that puts NVIDIA and Los Alamos in the same room as their engineers. Six months ago, none of that existed.

The thing about threshold moments is you usually only recognize them clearly in hindsight. The Meta deal might be one. Or it might be the high-water mark before a multi-year grind through delays and dilution. The options market is pricing both outcomes as genuinely possible, and right now, I think that’s actually the correct read.