The Double Dip: How Sam Altman Collected Two Government Contracts in 90 Days

The Setup

Sam Altman runs two companies that matter here. OpenAI — the world's most-valued AI firm, backed by Microsoft, customer to half of Silicon Valley, increasingly a defense contractor. And Oklo Inc. (NYSE: OKLO) — a zero-revenue advanced nuclear startup he founded with Jacob DeWitte and Caroline Cochran, took public via SPAC in 2024, and backed with an estimated 6% ownership stake.

They operate in different industries, before different agencies, regulated under separate statutory frameworks. But in a 90-day window from late February to late May 2026, both companies collected major U.S. government contracts. One from the Department of Defense. One from the Department of Energy. The same administration. The same 90-day window. The same man on both sides of both deals.

And the stock that was supposed to move on the DOE announcement had already moved 59% before the announcement was made.

Part One: The Anthropic Purge

On February 24, 2026, Defense Secretary Pete Hegseth gave Anthropic an ultimatum: remove AI safety guardrails that prevented Claude from being used for autonomous weapons systems and domestic mass surveillance, or face contract termination. Anthropic — maker of Claude, a direct competitor to OpenAI's GPT models — declined.

"Anthropic will continue to provide the Department of War its services for a period of no more than six months to allow for a seamless transition to a better and more patriotic service." — Hegseth, Feb 26, 2026

On February 26, Hegseth designated Anthropic a "supply chain risk to national security" — a formal designation that bars Anthropic from working with the military or any defense contractor. Every federal contract Anthropic held or was competing for was eliminated in a single administrative action. The company said the ban could cost it billions.

On February 27 — the next morning — Sam Altman announced that OpenAI had reached an agreement with the Department of Defense to deploy its models in classified military networks.

Twenty-four hours. The competing company was banned. The beneficiary was already signed.

Altman later admitted the deal "looked opportunistic and sloppy" and said OpenAI "shouldn't have rushed." He renegotiated to include technical safeguards. The contract stood.

The May 1 classified-network AI contract sweep included every major tech competitor except Anthropic. Nvidia — which would partner with Oklo just 22 days later — was on that list. Oracle — whose CEO Safra Catz scores 19/20 on this dashboard — was on that list. The exclusion of Anthropic from the defense AI ecosystem was total and permanent, and the network that filled the vacuum is the same network now accumulating Oklo's nuclear contracts. Suspects Dashboard →

Part Two: The Nuclear Pre-Position

While the Pentagon AI contracts were being finalized, a parallel sequence was unfolding at the Department of Energy. Oklo — Altman's nuclear company — was collecting DOE deals faster than any zero-revenue startup in the advanced nuclear sector.

The pattern is the DOE pre-announcement architecture. LANL and INL are not private partners. They are federal laboratories under DOE authority. When Oklo announced partnerships with LANL on April 23 and with INL on May 12 — both specifically targeting plutonium fuel technology under existing DOE programs — the market read the signal correctly: the formal contract was coming. The 59% surge happened before the announcement because the DOE's own employees had already told the story through partnership press releases.

Part Three: The Insider Exit

While the DOE deals accumulated, Oklo's founders were executing one of the most systematic insider exits in the advanced nuclear sector. The numbers are difficult to reconcile with a company that generates no revenue.

The 10b5-1 plans provide legal cover. A plan adopted on March 31, 2025 predates specific knowledge of any 2026 DOE contract by over a year. The SEC's 2023 rule amendments require a 90-day cooling-off period — both founders cleared that bar. The sales are technically lawful.

But there are two facts that 10b5-1 cannot explain. First: both co-founders adopted identical plans on the exact same day — March 31, 2025. Coordinated same-day plan adoption by the CEO and COO of a zero-revenue company, thirteen months before a sequence of government contracts that drove the stock 59% higher. Second: Caroline Cochran sold approximately $189 million in shares over six months. The company had no revenue. The only asset it had was government contract expectations — and those contracts were being assembled, in significant part, through partnerships with the government's own laboratories.

The 10b5-1 mechanism was designed for exactly this: a pre-arranged schedule that converts anticipated stock appreciation into realized gains, while maintaining plausible legal separation from the specific events that drive the appreciation.

Part Four: The Network

The actors in this story are not strangers to each other or to this dashboard.

This is the architecture of captured government procurement: the competing product is eliminated by executive action, the beneficiary captures the resulting contract within 24 hours, the secondary government agency provides contracts to the same principal's nuclear investment, and the whole sequence is pre-announced — not through leaks, but through partnerships with the government's own laboratories. When the official announcement comes, the stock has already moved. The public reads the announcement. The network read the LANL press release five weeks earlier. Network Map →

Part Five: The Kalshi Signal

Prediction markets have not generated the kind of anomalous spike around Oklo that this dashboard has documented for FRO options before Hormuz and Kalshi crude before the Iran strikes. But there is one market worth watching.

Octagon AI / Kalshi hosts a market on which nuclear power companies will achieve criticality before August 2026. Oklo currently sits at 18% — lower than several competitors. But Oklo's subsidiary Atomic Alchemy is targeting criticality at its Groves facility by July 4, 2026. The Kalshi market closes August 1. If Atomic Alchemy achieves criticality, the probability moves from 18% to 100% — a 5.5× payoff on a binary outcome that Oklo's own press releases suggest is on schedule.

The January 7 DOE agreement was the legal starting gun. The target date is public. Anyone tracking Oklo's DOE portfolio knows the July 4 target. Watch the Kalshi criticality market volume in June 2026. An anomalous accumulation of YES position on Oklo before a criticality announcement would be the same pattern documented five times in the crude oil series. Kalshi Signals →

What This Is

This is not a traditional insider trading case. The Form 4 sales are pre-arranged and technically lawful. There is no evidence of individual stock purchases by Altman or government officials ahead of specific announcements. The investigation here is structural, not transactional.

The structure is this: one official (Hegseth) eliminates a competitor (Anthropic) through an administrative designation, creating a captured market. The primary beneficiary (OpenAI / Altman) captures the resulting Pentagon AI contract within 24 hours. Through a separate but parallel government agency (DOE), the same principal's nuclear investment (Oklo) collects four contracts in five months, pre-announced through partnerships with government-employed scientists at DOE national laboratories. The stock pre-runs 59% before the official announcement. The founders collect $189M before the public reads the press release.

Each piece has a legal explanation. The totality does not.

All facts sourced from SEC Form 4 filings, DOE press releases, Pentagon announcements, and public reporting. This investigation reflects what is documented in the public record. It is not a legal conclusion. The pattern is the evidence; readers draw their own conclusions. Case Files → Watchlist → Donor Records →