TEXXR's knowledge graph tracks how entities are covered — not what they say about themselves, but how the world's tech writers describe their actions. Each quarter, it maps the distribution of predicates: what percentage of an entity's edges are about launches, financial moves, partnerships, competition, controversy. In Q1 2026, something happened for the first time. OpenAI and Anthropic — two companies founded on opposite principles — arrived at identical profiles. Both at 23% competitor. Both at 21% financial. Both with SpaceX in their entity networks. Both with "partnership" in decline. Two companies designed to be different, and the data can no longer distinguish them.
The Split
In 2021, Dario Amodei — OpenAI's VP of Research — left to found Anthropic. The founding premise was explicit: OpenAI was commercializing too fast. Anthropic would be the safety-first alternative, a public benefit corporation that put responsible scaling ahead of revenue growth. It published a Responsible Scaling Policy. It wrote a constitution for its AI. Its CEO published an essay called "Machines of Loving Grace" about how AI could help humanity rather than harm it.
OpenAI, meanwhile, moved in the other direction. Sam Altman told staff the company would become a for-profit. It completed a recapitalization that paved the way for an IPO. It signed a partnership with Microsoft worth billions, then began building an ad business inside ChatGPT.
One company was supposed to be the accelerator. The other, the brake. Their divergence was the founding story of the AI industry.
The Data
The knowledge graph tells a different story. Here is what it measured:
| Category | OpenAI Q1 2024 | Anthropic Q1 2024 | OpenAI Q1 2026 | Anthropic Q1 2026 |
|---|---|---|---|---|
| Competitor | 0% | 0% | 23% | 23% |
| Financial | 14% | 36% | 21% | 21% |
| Launch | 23% | 31% | varies | 36% |
| Controversy | 12% | 20% | 8% | 8% |
| Regulatory | 17% | 0% | 1% | 0% |
| Partnership | 13% | 8% | 6% | 2% |
| Personnel | 8% | 0% | 12% | 9% |
In early 2024, the two companies had distinct profiles. OpenAI was heavy on regulatory scrutiny (17%) and controversy (12%) — the aftermath of the board crisis and the Musk lawsuit. Anthropic was heavy on financial activity (36%) — the massive funding rounds. They were different companies covered in different ways.
By Q1 2026, the profiles converged. "Competitor" — a category that didn't exist for either company in 2024 — is now the dominant non-launch category for both, at exactly 23%. Financial coverage converged to 21% each. Controversy dropped for both. Partnership declined for both.
Two companies that were founded to be different are now described in identical terms. The knowledge graph doesn't have opinions. It has measurements.
The Regime Break
Anthropic had maintained a stable identity in the knowledge graph for six consecutive quarters — from Q3 2024 through Q4 2025. Its profile was consistent: launch-driven, financially active, with moderate regulatory and controversy signals. A safety lab that shipped products and raised money.
In Q1 2026, that regime broke. "Competitor" surged 23 percentage points. "Partnership" fell 13 points. "Regulatory" — the category that once defined the policy-forward company — fell to near zero.
OpenAI experienced its own shift. Microsoft, which had appeared in OpenAI's top entity network every single quarter from 2024 through 2025, disappeared in Q1 2026 — replaced by Anthropic and SpaceX. The defining partnership of the AI era stopped being one of OpenAI's defining relationships.
What the Market Built
The convergence wasn't an accident. It was structural. The forces that shaped both companies into the same profile were the same forces:
Enterprise revenue. Anthropic's revenue hit a $4 billion annual pace by mid-2025, then raised internal forecasts to $18 billion for 2026. Corporate AI makes up roughly 80% of Anthropic's business. Apple's internal development "runs on Anthropic at this point." The safety lab became an enterprise vendor, and enterprise vendors compete.
The IPO path. In December 2025, Anthropic tapped Wilson Sonsini to begin IPO work. By January 2026, sources said SpaceX, OpenAI, and Anthropic could all go public in 2026. This week, Anthropic executives discussed an IPO as soon as Q4. Public markets demand growth narratives, competitive positioning, and financial metrics — the same categories that now dominate both companies' edge profiles.
Defense procurement. The DOD designated Anthropic a supply chain risk for reserving the right to refuse military orders — then used Claude in combat anyway. Dario Amodei called OpenAI's DOD deal "safety theater." But the Pentagon's action didn't differentiate the companies — it pulled both into the same arena. Whether you embrace or resist military AI, the defense conversation shapes your coverage profile.
Talent competition. OpenAI co-founder John Schulman departed to join Anthropic to focus on AI alignment. Amodei says Anthropic has a better retention rate. The personnel category rose for both companies — from near zero to 9-12%. They're competing for the same people.
The Model Announcement
Today, Anthropic announced it's testing a model that represents a "step change" in performance. The language is indistinguishable from how OpenAI announces its models. "Step change" is not safety vocabulary. It's competitive vocabulary — the language of a company positioning against peers.
This is the same week Anthropic discussed an IPO timeline, the same quarter the knowledge graph registered the largest identity break in the company's recorded history. The safety lab didn't abandon its principles. It added competitive ones. And in the data, those competitive signals are now louder than any other category.
The Irony
Dario Amodei left OpenAI because he believed the company was moving too fast, prioritizing growth over safety. He founded Anthropic to prove a different path was possible — that you could build frontier AI responsibly, with published principles and external accountability.
Four years later, the knowledge graph classifies both companies identically. Not because Anthropic abandoned its principles — it still publishes scaling policies, still writes constitutional guidelines, still employs alignment researchers. But the market doesn't measure principles. It measures actions. And the actions of both companies — raising capital, shipping models, competing for enterprise contracts, preparing for public markets — produce the same signature in the data.
The founding article's headline contains the phrase "to explore AI safety." The Q1 2026 edge profile contains no safety category at all. The principle still exists. The data just can't see it anymore — not because safety doesn't matter, but because competition, finance, and product launches have grown loud enough to drown it out.
The convergence is the story of every industry that starts with ideological differentiation and ends with structural similarity. The market doesn't care why you were founded. It cares what you ship, who you compete with, and how much you're worth. By those measures — the only measures the knowledge graph can see — the accelerator and the brake are now the same machine.