In April 2022, Anthropic closed a $580 million Series B. The lead investor was Sam Bankman-Fried, then CEO of cryptocurrency exchange FTX. Less than eight months later, FTX collapsed into bankruptcy. Bankman-Fried was convicted of fraud in November 2023. The equity stake FTX held in Anthropic was effectively wiped out.
Anthropic survived. It raised billions more from Google, Amazon, and other backers at markedly higher valuations, and launched its Claude AI assistant in March 2023. The company has since grown into a direct competitor to OpenAI.
Anthropic was founded in 2021 by siblings Dario and Daniela Amodei alongside other former OpenAI employees. Its stated mission: build reliable, interpretable, and steerable AI systems. The Series B proceeds were earmarked to accelerate work on constitutional AI and mechanistic interpretability, two technical approaches designed to make large-scale systems safer and more transparent rather than merely more powerful.
The Size of the Round and What It Signaled
A $580 million Series B in early 2022 was large by any yardstick. It was especially striking given the company's stated goal of limiting the risks of the technology it was building. The round placed Anthropic among a handful of frontier labs with the resources to train models at the largest scale. It was a direct bet that safety work required the same capital intensity as capability work, and that investors were willing to fund the constraint side of the equation.
Bankman-Fried's involvement added weight. He was one of the most visible figures in crypto and had publicly discussed AI risk as a priority. His commitment signaled that money from outside the traditional venture ecosystem was flowing into safety work at an early stage. The round also gave Anthropic financial runway to hire talent and secure computing resources during a period when GPU access was becoming a bottleneck for AI labs worldwide.
From OpenAI to Anthropic: The Amodei Split
Anthropic's founding team came directly from OpenAI. Dario Amodei, the CEO, had been a research executive there and previously led its safety and policy efforts. Daniela Amodei, the president, had led policy and communications. They left, along with several other researchers, in 2021. Reporting at the time attributed the departure to disagreements over the pace and direction of OpenAI's commercialization under its Microsoft partnership.
Constitutional AI and Mechanistic Interpretability
The Series B financed two technical pillars. Constitutional AI uses a set of written principles, a constitution, to guide model behavior instead of relying solely on human feedback. The aim is models that reason about the rules they are given and apply them consistently. Mechanistic interpretability tries to understand, down to individual neurons and circuits, how a neural network produces its outputs. Both approaches were meant to make model behavior more predictable and auditable.
Competitive Positioning Against OpenAI and DeepMind
Anthropic's Series B positioned it, on paper, as a direct competitor to both OpenAI and DeepMind. At the time, OpenAI was valued above $20 billion and had already deployed GPT-3 and DALL-E 2. DeepMind, owned by Alphabet, operated with a budget that drew on Alphabet's broader resources. The open question was whether a startup could compete with those organizations while devoting substantial resources to safety studies that might slow product releases.
Anthropic's answer: safety and capability development are not separable. Building safe systems, the company argued, requires building systems you fully understand, which means training them at scale. The Series B was intended to let Anthropic train its own large models rather than borrow time on someone else's infrastructure. That meant developing a general-purpose assistant, which later became Claude.
The Broader AI Investment Landscape in 2022
Anthropic's Series B arrived during a period of intense capital deployment into AI. OpenAI was raising from Microsoft and others at valuations that would eventually surpass $30 billion. DeepMind continued to operate with large annual budgets from Alphabet. Smaller labs including Cohere, AI21 Labs, and Stability AI had also raised significant rounds. The unifying theme: the cost of training frontier models was climbing, and investors were willing to fund that escalation.
Anthropic's round stood out because the lead investor was not a traditional technology strategist or venture firm. It was a crypto executive. Bankman-Fried's bet was that AI safety would be the defining technical challenge of the decade and that Anthropic had the credibility to address it. FTX's collapse months later removed that particular backer but did not remove the thesis. Subsequent backing from Google and Amazon made clear that the underlying thesis had broad support.
What Happened After the Round Closed
The Series B closed in April 2022. In March 2023, Anthropic launched Claude, a conversational AI assistant it described as safer and more steerable than rivals. The product arrived first via an API; a consumer-facing chatbot followed. By late 2023, the company had raised additional rounds, including a reported investment from Google and later billions from Amazon, at valuations multiple outlets pegged in the tens of billions. (Exact figures vary by source; Anthropic has not disclosed the terms. Check the company's official announcements for confirmed amounts.)
FTX's bankruptcy created a significant complication. Bankman-Fried's stake sat on FTX's balance sheet. During the proceedings, the value of that stake was uncertain, and FTX's creditors pursued various strategies to recover assets. As of January 2025, the precise outcome has not been fully resolved, though FTX did sell a portion of its shares during the bankruptcy process. Anthropic itself was not a party to the FTX collapse and continued operations throughout.
The Double-Edged Nature of the Bankman-Fried Connection
Having Sam Bankman-Fried as the lead investor gave Anthropic a large, fast check at a critical moment. It also tied the company's name to a figure who became one of the most high-profile fraud convicts in technology history. The association has complicated Anthropic's public narrative. The company does not control who its investors are, but the identity of the Series B lead inevitably shaped perceptions of the round.
Anthropic has consistently framed its mission as independent of any single backer. The company's subsequent fundraising from Google and Amazon, and its continued growth, suggest that institutional money has not been deterred by the FTX connection. The Series B round itself now reads as a historical artifact of a specific moment, when both AI and crypto were peaking simultaneously in 2022.
