When you buy through affiliate links in our content, we may earn a commission at no extra cost to you. Learn 
Bankrupt cryptocurrency exchange FTX has filed a motion with the United States Bankruptcy Court to seek approval for the sale of its entire stake in the artificial intelligence (AI) firm Anthropic.
According to court documents submitted on February 3, FTX aims to sell Anthropic Series B preferred stock, along with any associated rights or interests, which is owned by its sister company, Alameda Research.
Former FTX CEO Sam Bankman-Fried had invested approximately $530 million in Anthropic in April 2022, months before FTX faced its collapse in November of the same year.
Notably, the investment in the AI startup was sourced from customer deposits held on FTX, as revealed during Bankman-Fried’s legal trial in October 2023.
Key Takeaways
- FTX plans to sell its 7.84% stake in AI startup Anthropic, which could be worth approximately $1.4 billion.
- FTX’s former CEO Sam Bankman-Fried invested $530 million in Anthropic in April 2022.
- The move is part of the exchange’s broader restructuring efforts to recover funds and fully reimburse its customers.
- Anthropic, led by former OpenAI researchers, recently unveiled Claude 2.1.
FTX’s Anthropic Stake Valued at $1.4 Billion
Alameda Research initially held around 13.56% of Anthropic after the completion of the Series B funding round in April 2022. However, subsequent funding rounds diluted Alameda’s stake to 7.84% as of January.
With Anthropic valued at $18 billion in December 2023, Alameda’s stake is estimated to be worth around $1.4 billion.
FTX is also seeking to expedite the review timeline for its sale motion, aiming to reach a resolution during the bankruptcy court’s upcoming meeting on February 22.
“The flexibility to adjust the sale timeline will help facilitate such cooperation, including by allowing the Debtors to capture excess demand for Anthropic’s equity securities channeled from any of Anthropic’s financing rounds,” the filing said.
“Further, given the significant number and value of Anthropic Shares held by the Debtors, the flexibility to sell portions of Anthropic Shares at different times will help the Debtors monetize their interest.”
FTX Seeks to Sell Claim Against Genesis
In addition to the sale of its stake in Anthropic, FTX has also filed a motion in a Delaware court to sell its $175 million claim against the bankrupt digital financial services firm Genesis Global Capital.
The claim, initially made by Alameda Research, FTX’s associated hedge fund, may be sold either entirely or in parts to take advantage of favorable market conditions.
The sale price must be at least 95% of the highest price quoted by leading market-makers for general unsecured claims of Genesis Global Capital within three days of the sale date.
FTX had initially sought to recover $3.9 billion from Genesis in May 2023 but later negotiated a reduced claim of $175 million in August 2023, which was approved by the court in October. The decision to reduce the claim was driven by the unpredictability of potential recoveries and the desire to avoid protracted and costly litigation.
Anthropic Makes Waves in the AI Industry
Anthropic has been making strides in the AI industry with ambitious plans and strategic collaborations. Founded by former OpenAI researchers, including siblings Dario and Daniela Amodei, Anthropic is gaining traction with its Claude AI models, competing directly with OpenAI’s GPT series.
The company has forecasted a substantial increase in its revenue, projecting more than $850 million in annualized revenue by the end of 2024.
In pursuit of its goals, Anthropic has formed strategic partnerships with major tech companies. For one, it has deepened its collaboration with Amazon Web Services (AWS), selecting AWS as its primary cloud provider.
More recently, Anthropic unveiled Claude 2.1, an advanced version of its AI model designed to enhance applications. Key features of Claude 2.1 include an expansion of the context window to 200,000 tokens, enabling the processing of documents up to 150,000 words.
This allows for robust summarization, Q&A, and analysis of lengthy and complex inputs, opening new possibilities for handling extensive documents like technical manuals or literary works.
Claude 2.1 (200K Tokens) – Pressure Testing Long Context Recall
We all love increasing context lengths – but what's performance like?
Anthropic reached out with early access to Claude 2.1 so I repeated the “needle in a haystack” analysis I did on GPT-4
Here's what I found:… pic.twitter.com/B36KnjtJmE
— Greg Kamradt (@GregKamradt) November 21, 2023
Moreover, Claude 2.1 boasts a reduction in hallucination rates by 50%, enhancing its reliability and accuracy. The improvement in honesty is expected to help enterprises deploy AI applications with greater trust, addressing concrete business problems more effectively.
The model also introduces tool use and function calling in its beta stage, offering greater flexibility and extended capabilities for users to integrate Claude with existing processes and APIs.
