Login / Signup

Access the Deck & Memo

Submission complete
Access Deck & Memo
Something went wrong while submitting

Confidential - Do not share without permission

Back

Performance

i

1.4x

-

Anthropic is an AI research company that builds reliable, interpretable, and steerable AI systems. Their first product is Claude, an AI assistant for tasks at any scale. Their research interests span multiple areas including natural language, human feedback, scaling laws, reinforcement learning, code generation, and interpretability.

Memo Highlights

0:00
0:00

Round

Secondary

Investors

Amazon, Google, Spark, Dustin Moskovitz, Eric Schmidt, Jaan Tallinn

Date

23 Dec

Questions

team@joinbeyond.co

Email

Submission received!

We'll email you regarding next steps.
Oops! Something went wrong while submitting the form.

Confidential
Do not share

Deck only available via Desktop

Highlights

0:00
0:00

We're very excited to invest in Anthropic, one of few companies leading the AI race. Access into this company is incredibly hard as sellers are rare, so much so that we've seen large blocks sold at over $70/share and bidders prepared to pay $80-100/share (please review secondary platforms).

Deal

  • Type: Secondary transaction of common shares
  • Price: $55/share (~$34B valuation)
  • Structure: 2nd & 3rd layer options
  • Fees: 0.75% mgmt fee, 10-15% carry
  • Closing: 20th December

Latest update

  • Amazon invested $4B in November (article mentions their new round at $40B valuation target, could be higher than thus).
  • They're outperforming all major players in a number of key criteria and their recent launch of 'computer use' & ability to run & write code showcases how AI will start to expand capabilities and revenue potential.
  • Launches Claude 3.5
  • If they retain their current technical & revenue trajectories ($1B ARR, 1,100% YoY) they could soon catch up with OpenAI's valuation of $157B.

Why we’re investing..

Leading AI Technology:  Anthropic has developed highly capable AI models like Claude that are competitive with industry leaders like OpenAI's GPT-4. Claude 3, Anthropic's latest model, reportedly outperformed GPT-4 on industry benchmarks. Anthropic's focus on multimodal AI with photo and video capabilities also sets it apart

Experienced Team:  Anthropic was founded by former OpenAI employees who are some of the top AI researchers in the world. This deep technical expertise has enabled rapid progress on the AI frontier.

Significant Funding:  Anthropic has raised over $7B in funding since 2021, including major investments from tech giants like Amazon, Google, and Salesforce. Amazon alone has committed $4.5B to Anthropic, making it Amazon's largest venture investment ever. This funding has fueled Anthropic's rapid growth and development.

Valuation Growth:  Anthropic's valuation has skyrocketed from $4.1B in May 2023 to $18.4B as of March 2024, a nearly 5x increase in less than a year. This rapid appreciation reflects Anthropic's strong competitive position and future potential in the booming AI market.

Strategic Partnerships:  Anthropic has formed key partnerships with major tech companies to integrate its AI technology into their products and services. This gives Anthropic broad distribution and exposure to drive adoption of its models.

Potential Growth:  Anthropic believes there will be just a few dominant players in the foundational model space. As Google ($2T), Apple ($2.5T) & Microsoft ($3T) dominated cloud & mobile, one of the few companies leading the next race are OpenAI ($80B) & Anthropic ($19B). Foundational AI has the potential to deliver 10-20x value of cloud & mobile. So even if Anthropic is able to take a fraction of this market, and reach the valuations of the previous dominant players, this still would represent significant upside.

Product

Anthropic's main product is Claude, an AI assistant and language model.

  • Anthropic has developed a "family of foundational AI models" called Claude, which can be used for a variety of applications like conversational tasks, image analysis, coding, and language translation.
  • Claude is available in different versions - Claude, Claude Instant, and Claude Team - with varying levels of performance and capabilities.
  • Anthropic is offering access to Claude through a chat interface at claude.ai as well as through an API for developers and businesses to integrate into their own products and workflows. They have also just launched native app experiences.
  • Anthropic describes Claude as a "next-generation AI assistant" that is more reliable, predictable, and steerable compared to other language models, with a focus on being helpful, honest, and harmless.
  • Anthropic has partnered with companies like Notion, Quora, DuckDuckGo, and Robin AI to integrate Claude into their products and services

Anthropic has identified that enterprises need models that are: capable, safe to all, custom, fast, cost-effective. Have have built Claude 3 to ensure it: 

  • 10x harder to jailbreak than competition
  • Is rapid & effective
  • Can ingest large token counts, 
  • Has embedded safety in entire model - SOC 2 & HIPAA compliant 

Anthropic believes it is imminent that AI Agents will be able to complete complex end to end tasks and aim to spearhead this development.

Traction 

- 2023 revenue finished between $80-100M ARR (with sources stating ~$300M ARR in March 2024)

- 2024 revenue is projected to reach $1B ARR (>1000% YoY growth)

- From formation, it took OpenAI ~8yrs to reach $1B in ARR, Anthropic is on track to reach this within 3 years, which would make it the fastest company to ever hit this milestone (most take 10+ years)

- Anthropic predict a 256x increase in performance by 2030

- Customers include: Slack, Quora, Midjourney, Bridgewater, many more

Distribution

To acquire and grow enterprise customers, which is the biggest opportunity in AI they have 3 key investors:

  • Amazon: Who own >30% of the cloud infrastructure market - Enabling Anthropic to access customers through AWS Bedrock (fully managed service for foundational models)
  • Salesforce: Who own almost 30% of the CRM market and over 150k enterprise customers
  • Google: >91% of the market for search 

Market & Competition 

AI & LLMs are projected to be worth $20T (20x size of cloud/mobile) and there are significant barriers to entry (capital & talent) that only a few have (OpenAI, Google, Anthropic).

Anthropic has several unique aspects that differentiate it from competitors in the AI industry:

  • Commitment to AI Safety and Ethics - Anthropic has an AI safety team and takes measures like red-teaming to ensure its AI systems are reliable and aligned with human value. The company focuses on developing "Constitutional AI" that is helpful, harmless, and honest from the outset.
  • Innovative Training Techniques - Anthropic leverages unique techniques like "Hybrid Training" that combines supervised, unsupervised, and reinforcement learning to create more capable AI models. They also use "Constitutional AI" to train models to be safe and beneficial. 
  • Broad Applicability - Rather than specializing in vertical domains, Anthropic aims to develop AI technology with broad applicability across many fields. Their flagship model Claude can handle a wide range of tasks like text generation, code creation, and intelligent assistance
  • Large Context Window - Anthropic has an extremely large context window of 100,000 tokens per prompt. This makes it well-suited for tasks requiring deep technical understanding and optimized code generation compared to competitors like ChatGPT
  • Commitment to Transparency - Anthropic has a strong commitment to transparency in its AI development, in contrast to the more closed-door approach of some rivals. The company has filed over 25 patents around methods for robust and safe AI systems

Anthropic is built to be at parity with GPT4 but achieved this with just 20% of the capital raised. Claude 3 is outperforming ChatGPT4 in every spec and has a superior attack success rate (GPT4: 47%, Palm 2: 66%, Claude 2: 2%). It is also 20% cheaper than GPT4 turbo.

Intelligence performance

Vision Capabilities

Fewer refusals

Accuracy

Team

Founded by Dario Amodei, former Vice President of Research at OpenAI, and his sister Daniela Amodei, former Vice President of Safety and Policy at OpenAI. The other co-founders are Jared Kaplan, Benjamin Mann, Jack Clark, Sam McCandlish, and Tom Brown, all of whom were also former employees of OpenAI.

Investors 

Press:

Memo

We're very excited to invest in Anthropic, one of few companies leading the AI race. Access into this company is incredibly hard as sellers are rare, so much so that we've seen large blocks sold at over $70/share and bidders prepared to pay $80-100/share (please review secondary platforms).

Deal

  • Type: Secondary transaction of common shares
  • Price: $55/share (~$34B valuation)
  • Structure: 2nd & 3rd layer options
  • Fees: 0.75% mgmt fee, 10-15% carry
  • Closing: 20th December

Latest update

  • Amazon invested $4B in November (article mentions their new round at $40B valuation target, could be higher than thus).
  • They're outperforming all major players in a number of key criteria and their recent launch of 'computer use' & ability to run & write code showcases how AI will start to expand capabilities and revenue potential.
  • Launches Claude 3.5
  • If they retain their current technical & revenue trajectories ($1B ARR, 1,100% YoY) they could soon catch up with OpenAI's valuation of $157B.

Why we’re investing..

Leading AI Technology:  Anthropic has developed highly capable AI models like Claude that are competitive with industry leaders like OpenAI's GPT-4. Claude 3, Anthropic's latest model, reportedly outperformed GPT-4 on industry benchmarks. Anthropic's focus on multimodal AI with photo and video capabilities also sets it apart

Experienced Team:  Anthropic was founded by former OpenAI employees who are some of the top AI researchers in the world. This deep technical expertise has enabled rapid progress on the AI frontier.

Significant Funding:  Anthropic has raised over $7B in funding since 2021, including major investments from tech giants like Amazon, Google, and Salesforce. Amazon alone has committed $4.5B to Anthropic, making it Amazon's largest venture investment ever. This funding has fueled Anthropic's rapid growth and development.

Valuation Growth:  Anthropic's valuation has skyrocketed from $4.1B in May 2023 to $18.4B as of March 2024, a nearly 5x increase in less than a year. This rapid appreciation reflects Anthropic's strong competitive position and future potential in the booming AI market.

Strategic Partnerships:  Anthropic has formed key partnerships with major tech companies to integrate its AI technology into their products and services. This gives Anthropic broad distribution and exposure to drive adoption of its models.

Potential Growth:  Anthropic believes there will be just a few dominant players in the foundational model space. As Google ($2T), Apple ($2.5T) & Microsoft ($3T) dominated cloud & mobile, one of the few companies leading the next race are OpenAI ($80B) & Anthropic ($19B). Foundational AI has the potential to deliver 10-20x value of cloud & mobile. So even if Anthropic is able to take a fraction of this market, and reach the valuations of the previous dominant players, this still would represent significant upside.

Product

Anthropic's main product is Claude, an AI assistant and language model.

  • Anthropic has developed a "family of foundational AI models" called Claude, which can be used for a variety of applications like conversational tasks, image analysis, coding, and language translation.
  • Claude is available in different versions - Claude, Claude Instant, and Claude Team - with varying levels of performance and capabilities.
  • Anthropic is offering access to Claude through a chat interface at claude.ai as well as through an API for developers and businesses to integrate into their own products and workflows. They have also just launched native app experiences.
  • Anthropic describes Claude as a "next-generation AI assistant" that is more reliable, predictable, and steerable compared to other language models, with a focus on being helpful, honest, and harmless.
  • Anthropic has partnered with companies like Notion, Quora, DuckDuckGo, and Robin AI to integrate Claude into their products and services

Anthropic has identified that enterprises need models that are: capable, safe to all, custom, fast, cost-effective. Have have built Claude 3 to ensure it: 

  • 10x harder to jailbreak than competition
  • Is rapid & effective
  • Can ingest large token counts, 
  • Has embedded safety in entire model - SOC 2 & HIPAA compliant 

Anthropic believes it is imminent that AI Agents will be able to complete complex end to end tasks and aim to spearhead this development.

Traction 

- 2023 revenue finished between $80-100M ARR (with sources stating ~$300M ARR in March 2024)

- 2024 revenue is projected to reach $1B ARR (>1000% YoY growth)

- From formation, it took OpenAI ~8yrs to reach $1B in ARR, Anthropic is on track to reach this within 3 years, which would make it the fastest company to ever hit this milestone (most take 10+ years)

- Anthropic predict a 256x increase in performance by 2030

- Customers include: Slack, Quora, Midjourney, Bridgewater, many more

Distribution

To acquire and grow enterprise customers, which is the biggest opportunity in AI they have 3 key investors:

  • Amazon: Who own >30% of the cloud infrastructure market - Enabling Anthropic to access customers through AWS Bedrock (fully managed service for foundational models)
  • Salesforce: Who own almost 30% of the CRM market and over 150k enterprise customers
  • Google: >91% of the market for search 

Market & Competition 

AI & LLMs are projected to be worth $20T (20x size of cloud/mobile) and there are significant barriers to entry (capital & talent) that only a few have (OpenAI, Google, Anthropic).

Anthropic has several unique aspects that differentiate it from competitors in the AI industry:

  • Commitment to AI Safety and Ethics - Anthropic has an AI safety team and takes measures like red-teaming to ensure its AI systems are reliable and aligned with human value. The company focuses on developing "Constitutional AI" that is helpful, harmless, and honest from the outset.
  • Innovative Training Techniques - Anthropic leverages unique techniques like "Hybrid Training" that combines supervised, unsupervised, and reinforcement learning to create more capable AI models. They also use "Constitutional AI" to train models to be safe and beneficial. 
  • Broad Applicability - Rather than specializing in vertical domains, Anthropic aims to develop AI technology with broad applicability across many fields. Their flagship model Claude can handle a wide range of tasks like text generation, code creation, and intelligent assistance
  • Large Context Window - Anthropic has an extremely large context window of 100,000 tokens per prompt. This makes it well-suited for tasks requiring deep technical understanding and optimized code generation compared to competitors like ChatGPT
  • Commitment to Transparency - Anthropic has a strong commitment to transparency in its AI development, in contrast to the more closed-door approach of some rivals. The company has filed over 25 patents around methods for robust and safe AI systems

Anthropic is built to be at parity with GPT4 but achieved this with just 20% of the capital raised. Claude 3 is outperforming ChatGPT4 in every spec and has a superior attack success rate (GPT4: 47%, Palm 2: 66%, Claude 2: 2%). It is also 20% cheaper than GPT4 turbo.

Intelligence performance

Vision Capabilities

Fewer refusals

Accuracy

Team

Founded by Dario Amodei, former Vice President of Research at OpenAI, and his sister Daniela Amodei, former Vice President of Safety and Policy at OpenAI. The other co-founders are Jared Kaplan, Benjamin Mann, Jack Clark, Sam McCandlish, and Tom Brown, all of whom were also former employees of OpenAI.

Investors 

Press:

No Bounties currently