Issue 36: Anthropic, the rising star

Last October (’23), we bought shares of Anthropic, at a whopping $18B in valuation. At first, I thought it was very expensive, but compared to its peers and its promise, it was easy to be convinced that perhaps it was cheap. In May, Elon Musks xAI closed a new round at $24B post money. Fast forward to this October (’24), Anthropic is in discussion for a new funding round at approximately $40B in valuation. ?Putting it in perspective, OpenAI went from $80B to $157B in 12 months, and Anthropic is the next rising star – see chart below. Unconfirmed until round closes, but that is phenomenal and worthy of a deep dive.

I wanted to cover them in a newsletter, but ran into one from Sabrina Wu and Vivek Ramaswami : Anthropic: Closing the Gap? Thought I’d come back to it in a bit, and here I am. This is written with a different lens - educational vs analytical.

Let us start with a view of the top Gen AI startups as of November 2024:

Anthropic Overview

Overview Anthropic is an AI safety and research company, founded in 2021 by former OpenAI employees including siblings Dario Amodei and Daniela Amodei, along with Jared Kaplan, Benjamin Mann, Jack Clark, Sam McCandlish, and Tom Brown. The company is headquartered in San Francisco, California, and is structured as a Delaware Public Benefit Corporation (PBC), which mandates a balance between private and public interests in its operations

Mission and Vision Anthropic's mission centers on creating AI systems that are both useful and safe. Anthropic differentiates itself from competition, by a) focusing on the enterprise market (while OpenAI has targeted consumers with its?GPT models) and b) pricing their models based on factors that are important to enterprises.

Anthropic introduced "Constitutional AI," a framework that guides behaviors of AI systems, using a predefined set of rules called the "constitution." This allows the AI to evaluate its own outputs and adapt its behavior in accordance with principles.

Products and Technology Anthropic's primary AI model is known as "Claude" and its evolutions (Claude Instant, Claude 2, and Claude 3). Anthropic has an interesting naming convention for its models – Haiku is the smallest, then Sonnet and then Opus. With Claude 3.5 Sonnet they introduced features to understand humor, handle complex workflows and interpret charts and graphs. Following 3.5 Sonnet, there will be Opus 3.5 and Haiku 4.0. With Opus they have capabilities such as interpreting images and complex code generation.

Funding and Financials Anthropic has raised significant capital since its inception, with key investments from tech giants like Amazon and Google. As of 2024, Amazon has invested $4 billion, while Google has contributed $2 billion. Other notable investors include Menlo Ventures and various venture capital firms. The company has raised approximately $8 billion to date and has a valuation of around $18.4 billion and is in current talks to raise the next round.

Competitive Landscape Anthropic competes with several other AI-focused companies, including OpenAI, and others described in the table above. Despite its relatively recent establishment, it has positioned itself as a leader in AI safety and interpretability research. The company's long-term strategy includes expanding its research portfolio and commercial applications through partnerships and new product offerings.

Corporate Structure and Governance Anthropic has established a unique corporate governance model. It is structured as a "Long-Term Benefit Trust," which guides the company’s priorities toward public benefits rather than short-term profitability. The board of trustees includes leaders from prominent organizations in AI safety and public health, such as the Clinton Health Access Initiative and the Alignment Research Center.

Future Outlook Anthropic aims to continue advancing AI safety research and developing models that are interpretable and reliable. With substantial funding and strategic partnerships, the company plans to expand its product offerings, which now include Claude enterprise plans and integrations with cloud platforms like Amazon Web Services (AWS).

Anthropic's rapid valuation acceleration is justified by several key factors:

1. Revenue Growth: Anthropic's annual recurring revenue (ARR) jumped from $22 million in 2023 to an estimated $316 million in 2024, marking a 1,341% growth. This growth, along with strategic pricing for its AI models (ranging from $3 to $75 per million tokens), supports its high valuation.
2. Focus on Safety and Ethical AI: As a Public Benefit Corporation (PBC), Anthropic's unique governance model ensures that safety and ethical considerations are integral to its operations. This positions Anthropic as a responsible AI provider, appealing to stakeholders and customers who prioritize AI ethics and safety.
3. Strategic Partnerships:

Anthropic has several strategic partnerships that are shaping its product roadmap and direction, per their annual investor report. The key partners listed inlcude:

·??????? Amazon Bedrock?

·??????? Quora Poe

·??????? Google

·??????? SAP?

·??????? SK Telecom in South Korea?

Products

Anthropic has developed a suite of products centered around its advanced AI models, collectively known as the Claude family. These models are designed to be reliable, interpretable, and steerable, catering to various applications across industries.

The Claude family comprises several models tailored to different performance and speed requirements:

• Claude 3.5 Sonnet: Introduced in October 2024, Claude 3.5 Sonnet is Anthropic's most advanced model to date. It excels in complex cognitive tasks, including math, coding, and logic, and offers enhanced language understanding. This model is available via Anthropic's API and through platforms like AWS and GCP.
• Claude 3.5 Haiku: A faster variant designed for lightweight tasks, Claude 3.5 Haiku provides rapid responses, making it suitable for applications requiring quick interactions.
• Claude 3.5 Opus: This model offers the highest performance within the Claude family, capable of handling complex analyses, multi-step tasks, and advanced coding requirements.

Anthropic's products are designed to serve businesses of various sizes, from small startups to large enterprises and of course developers. Here's a breakdown of Anthropic's product offerings based on the size of the business they target:

?

1. Products for Small Businesses and Startups

• Claude API (Basic Access): Claude’s basic API access allows small businesses and startups to integrate Anthropic’s AI capabilities into their applications with flexibility in pricing and scalability. This version focuses on quick setup and real-time interaction for tasks such as customer support, content generation, and automation, while being cost-effective for smaller operations. Use Cases: Chatbots for customer support, small-scale content creation, basic automation tasks.
• Claude Lite Target Users: Freelancers, solo entrepreneurs, and small businesses. Description: A lightweight version of the Claude AI models, Claude Lite offers lower compute requirements, making it suitable for smaller businesses with limited resources or less complex needs. Use Cases: Website chatbots, social media automation, small-scale e-commerce support.

2. Products for Medium-Sized Businesses

• Claude API (Pro Access): Medium-sized businesses with higher-scale needs but without enterprise-level infrastructure. Description: Pro API access provides enhanced performance, allowing medium-sized businesses to leverage more sophisticated natural language understanding, faster response times, and slightly extended context windows compared to the basic API. This version is ideal for businesses that need to scale up their AI-driven solutions but don’t require full enterprise features. Use Cases: Advanced customer support, sales automation, personalized content generation for marketing, and internal task automation.
• Custom Claude Models (Mid-Tier) Target Users: Medium-sized businesses that need industry-specific AI models. Description: Anthropic offers fine-tuned versions of their AI models tailored to specific industry needs, such as legal, healthcare, or finance. This service helps medium businesses leverage Claude’s capabilities for niche applications without the complexity of full enterprise solutions. Use Cases: Customized industry-specific chatbots, legal document analysis, healthcare data automation.
• Claude Assistant (Standard) Target Users: Medium-sized businesses using AI for internal productivity. Description: Claude Assistant helps businesses streamline their internal processes, such as managing workflows, automating customer interactions, and generating reports. The standard version provides an affordable solution for businesses needing to enhance team productivity with AI-driven insights and recommendations. Use Cases: Project management assistance, smart document handling, automated data summarization.

3. Products for Large Enterprises

• Claude API (Enterprise Access): Large enterprises and corporations with extensive AI needs. Description: This version of Claude’s API is designed for enterprises that require massive scale, longer context windows, and the ability to handle complex, multi-turn conversations. Enterprise access includes features such as dedicated support, enhanced security, and the ability to handle larger datasets for more sophisticated tasks. Use Cases: Large-scale customer service automation, enterprise knowledge management, advanced AI-driven analytics, and enterprise resource planning (ERP) automation.
• Claude Enterprise Suite Target Users: Large corporations with multiple departments or divisions looking to centralize AI usage. Description: A comprehensive suite of tools designed for large enterprises, the Enterprise Suite integrates Claude’s models into a range of business processes, such as HR, sales, marketing, operations, and customer support. It includes higher-level AI tools for analytics, automation, and decision-making. Use Cases: Full business process automation, enterprise-level AI assistants for decision-making, large-scale document generation, and smart customer relationship management (CRM) systems.
• Custom Claude Models (Enterprise-Grade) Target Users: Large enterprises in industries with specific regulatory and operational needs (e.g., finance, healthcare, legal). Description: Anthropic provides custom AI model development tailored to the exact needs of large enterprises. These models are designed to operate within strict regulatory frameworks, ensuring compliance while delivering sophisticated AI-powered automation and analytics. Use Cases: Legal and regulatory document processing, AI-driven financial analysis, healthcare diagnostics automation, and personalized marketing at scale.
• Claude AI for Enterprises (AI Assistants) Target Users: Large corporations and multinational organizations. Description: Claude AI for Enterprises provides highly scalable AI assistants capable of handling internal and external communications, managing large datasets, and providing actionable insights across departments. This product is often paired with extensive security features and customizable workflows. Use Cases: AI-driven customer service, large-scale automation of internal workflows, AI-driven HR and recruitment tools, enterprise-wide analytics and decision support.

4. Products for Developers and Independent Creators

• Claude Developer Tools Target Users: Independent developers and software engineers building AI tools for business or personal use. This suite provides APIs, SDKs, and development environments to experiment with and build upon Claude’s models. It’s ideal for developers who want to create customized AI-driven applications or services but do not have large business infrastructure. Use Cases: Prototyping AI tools, building chatbots, or creating personalized AI-driven apps for independent projects or small-scale business use.

Strategic Partnerships & Key Customers

Anthropic has primarily been working with organizations focused on responsible AI development, improving large language models, and creating safe AI systems. Some early customers and collaborators have used Anthropic’s AI models to shape the company's roadmap and direction:

1. Quora – Integrated Anthropic’s AI models in Poe, its AI chatbot platform, to enhance question-answering and conversational interactions.
2. DuckDuckGo – Partnered with Anthropic to improve DuckAssist, an AI-powered tool to help answer search queries based on Wikipedia and other sources, enhancing search experiences.
3. Slack – Used Anthropic’s AI models in Slack GPT to improve natural language processing in customer communication and automate workflows.
4. Zoom – Integrated Anthropic’s models into its platform for advanced AI-powered meeting summaries, transcription, and smart suggestions for more effective meeting management.
5. Notion – Utilized Anthropic's AI for content generation and assistance within its platform, especially around smart summaries, document creation, and automating workflows.
6. RStudio (Posit) – Leveraged Anthropic’s AI for enhanced AI-driven code suggestions, supporting data scientists in coding and automation within RStudio.
7. Shopify – Applied Anthropic’s models to power conversational agents for customer support, integrating them into Shopify’s backend to automate routine support inquiries.
8. Jasper – Incorporated Anthropic’s language models to create improved content generation tools, focusing on marketing, copywriting, and social media management.
9. HubSpot – Used Anthropic’s AI to develop smarter chatbots and customer management tools to optimize customer interactions and lead management processes.
10. Midjourney – Though mostly visual AI-based, Midjourney has collaborated with language models like Anthropic’s to explore multimodal AI use cases, enhancing the narrative aspects of generated art.

Competitive landscape?

Anthropic operates in a highly competitive AI landscape, contending with major players such as:

1. OpenAI: A direct competitor with models like GPT-4, OpenAI focuses on commercial AI products and partnerships, including a significant investment from Microsoft.
2. Google DeepMind: Recently merged its AI divisions to create multi-modal models like Gemini, leveraging its vast compute and data resources.
3. Cohere: Competes by emphasizing robust data privacy and serving enterprises.
4. AI21 Labs: Known for its "Jurassic" models, primarily targeting language-based applications.

?

Setbacks and controversies

Anthropic has encountered a few setbacks and controversies since its inception. Here are some notable instances:

1.?Disagreement Leading to Founding

The initial founding of Anthropic itself was rooted in a controversy. The company was established in 2021 by former OpenAI researchers who left due to internal disagreements over OpenAI’s direction, particularly its close partnership with Microsoft. This move reflected a divide within the AI research community over corporate influence and commercialization strategies for AI technology

2.?Data Leak (2024)

In January 2024, Anthropic confirmed that it had suffered a data leak, which exposed sensitive information about its research and internal operations. This incident raised concerns about the security and safety measures the company had in place to protect its proprietary information and the ethical implications of handling such breaches

3.?Regulatory Scrutiny

Anthropic has faced scrutiny from regulators and industry stakeholders regarding the safety and potential risks associated with advanced AI models. As Anthropic’s research focuses on safety and reliability, it has been involved in discussions about the ethical use of AI and the potential for misuse. In particular, Anthropic’s approach to "Constitutional AI" has sparked debate about how ethical principles are encoded and enforced in AI systems, and whether these principles truly align with diverse global perspectives

4.?Funding Controversies

Anthropic has received funding from multiple large corporations, including Amazon and Google. While this has enabled rapid growth and development, it has also led to questions about the influence these corporations may have on the company’s research direction and priorities. Given Anthropic's focus on AI safety, some industry observers have expressed concerns that the company could be pressured to prioritize commercial objectives over ethical considerations

5.?Dispute over AI Safety Standards

As a company deeply involved in AI safety research, Anthropic has occasionally been at odds with other AI labs and research institutions over what constitutes appropriate safety standards. This includes debates around the transparency and interpretability of AI models and how to best implement safeguards against misuse or unintended consequences of AI technologies

These controversies reflect both the challenges and growing pains that Anthropic faces as it navigates its role as a leader in AI safety while balancing commercial and ethical considerations.

Risks to future?

Anthropic faces several risks to its future growth and stability:

1. Regulatory and Compliance Risks: Increased scrutiny from regulatory bodies like the FTC could impact Anthropic’s operations and partnerships with hyperscalers like Amazon and Google.
2. Intense Competition: Competitors such as OpenAI, Google, and emerging startups like Cohere and AI21 Labs pose significant challenges, particularly in technology innovation and market share.
3. Dependence on Hyperscaler Partnerships: Heavy reliance on cloud providers like AWS and Google Cloud means any disruption or strategic shift by these partners could impact Anthropic’s capabilities and offerings.
4. Financial Sustainability: High R&D costs and reliance on external funding may challenge profitability and long-term financial health.
5. AI Safety Concerns: Any safety incidents or ethical missteps could damage Anthropic’s reputation, given its focus on AI safety and ethics.

Anthology Fund

One of the most interesting recent PR announcements by Anthropic is the launch of the?Anthology Fund, a $100 million initiative in partnership with Menlo Ventures. The fund aims to support startups that leverage Anthropic's technology to develop AI solutions across key areas such as healthcare, legal services, education, energy, and scientific research. The initiative provides startups with access to Anthropic's AI models, $25,000 in credits, and venture support, helping drive responsible AI innovation and growth in the AI ecosystem.

Closing

Anthropic stands at the forefront of a transformative era in AI, with its unwavering focus on safety, alignment, and ethical AI development. By addressing critical concerns around the responsible deployment of artificial intelligence, Anthropic has carved out a distinct niche in an increasingly competitive landscape. As the world continues to grapple with the rapid integration of AI into everyday life, Anthropic’s mission provides a crucial counterbalance to the industry’s pace, ensuring that progress does not come at the expense of ethical standards. The company's commitment to transparency and collaboration underscores its potential to be a guiding light for the future of AI—one that is not only powerful but also profoundly human-centric.