Our Investment in Anthropic: The Foundation Layer for Generative AI

We stand on the precipice of an extraordinary platform shift catalyzed by groundbreaking advancements in generative artificial intelligence (AI). The generative AI gold rush is already underway—and at a furious pace of adoption that exceeds anything we’ve seen before.

As we experienced with earlier shifts like mobile, cloud, and the web, incredible businesses will be born during this shift, creating massive long-term economic and societal value. We expect this new wave of intelligence to eat into—and expand—the entirety of today’s $140 billion consumer and enterprise software markets.

Products are now able to understand natural language, search through unstructured data, reason with context, and generate text and images that are actionable and instructive. At the foundation of this new software stack is the large language model (LLM), the pioneering building block (AI primitive) every company needs to power its generative AI capabilities.

Today, we’re excited to announce our investment in Anthropic, one of the leading language model providers at the foundation layer, with a clear focus on building aligned AI systems. 

Backing One of the Best Teams in AI

The Anthropic team has been at the forefront of almost all the innovations in generative AI since its earliest days. Anthropic CEO Dario Amodei led OpenAI’s GPT-2 and GPT-3 projects as the former VP of Research, and the founding team at Anthropic included many of the most significant contributors to the two pioneering models across their research, architecture, and engineering. Anthropic was founded as a focused bet on safety research; the team wanted to create a race to the top on safety and build reliable, interpretable, and steerable AI systems.

A Powerful Platform Providing the Foundation for a New Generation of AI Applications

Like other large language models, Claude—Anthropic’s model family—is trained on massive datasets with billions of parameters, allowing it to develop a rich understanding of the relationships between words, concepts, and the nuance of languages.

But the AI model also includes key algorithmic innovations that continue to push the frontier of AI research, including breakthroughs like constitutional AI and reinforcement learning from AI feedback (RLAIF) that seek to make progress on prevalent problems like model hallucinations and toxicity.

Combined with flexible deployment options that include both a plug-and-play API call and the option to host custom models on dedicated compute, it’s easy to understand why Anthropic is already helping to power generative AI solutions of leading enterprise products, including Slack, Notion, and Quora’s Poe chatbot.

Building on a Portfolio of AI Infrastructure and Application Layer Leaders

For Menlo Ventures, our partnership with Anthropic is the latest in years of investing in AI at the infrastructure and application layers.

Menlo’s other investments in the AI infra layer include Pinecone, a vector-embedding database that has emerged as the data plane for the modern AI software stack; TruEra (ML observability); and Clarifai (AI development and operationalization platform). 

While infrastructure creates the plumbing and tooling that make it all possible, the application layer is where use cases come to life. We’re equally active in the massive new wave of GenAI-first applications, including Typeface, an enterprise-grade generative AI content creation platform; Sana Labs, an AI-powered learning platform for corporate L&D departments; and Lindy, an AI assistant for daily tasks ranging from calendar management to email drafting.

Anthropic’s innovative constitutional approach to generative AI will spark a new generation of companies. This rare and massive platform shift will crown new decacorns, displace multibillion-dollar incumbents, and transform industries. Anthropic will continue to play a vital role in driving it forward. Today, we proudly and warmly welcome Dario and the Anthropic team to a growing portfolio of companies propelling the industry forward. If you are building AI, we would love to hear from you.