Infra

When people think about startups working with the government, the phrase “black box” often comes up. But what if that box is finally being pried open?

Convex cofounder and Chief Scientist Sujay Jayakar and a16z General Partner Martin Casado discuss the real challenges of autonomous software development and why robust evals — and guardrails like type safety — matter more than ever.

Resend's Zeno Rocha and a16z's Yoko Li discuss how AI agents and MCP are reshaping the developer experience for email creation and programming overall.

The global race for AI leadership is no longer just about companies—it’s about nations. AI isn’t just computing infrastructure; it’s cultural infrastructure, economic strategy, and national security all rolled into one.

Sesame's Ankit Kumar and a16z's Anjney Midha do a deep dive into the research and engineering behind Sesame's cutting-edge AI model for conversations.

AI is reshaping web development. At the center is agent experience (AX), a new paradigm where AI agents open new capabilities in web design and experience.

a16z partners Guido Appenzeller and Marco Mascorro demystify how DeepSeek works and explain what the era of reasoning models means for AI.

In this episode of AI + a16z, a trio of security experts join a16z partner Joel de la Garza to discuss the security implications of the DeepSeek reasoning model that made waves recently. It's three separate discussions,...

Critical materials like copper, lithium, and gallium have been mined for decades, but their role in core technologies, geopolitics, and the energy transition have come to a height in recent years.

Fivetran cofounder and CEO George Fraser and a16z partner Guido Appenzeller discuss how LLMs fit into the data management picture within large enterprises.

Two words have caught the Internet by storm. DeepSeek.

Healthcare is a $4 trillion industry, making up nearly a fifth of the U.S. economy—but despite having some of the best doctors and advanced technology, the system often delivers poor outcomes at skyrocketing costs. Why i...

Martin Casado and Alan Nichol discuss how to build reliable AI agents and chatbots — from early NLP approaches to integrating LLMs with business logic.

A 2024 highlight reel from the AI + a16z podcast, featuring insights, advice, and experiences from building AI companies.

The AI race is on, and 2025 could be its most transformative year yet.

Replicate CEO Ben Firshman and a16z partner Matt Bornstein discuss the art of building productss — from PaaS to AI — that appeal to software developers.

Pedro Domingos joins Martin Casado to discuss the state of AI, whether we're really on a path toward AGI, and the value of expressing unpopular opinions.

a16z General Partner Anjney Midha discusses the Nobel Prizes awarded to AI researchers, and how work dating back decades led to today's AI-powered world.

What does Rich Sutton’s "Bitter Lesson" reveal about the decisions Tesla is making in its pursuit of autonomy?

Oxygen is an a16z program that helps AI startup founders navigate their compute challenges and access the GPUs and other resources they require.

Bowen Peng and Jeffrey Quesnelle of Nous Research discuss their mission to accelerate open source AI research, including with a new project called DisTrO.

Ambience cofounder Nikhil Buduma discusses how to build vertical applications with AI models, including in health care, and why tech expertise isn't enough.

MotherDuck CEO Jordan Tigani discusses DuckDB's spiking popularity as the era of big data wanes, as well as the nexus of SQL and LLMs.

Black Forest Labs founders Robin Rombach, Andreas Blattmann, and Patrick Esser discuss their new company building state-of-the-art image and video models.

Vijay Pande walks us through two decades of applying software engineering and AI to biotech and health care — from Folding@Home through AlphaFold and more.

PromptFoo creator Ian Webster discusses the importance of red-teaming for AI safety and security, and of bringing those capabilities to more organizations.

Command Zero CTO Dean de Beer discusses how large language models can help with cybersecurity incident response, and how to build products on LLMs.

Anyscale's Robert Nishihara discusses the challenges of training and running AI models at scale, and how a focus on video data will change generative AI.

Jiaming Song and Anjney Midha discuss Luma's Dream Machine 3D model that shows abilities to reason about the world across a variety of aspects.

Alasdair Monk discusses how generative AI is changing how developers — and the those building for developers — interact with the tools of their trade.

In this AI + a16z episode, Inngest CEO Tony Holdstock-Brown discusses the reality of running AI agents and multistep AI workflows in production.

In this AI + a16z podcast episode, Mohammad Norouzi shares his story of building influential text-to-image models at Google and cofounding Ideogram.

We stitched together two episodes from the a16z Podcast, featuring Anjney Midha interviewing Arthur Mensch (Mistral) and Stefano Ermon (Stanford).

In this episode of the AI + a16z podcast, a16z partners Guido Appenzeller and Matt Bornstein discuss the state of the generative AI market, about 18 months after it really kicked into high gear with the release of ChatGP...

Security-startup founders Dean De Beer (Command Zero), Kevin Tian (Doppel), and Travis McPeak (Resourcely) share their thoughts on generative AI.

In this AI + a16z episode, a16z's Zane Lackey and Joel de la Garza discuss how generative AI and LLMs could effect profound change in cybersecurity.

In this episode of the AI + a16z podcast, Socket's Feross Aboukhadijeh and a16z's Joel de la Garza discuss the open-source software supply chain.

Pinecone Founder and CEO Edo Liberty discusses the promises, challenges, and opportunities for vector databases and retrieval augmented generation (RAG).

Anjney Midha shares his thoughts on how hardware for AI might evolve over the years to come as we place more emphasis on AI inference workloads.

Naveen Rao of Databricks joins a16z's Matt Bornstein and Derrick Harris to discuss where we're at in terms of large language model (LLM) adoption.

This episode of the AI + a16z podcast features a panel discussion from back in February, focused on the state — and future — of open source AI models.

Sesame believes the promise of augmented reality is in the audio, not the display, and it has trained a novel Conversational Speech Model to deliver on this vision.

Fal's inference engine is designed for the complexities of multi-modal AI workloads, focusing on performance, cost, and latency.

What products are builders using to make websites and web apps with AI? A deep dive into how these products work, their limitations, and the features on the horizon. 

Grepr is an observability platform that helps companies slash their spend by using AI to analyze usage patterns and route only the data that truly matters.

There actually is a reasonable policy position on AI regulation: focus on marginal risk and apply our regulatory energy there.

Stainless is revolutionizing API development and SDK generation for the AI era, via its unique approach to handling custom code and API best practices.

WaveForms AI is training an end-to-end audio language model with the goal of solving the speech Turing test — an AI that feels like talking to a human.

An external "AI brain." Big swings in biopharma. Infinite games. A nuclear resurgence. "Faceless" creators. Google search challengers. Battlefield AI. We asked 50 a16z partners to preview one big idea that will spur innovation in 2025.

Resend is transforming how businesses communicate with their customers, and setting new standards for the developer experience around email sending. 

For LLM of equivalent performance, the inference cost is decreasing by 10x every year. What cost $60/million tokens in 2021 costs $.06/million tokens today.

Braintrust is a devtool platform for any product effort — from simple apps to sophisticated products — based around large language models.

Fei-Fei Li and are team of experts at World Labs are building a spatial intelligence model that can generate 3D worlds that users can interact with.

Mintlify lets companies publish quality software documentation that aligns with their brands, including guides, API reference, and examples.

Pylon provides a single view of customer issues, wherever they are happening — Slack, Teams, chat widget, ticket forms, or emails.

Cursor is a fork of VS Code that’s heavily customized for AI-assisted programming, with loads of features designed to integrate AI into developer workflows.

We're excited to announce our investment in Black Forest Labs (BFL), which is building the world’s best open visual models for developers.

Martin Casado and Ion Stoica argue that open-source models will power innovation without compromising security.

Senate Bill 1047 is designed to apply to models trained above certain compute and cost thresholds. It also holds developers legally liable for the downstream use or modification of their models.

The stakes are high. The opportunities are profound. From the creation of new medicines to bolstering national defense, this is our vision for the AI-enabled future.

Mature code-generation technology, coupled with advanced generative AI image models, has shortened the journey from idea to fully operational application.

Jennifer Li is being promoted to General Partner at a16z. She will continue to invest broadly within the enterprise space, focusing on infra.

a16z's Martin Casado lays out the case for AI as a driving force behind incredible advancements in tech, creativity, and the human experience.

Senator Todd Young speaks about the importance of open innovation and US leadership in AI, and why we need to support AI research.

a16z announces its second batch of a16z Open Source AI Grant recipients. This cohort focuses mainly on LLMs and visual AI models.

No one knows how generative AI will play out from a product perspective. The speakers at our Connect/Enterprise event shared their thoughts and experiences.

a16z announces its Open Source AI Grant program, which will support a small group of open source developers through grant funding.

With generative AI, we’re already seeing use cases with orders-of-magnitude improvement in time, cost, and performance over previous AI waves.

Chatbots have been around for decades, but this time is different. Today’s bots are making inroads into our social lives.

A very simple “getting started with AI” template for those who want to play with core technologies, but not have to think too much about tooling. 

A reference architecture for the LLM app stack. It shows the most common systems, tools, and design patterns used by AI startups and tech companies.

A curated list of resources we’ve relied on to get smarter about modern AI, including generative AI, LLMs, and transformer models.

Composable customer data platforms are taking advantage of a shift in data infrastructure and embracing a “warehouse-first” architecture.

Highlights from the a16z Data and AI Forum, featuring founders building products across the spectrum of data and AI use cases.

The generative AI boom is compute-bound and, as a result, a predominant factor driving the industry is simply the cost of training and inference. 

Transactional databases have long been the most critical component of application design. Why? Because a steadfast database is generally the ultimate enforcement point for correctness in a messy, distributed world. Witho...

We're starting to see the very early stages of a tech stack emerge in generative artificial intelligence (AI). Hundreds of new startups are rushing into the market to develop foundation models, build AI-native apps, and...

Understanding how to build a company in the face of a new, immature, or non-existent market is a topic startups should obsess about.

Why AI models will replace artists long before they'll replace programmers

Vertical clouds are on the rise, as traditional clouds give way to specialization.

The pressure the cloud puts on margins can start to outweigh the benefits you scale and growth slows. Understand how much market cap is being suppressed by the cloud to help inform the decision-making framework on managing infrastructure as companies scale.

New technologies enable activities that fall into one of two categories: 1) doing things you could already do but can now do better because they are faster, cheaper, easier, higher quality, etc. 2) doing brand new things...

To help data teams stay on top of the changes happening in the industry, this article reviews an updated set of data infrastructure architectures.

0/ Is there an Enterprise Margin Crisis? It's not uncommon to see software startups with surprisingly low margins (30-40%). I believe there is a broader trend going on here, which I explore in this thread.

Editor’s note: It’s Summit week at a16z, so each day we’re sharing some of our favorite talks from the last few years. a16z Summit is an annual, invite-only event bringing together thinkers, builders, and innovators to e...

Attackers look for the path of least resistance. Recently, that path has shifted from enterprise networks to you and your devices for two reasons. First, as we have built more secure software and systems, it has made it...

The recent Capital One data hack and breach that compromised sensitive information for 106 million people, including 140,000 Social Security numbers and 80,000 bank account numbers, was executed by a single hacker who ex...

watch time: 18 minutes

[vimeo 179360082 w=720 h=404]

BY BENEDICT EVANS

[vimeo 88682226 w=720 h=404]

There is a perfect storm of three distinct disruptive forces that has the potential to topple nearly every major enterprise software incumbent. And the traditional approach of dealing with technology shifts – through acq...

The Internet dramatically reduced the costs of a number of creative activities. For example, publishing written work used to require a large upfront investment. As a result, publishing was controlled by corporations who...