AI in 2026: 3 Predictions For What’s To Come (a16z Big Ideas)

1. 0:00 – 0:28

   Big Ideas for 2026

   1. ETErik Torenberg0:00

      Welcome to part three of our 2026 Big Idea Series. Oliver Hsu explores how autonomous labs and AI are revolutionizing scientific discovery and changing how we conduct research and accelerate breakthroughs. Bryan Kim reveals how AI is evolving beyond mere productivity tools to become the connective tissue in consumer applications, transforming how we interact and engage. And David Haber discusses how AI is reinforcing business models, creating compounding advantages that separate leaders from followers.

2. 0:28 – 3:55

   Autonomous Labs and AI in Scientific Discovery

   1. OHOliver Hsu0:28

      My name is Oliver Hsu. I'm a partner on the American Dynamism team here at a16z, and my big idea is that advances in AI reasoning capabilities and in robot learning will help accelerate scientific progress by moving us closer towards autonomous labs. Laboratory automation is something that's existed for a long time. Like, that is not new, this idea of having robots that you can pre-program to assist in some of the motions involved in a lab. What is new and what is emerging right now is the combination of reasoning capabilities, um, and experiment planning and, uh, the physical element of lab automation. So what that might look like in the near term is collaboration between a scientist and a system that involves both an AI application and a robot, um, and having that be a much more collaborative process, uh, in the near term in many different kinds of labs and many different kinds of scientific processes, whether that's in the life sciences, in the chemicals industry, in the material science, uh, research sphere, uh, and so on and so forth. One of the things that I think is important in the near term though is around interpretability. So, you know, if you think about AI systems as non-deterministic computers, one of the things that really matters for research is you want to really understand why the system is doing what it's doing, why it's planning on, uh, iterating on an experiment in a given way, why it's planning on doing this particular thing. And I think, uh, you know, systems that are purpose-built for scientific research are probably going to focus a lot on that, on the interpretability, on recording what exactly is, is, uh, is happening throughout each step of the process as it collaborates with a human scientist. I think this concept of fully self-driving science, right? Like a, a closed loop where you have AI that iterates on itself and then carries out an experiment, then continues to, to, to iterate without human intervention, I think this is further out. This is what I would consider the destination, uh, for this idea of, uh, of autonomous science. I think where we are right now is that there's a lot of work being done, uh, to form the foundations of autonomous science. And there, uh, if you consider science as, you know, broadly speaking, some combination of theory, of computation, and of experimentation, there's work being done in the AI ecosystem across areas like mathematical reasoning, uh, physical reasoning, simulation and world models, and robot learning. And all of these things eventually, as these capabilities improve, can be applied to, uh, closing this loop. But progress across all these fields is of course uneven, and you kinda have to wait for the capabilities to get to the point where, um, they're ready to be applied to this close, cl- the, this closed loop. And I think that's the destination, and in the near term, incrementally we'll make progress on, you know, lab automation, on the reasoning pieces of this, um, but, uh, ultimately the final destination I think would be around this idea of a, of, of a self-driving lab or of autonomous

3. 3:55 – 5:08

   Market Dynamics and Early Adopters in Autonomous Science

   1. OHOliver Hsu3:55

      science. Part of this is gonna be driven by the market dynamic, uh, in which the research is conducted. So I think there are certain categories of science where there is just a much more mature, uh, demand side market for the outputs of research. And examples include of course life sciences and pharma, um, the chemicals industry, um, facets of the, uh, material science industry. I think these are areas where there is a ready, uh, and willing buyer for a lot of the outputs of this research, and the, uh, the s- the increase in speed and capability, uh, as well as any cost advantage that might, um, that might accrue, all of these things are gonna matter more to, uh, for markets where there is a well-established buyer of, of, of research output. And so I think the, where you see autonomous labs and autonomous science being adopted first is probably more of a function of the market that, um, that, that, that it's operating in. I think Periodic Labs is a great example of, uh, uh, uh, of a team taking a swing, um, at, uh, at autonomous science. I think, you know, uh, when you look at

4. 5:08 – 6:21

   Public-Private Partnerships Accelerating AI-Driven Science

   1. OHOliver Hsu5:08

      the early stage startup landscape, there's companies like Medra that are focused on the life sciences and, uh, pharma market. There's companies like, uh, ChemFi and Yoneda Labs that are focused on the, uh, uh, on, on the chemistry industry. Um, and then there's, zooming out a bit, there's collaborations between government and industry that are really focused on this intersection of AI and science. Uh, you know, there is the Genesis Mission led by the Department of Energy, uh, that brings together, you know, academia, government, and the national labs, as well as, uh, leading AI companies, uh, to pursue AI-driven science. I think just today, DeepMind announced a partnership with the UK government, um, to collaborate on areas of scientific discovery. So I think there's, you know, there's startups that are working on lab automation, there's startups that are working on building an AI scientist, and that-Work is happening against the backdrop of a broader collaboration between both the public sector and the private sector and academia to really accelerate AI-driven scientific discovery.

   2. BKBryan Kim6:16

      [instrumental music]

5. 6:21 – 7:08

   AI in Consumer Applications: From Productivity to Connectivity

   1. BKBryan Kim6:21

      Hi, I'm Bryan Kim. I'm a partner at a16z's AI Applications investing team. 2026 marks the year where major consumer AI application products shift from productivity, helping you work, to connectivity, helping you stay connected. Instead of helping you just do work, AI allows you to see yourself clearly and help build relationships with people you love. AI has been incredibly useful for productivity, and I think we'll, we'll start seeing AI actually take more mindshare and time from traditional products versus AI productivity tools. There will be folks who use it to really augment and actually get that connection that they feel that they need from others digitally. I think there will be a group of people who really use AI to facilitate their existing relationships in person. We're

6. 7:08 – 7:47

   AI and Human Connection: Startups vs. Incumbents

   1. BKBryan Kim7:08

      all social animals, and I believe AI has a real place in helping us stay connected with others and help us feel like we're seen by others. Can startups compete with the large incumbent platforms? The incumbents have the platform, they have the network. AI brings a net new user interaction that may be difficult to replicate and may not natively live in the platform of the product. And insofar as there are net new user interaction models, insofar as there is net new creative outlets and atomic units that look different from what's available in current platforms, my strong belief is that startups can absolutely

7. 7:47 – 8:39

   AI as a Relationship Facilitator

   1. BKBryan Kim7:47

      win. Increasingly, we're sharing so much more of our inner life with AI. What I get really excited about is people's willingness to share is deepening with AI. What happens when I'm okay with my AI coming to your AI, my guy talking to your guy and say, "Look, have you checked in on him? Do you want to talk about ABC?" I think those would be an opener for net new relationship, net new conversations that we wouldn't have otherwise, and I'm very excited for AI to actually finally help people be seen by others. The mantra in consumer products is, look, always try to actually address the core emotion. The core emotion again here is wanting to be seen, wanting to feel connected to others. And in order for the first step to happen, I think it's, it's helpful for the AI product to be able to understand who you are.

8. 8:39 – 9:31

   Personalization and the Future of Consumer AI

   1. BKBryan Kim8:39

      So then the question is, what would be the best mechanism for a product to understand you quickly without you narrating your life story? Perhaps it's, uh, ingestion of your digital footprint. Perhaps it's ingestion of some of the things that you talked about online or offline. Perhaps it's looking through your photo roll. With artificial intelligence or gen AI, we have a net new wave of companies that really help you do work better, think better, and get information easier. We have been blessed by an incredible revolution in AI today. What I get really excited about is what is the next steps and what can be done. I get very excited to think about the next suite of products that will start addressing and helping people feel like they're being seen by others. [instrumental music]

9. 9:31 – 10:05

   AI Reinforcing Business Models

   1. DHDavid Haber9:31

      Hey, I'm David Haber, general partner here at a16z, and I help co-lead the AI Apps Fund. My big idea for 2026 is looking for companies where AI reinforces the business model. You know, I think there's a lot of narrative around AI helping automate work and reducing cost, but I think in instances where AI is actually reinforcing the business model in driving revenue, there's really no limit to the amount that customers may want to adopt that technology. And so the market pull in examples like that are just, you know, so much stronger than, than those where it's just a cost reduction story. I sit on

10. 10:05 – 11:26

    Case Study: AI in Plaintiff Law and Lending

    1. DHDavid Haber10:05

       the board of a company called Eve, which operates in the plaintiff law space. And what's unique about plaintiff law is that those attorneys don't charge by the hour. They operate on a contingency basis, which, which means that they only get paid if they win. And so again, while AI is helping automate a lot of the drafting and reasoning work that they do, ultimately, it's, it's really about enabling them to take on more clients and make more money. So it doesn't erode, you know, the billable hour. It really reinforces their business model. And as a result, the market pull for Eve's kind of AI workspace has just been tremendous. Another example in our portfolio is a company called Salient, which operates in the, uh, loan servicing space. So they're applying voice agents to... They started in, uh, auto lending, but they've expanded to a whole ecosystem of kind of consumer lending products where, you know, a voice agent can speak in fifty languages fully compliantly, track UDAP, do welcome calls, and payment re-reminders. And obviously, you know, there is a cost reduction story in that, right? It is helping drive efficiencies in many of these bank and non-bank lenders who have large call centers. But I think what's-what they found, which is so remarkable, is that the voice agents are actually driving better collection rates, right? So it's not just a cost reduction story, it's actually delivering, you know, better outcomes, you know, for their end customers. And as a result, um, it's reinforcing, you know, the lender business model.

11. 11:26 – 12:29

    Compounding Advantages and Proprietary Data

    1. DHDavid Haber11:26

       Ultimately, where do the sources of compounding competitive advantage, you know, reside in, in AI applications? And I think Eve is a really, uh, unique kind of example and case study for this. You know, ultimately, the founders of Eve had a vision for, you know, owning the kind of end-to-end workflow from intake, you know, to outcome. And I think, you know, deeply embedder-embedding yourself within your customer, having them, you know, live within the product, you know, every day is a source of defensibility. I think they are also creating a, a really unique data asset, right? Ultimately, by being able to process cases, again, from intake all the way to outcomes, that outcomes data is not public, right? That is not a, a source of information that, you know, model companies and labs can actually train on in, you know, on the public internet. And so, you know, ultimately that, that outcomes data is, is used to better inform smarter intake so that Eve can tell their, their customers, "Look, this case has these characteristics to potentially be worth, you know, fifty thousand dollars. This case is potentially worth five million dollars. Here's how you may want to triage, you know, your labor and your time." And ultimately,

12. 12:29 – 12:55

    Smarter Outcomes and the Future of AI-Driven Platforms

    1. DHDavid Haber12:29

       you know, given this counterparty, you know, what are the characteristics that you may want to put into a demand letter to actually affect better outcomes? And so I think the more cases that Eve's processes, you know, the smarter and more powerful the platform becomes. Again, ultimately reinforcing the business model for their clients because, you know, they only get paid if they win.

    2. BKBryan Kim12:47

       [outro music]

Episode duration: 12:55