jakiro
Author:a16z New Media,compile：Block unicorn
Yesterday, we shared the first part of our Big Ideas series, which covers infrastructure, growth, bio + health, and the challenges our Speedrun team members believe startups will face in 2026.
Today, we’re continuing the second part of the series with contributions from American Dynamism (an investment group a16z set up specifically in 2021) and the Apps team.
American Dynamics
David Ulevitch: Building an AI-native industrial foundation
America is rebuilding those components of the economy that truly give the country its strength.Energy, manufacturing, logistics and infrastructure are once again in the spotlight, and the most important shift is the rise of a truly software-first industrial base with artificial intelligence as its native foundation.These companies start with simulation, automated design, and AI-driven operations.They are not modernizing the past but building the future.
This is opening up huge opportunities in areas such as advanced energy systems, heavy-duty robotic manufacturing, next-generation mining, and biological and enzymatic processes that produce the precursor chemicals that various industries rely on.AI can design cleaner reactors, optimize mining, engineer better enzymes, and coordinate clusters of autonomous machines with insights beyond the reach of traditional operators.
The same changes are reshaping the world outside the factory.Autonomous sensors, drones, and modern artificial intelligence models now enable continuous monitoring of ports, railroads, power lines, pipelines, military bases, data centers, and other critical systems that were once too large to fully manage.
The real world requires new software.The founders who build these software will shape America’s prosperity for the next century.
Erin Price-Wright: The resurgence of the American factory
America’s first great century was built on great industrial might, but as we all know, we’ve lost much of it—partly through offshoring, partly through society’s deliberate lack of constructiveness.However, the rusty machinery is turning again and we are witnessing a renaissance of the American factory with software and artificial intelligence at its core.
I believe that by 2026 we will see companies adopting a factory mindset to address challenges in areas such as energy, mining, construction and manufacturing.This means combining artificial intelligence and automation technology with skilled workers to make complex, customized processes run as efficiently as an assembly line.Specifically include:
-
Navigate complex regulatory and permitting processes quickly and iteratively
-
Accelerate design cycles and design for manufacturability from the start
-
Better manage large-scale project coordination
-
Deploy autonomous systems to accelerate tasks that are difficult or dangerous for humans
By applying technology developed by Henry Ford a century ago, planning for scale and repeatability from the start, and incorporating the latest advances in artificial intelligence, we will soon be able to mass-produce nuclear reactors, build housing to meet the nation’s needs, build data centers at breakneck speed, and enter a new golden age of industrial strength.As Elon Musk says, “The factory is the product.”
Zabie Elmgren: The next wave of observability will be physical, not digital
Over the past decade, software observability has transformed the way we monitor digital systems, making codebases and servers transparent through logs, metrics, and traces.The same transformation is about to sweep through the physical world.
As cities across the U.S. deploy more than a billion connected cameras and sensors, physical observability—real-time visibility into how cities, power grids, and other infrastructure are performing—is becoming both urgent and feasible.This new level of perception will also drive the next frontier in robotics and autonomy, where machines rely on a common framework that makes the physical world as observable as code.
Of course, this shift also carries real risks: tools that can detect wildfires or prevent accidents on the job site could also trigger dystopian nightmares.The winners of the next wave will be those companies that earn the public’s trust and build privacy-preserving, interoperable systems that natively support artificial intelligence, thereby increasing transparency in society without compromising social freedoms.Whoever builds this trustworthy framework will be able to define where observability will go in the next decade.
Ryan McEntush: Electronic industrial architecture will change the world
The next industrial revolution won’t just happen in factories, but inside the machines that power them.
Software has revolutionized the way we think, design and communicate.Today, it is changing the way we move, build and produce.Advances in electrification, materials and artificial intelligence are converging to enable software to truly control the physical world.Machines are beginning to be able to sense, learn and act autonomously.
This is the rise of the electronics industrial stack—the integrated technologies that power electric vehicles, drones, data centers, and modern manufacturing.It connects the atoms that drive the world to the bits that control it: from minerals refined into components, energy stored in batteries, electricity controlled by electronics, movement achieved through precision motors, all orchestrated by software.It’s the invisible foundation behind every breakthrough in physical automation; it determines whether software merely summons a cab or actually takes the wheel.
However, the ability to build this stack, from refining critical materials to manufacturing advanced chips, is being lost.If the United States wants to lead the next industrial era, it must build the hardware that will support it.Countries that master the electronics industrial stack will define the future of industrial and military technology.
Software eats the world.Now, it will move the world forward.
Oliver Hsu: Autonomous labs accelerate scientific discovery
With the advancement of model capabilities in multi-modality and the continuous improvement of robot operation capabilities, the team will accelerate autonomous scientific discovery.These parallel technologies will give rise to autonomous laboratories that can achieve a closed loop of scientific discovery—from hypothesis formulation to experimental design and execution, to inference, result analysis, and iteration of future research directions.The teams building these laboratories will be interdisciplinary in nature and will integrate expertise in artificial intelligence, robotics, physics and life sciences, manufacturing, operations and other fields to achieve continuous cross-domain experimentation and discovery through unattended laboratories.
Will Bitsky: The data journey for key industries
In 2025, the zeitgeist of artificial intelligence will be defined by the constraints of computing resources and the construction of data centers.And by 2026, it will be defined by the constraints of data resources and the next frontier in the data journey: our key industries.
Our critical industries remain treasure troves of latent, unstructured data.Every truck dispatch, every meter reading, every maintenance job, every production run, every assembly, and every test run is the material for model training.However, neither data collection, annotation, nor model training are commonly used terms in the industry.
The demand for this type of data is endless.Companies like Scale, Mercor, and Artificial Intelligence Research Labs are working tirelessly to collect process data (not just “what” is done, but “how” it is done).They pay exorbitant amounts for every piece of “sweatshop data.”
Industrial companies with existing physical infrastructure and workforce have a comparative advantage in data collection and will begin to exploit this advantage.Their operations generate vast amounts of data that can be captured at almost zero marginal cost and used to train their own models or licensed to third parties.
We should also expect that startups will emerge to help.Startups will provide the orchestration stack: software tools for collection, labeling, and authorization; sensor hardware and software development kits (SDKs); reinforcement learning (RL) environments and training pipelines; and, ultimately, their own smart machines.
Applications (Apps) Team
David Haber: Artificial Intelligence Enhances Business Models
The best AI startups are not just automating tasks; they are amplifying the economic benefits for their customers.For example, in a win-share-based law, a law firm receives revenue only if it wins a lawsuit.Companies like Eve leverage proprietary outcome data to predict case success rates, helping law firms choose more appropriate cases, serve more clients, and improve win rates.
Artificial intelligence alone can enhance business models.It not only reduces costs but also brings in more revenue.By 2026, we will see this logic expand across industries as AI systems will more deeply align with customer incentives and create compounding benefits that traditional software cannot match.
Anish Acharya: ChatGPT will become an artificial intelligence application store
The consumer product cycle requires three elements to be successful: new technologies, new consumer behaviors, and new distribution channels.
Until recently, the AI wave met the first two criteria but lacked new native distribution channels.Most products thrive on existing networks like X or word of mouth.
However, with the release of the OpenAI Apps SDK, Apple’s support for mini programs, and ChatGPT’s launch of group chat capabilities, consumer developers can now directly tap ChatGPT’s 900 million user base and grow with new mini program networks like Wabi.As the final link in the consumer product life cycle, this new distribution channel is expected to kick off a once-in-a-decade consumer technology gold rush in 2026.Ignore it at your own risk.
Olivia Moore: Voice agents are starting to take hold
Over the past 18 months, the idea of artificially intelligent agents handling real-life interactions for businesses has gone from science fiction to reality.Thousands of companies, from SMEs to large enterprises, are using voice AI to schedule appointments, fulfill bookings, conduct surveys, conduct customer information collection, and more.Not only do these agents save businesses costs and generate additional revenue, they also free up employees to do more valuable work—and more interesting work.
But because the field is still in its infancy, many companies are still in the “voice as entry point” phase, offering only one or a few types of calls as a single solution.I’d be excited to see voice assistants expand to handle entire workflows (possibly multimodal) and even manage the full customer relationship cycle.
This will most likely mean that agents will be more deeply integrated into business systems and given the freedom to handle more complex types of interactions.As the underlying models continue to improve—agents can now invoke tools and operate across disparate systems—every company should be deploying voice-led AI products and using them to optimize key aspects of their business.
Marc Andrusko: Proactive apps without prompts are coming
In 2026, mainstream users will say goodbye to prompt boxes.The next generation of AI apps will show no prompts at all—they will observe what you do and proactively provide you with action suggestions for your reference.Your integrated development environment (IDE) will suggest refactorings before you even ask a question.Your customer relationship management system (CRM) automatically generates a follow-up email after you end the call.Your design tools will generate solutions as you work.The chat interface is only an auxiliary tool.Today, AI will become invisible scaffolding throughout every workflow, activated by user intent rather than command.
Angela Strange: Artificial Intelligence will eventually upgrade banking and insurance infrastructure
Many banks and insurance companies have integrated AI capabilities such as document import and AI voice agents on their legacy systems, but AI can only truly transform the financial services industry by rebuilding the infrastructure that supports it.
By 2026, the risk of failing to modernize and fully leverage AI will outweigh the risk of failure, at which point we will see large financial institutions abandoning contracts with legacy vendors in favor of implementing newer, more AI-native alternatives.These companies are breaking free from the shackles of past classifications and becoming platforms that centralize, normalize and enrich underlying data from legacy systems and external sources.
What was the result?
-
Workflows will be significantly simplified and parallelized.No more switching back and forth between different systems and screens.Think about it: you can see and process hundreds of pending tasks in your mortgage origination system (LOS) at once and in parallel, with agents completing even the more tedious parts of them.
-
Classifications that we are familiar with will be combined to form larger classifications.For example, customer KYC, account opening and transaction monitoring data can now be unified in a single risk platform.
-
The winners in these new categories will be 10 times larger than the incumbents: the categories are wider, and the software market is gobbling up the labor force.
The future of financial services is not about applying artificial intelligence to old systems, but about building a new operating system based on artificial intelligence.
Joe Schmidt: Forward-thinking strategies to bring AI to 99% of businesses
Artificial intelligence is the most exciting technological breakthrough in our lifetimes.So far, however, most of the proceeds from new startups have gone to Silicon Valley’s 1% of companies—either companies that are actually based in the Bay Area or part of its vast network.It’s understandable: Entrepreneurs want to sell to companies they know well and have easy access to, whether in person at their offices or through connections with VCs on their boards.
By 2026, this will completely change.Businesses will realize that the vast majority of AI opportunities exist outside of Silicon Valley, and we will see new startups leveraging forward-thinking strategies to uncover more opportunities hidden within large traditional verticals.There are huge opportunities for AI in traditional consulting and service industries, such as systems integrators and implementation companies, as well as in slower-moving industries such as manufacturing.
Seema Amble: AI creates new layers of coordination and new roles in Fortune 500 companies
By 2026, enterprises will further move away from siled AI tools toward multi-agent systems that need to operate like coordinated digital teams.As agents begin to manage complex and interdependent workflows (such as joint planning, analysis, and execution), enterprises need to rethink how work is structured and how context flows between systems.We’re already seeing this shift happening with companies like AskLio and HappyRobot, which deploy agents across entire processes rather than individual tasks.
This shift will be felt most acutely by Fortune 500 companies: they command the largest reservoirs of siled data, institutional knowledge, and operational complexity, much of which resides in the minds of their employees.Transforming this information into a shared foundation for autonomous workers will unlock faster decision-making, shorter cycle times, and end-to-end processes that no longer rely on constant manual micromanagement.
This shift will also force leaders to reimagine roles and software.New functions will emerge, such as AI workflow designers, agency supervisors, and governance leaders responsible for coordinating and reviewing collaborative digital workers.In addition to existing systems of record, enterprises need orchestration systems: new layers to manage multi-agent interactions, determine context, and ensure the reliability of autonomous workflows.Humans will focus on handling edge problems and the most complex situations.The rise of multi-agent systems is more than just another step in the automation journey; it represents a reimagining of how businesses operate, how they make decisions, and ultimately how they create value.
Bryan Kim: Consumer AI shifts from “helping me” to “understanding me”
2026 marks the year when the functionality of mainstream consumer AI products will shift from improving productivity to enhancing human connections.Artificial intelligence is no longer just about helping you get your work done, but about giving you a clearer understanding of yourself and helping you build stronger relationships.
To be clear: this won’t be easy.Many social AI products have been launched but ultimately failed.However, thanks to multimodal context windows and falling inference costs, AI products can now learn from all aspects of your life, not just what you tell the chatbot.Imagine your phone photo album showing real emotional moments, one-on-one messages and group chats changing based on who you’re chatting with, and your daily habits changing under stress.
Once these products actually become available, they will become part of our daily lives.Generally speaking, “get to know me” products have better user retention mechanisms than “help me” products.”Help me” products make money through users’ high willingness to pay for specific tasks and strive to improve user retention.”Follow me” products make money through continuous daily interactions: users are less willing to pay, but have higher user retention rates.
People are constantly exchanging data for value: the question is whether what they get in return is worth it.And the answer will soon be revealed.
Kimberly Tan: New model primitives enable unprecedented companies
By 2026, we will witness the rise of companies that simply could not exist before breakthrough advances in reasoning, multimodality, and computer applications.To date, many industries (such as legal or customer service) have leveraged improved reasoning techniques to enhance existing products.But we are only now starting to see companies whose core product functionality fundamentally relies on these new model primitives.
Advances in reasoning can lead to new capabilities for evaluating complex financial claims or acting on the results of intensive academic or analyst research (e.g., adjudicating billing disputes).Multimodal models make it possible to extract underlying video data from the physical world (e.g., cameras on a manufacturing site).The advent of computers has made it possible to automate large industries whose value had historically been shackled by desktop software, poor APIs, and fragmented workflows.
James da Costa: AI startups achieve scale by selling to other AI startups
We are in the midst of an unprecedented wave of company creation, driven largely by the current AI product cycle.But unlike previous product cycles, incumbents are not sitting on the sidelines; they are also actively adopting AI.So, how do startups win?
One of the most effective and underrated ways for startups to outperform incumbents in distribution channels is to serve them at their inception: greenfield companies that are just getting started (i.e., brand new businesses).If you can attract all new companies and grow with them, as your customers grow, you will become a big company too.Companies like Stripe, Deel, Mercury, Ramp, and others have all followed this strategy.In fact, many of Stripe’s customers didn’t even exist when Stripe was founded.
In 2026, we will see startups that started from scratch achieve scale in many areas of enterprise software.They just need to build better products and go all out to develop new customers who are not yet locked into existing vendors.
