I’ve struggled with CRM systems for years, they were either missing basic features or were too expensive or bloated pieces of software. Then 18 months ago I found Attio, and its been smooth sailing since then! Yes, this is an ad, but I actually use this CRM daily! For the people I want to meet, to the people I invite to cover their founder stories.

“When I first opened Attio, I instantly got the feeling this was the next generation of CRM.”
— Margaret Shen, Head of GTM at Modal

Attio is the AI-native CRM for modern teams. With automatic enrichment, call intelligence, AI agents, flexible workflows and more, Attio works for any business and only takes minutes to set up.

Join industry leaders like Granola, Taskrabbit, Flatfile and more.

👉Start for free today

Estimated read time: 6 minutes

Every December, Andreessen Horowitz asks their investment teams to share one big idea they think builders will tackle in the coming year. This year, they published predictions across Infrastructure, Growth, Apps, American Dynamism, and Crypto, spanning everything from agent-native infrastructure to the first AI-native university.

This matters even if you never plan to pitch A16Z. They're the largest, most influential VC firm in tech. Understanding their lens helps you understand where capital is flowing, what narratives are gaining traction, and how to position your company if you're fundraising in these spaces.

A quick note before we dig in: VCs talk their book. Everyone does. They make bold predictions, celebrate the ones that hit, and quietly forget the misses. This isn't criticism; it's the game. The key is to use their predictions as market intelligence, not gospel. You're building the company. They're placing bets on what might work. Those are very different jobs.

The predictions skew heavily toward infrastructure. Agent-native infrastructure. Data stack evolution. Cybersecurity automation. Systems of record being rebuilt from scratch.

If you're building AI applications, you're entering while the infrastructure layer is still being built. This isn't a timing risk; it's reality. You need to be agile and nimble, figuring it out as you go since the tooling will continue to change right alongside your startup. The winners will be the ones who can adapt as the foundation shifts beneath them.

The vast majority of these predictions focus on enterprise and B2B opportunities. Consumer AI gets mentioned, but sparingly and in specific contexts like new distribution platforms or voice-based interfaces.

Here's the thing about consumer AI. There's always an element of luck involved in consumer. VCs, unless they specifically focus only on consumer, will always largely talk about enterprise and B2B because it's more predictable. Repeatable playbooks. Clear value propositions. Measurable ROI.

Consumer AI hasn't failed to prove itself. It's just that most VCs prefer the safer bet.

There's a clear pattern across nearly every prediction: don't retrofit AI onto old systems. Rebuild from scratch.

Banking infrastructure rebuilt with AI-native platforms. Agent-native infrastructure that treats "thundering herd" patterns as the default state. Industrial bases that start with simulation and AI-driven operations.

For fundraising, this matters. "We added AI to X" won't excite investors looking at this thesis. "We rebuilt X for an AI-native world" might. It signals you understand the shift isn't about adding features but reimagining what's possible from the ground up.

We're moving from text prompts to multimodal inputs and outputs. We're moving from chat interfaces to proactive AI that acts without being asked.

This isn't three years out. It's happening now in products people use daily. You're already seeing it: AI that can process images, video, and text together. Interfaces that suggest the next action before you think to ask.

The early opportunity? UX patterns for prompt-free interfaces are still being figured out. The winners will be the teams who crack how AI should surface insights and take action without overwhelming users or breaking their workflows.

Palantir pioneered the forward-deployed model. YC has been pushing it hard in their content and who they're backing. Now A16Z is calling it out explicitly.

For early founders, this might be your best competitive edge. Going deep into unsexy, non-SF industries. Understanding workflows that Silicon Valley VCs will never see firsthand.

Middle America has document-heavy, workflow-heavy businesses that don't need venture outcomes but could be completely transformed by AI. Legal services in small towns. Regional manufacturing operations. Local government procurement.

These aren't sexy. They won't make TechCrunch. But they're real businesses solving real problems, and the barriers to entry are understanding the domain deeply enough that you can't fake it.

There's renewed enthusiasm for hardware, robotics, manufacturing, and energy. AI-native industrial bases. Autonomous labs accelerating scientific discovery. The renaissance of the American factory.

As someone who studied robotics in undergrad, I'm genuinely excited about this. We're seeing feedback loops shorten dramatically: simulation, AI-driven design, automated operations. The timeline is still longer than SaaS, but the barriers to entry are dropping fast.

If this interests you, know that it won't move at software speed. But the opportunities are massive, and the moats are real once you build them.

One of A16Z's crypto partners made an observation that applies beyond crypto: every crypto company that's doing well is pivoting to trading. But if everyone becomes a trading platform, where does that leave everyone?

The broader lesson for early founders: chasing immediate revenue or engagement can kill long-term defensibility. Product-market fit is important, but so is building something durable.

Validate relentlessly. Listen to your users. But don't let short-term metrics override building something that compounds over time. The marshmallow test applies to startups too.

A16Z predicts we'll see the birth of the first AI-native university in 2026. An institution built from the ground up around intelligent systems, where courses adapt in real time, reading lists evolve nightly, learning paths shift to meet each student's pace and context.

Having worked in EdTech, this feels closer than it ever has. Personalized learning paths. Real-time curriculum adaptation. AI tutors that adapt to each student's pace and curiosity.

The opportunity here isn't just universities. It's the entire education infrastructure, from K-12 to corporate training to continuing education. Everything is ripe for complete reimagining.

For Aspiring Founders: Mine these predictions for ideas, but don't treat them as a startup shopping list. The best opportunities might be in the gaps: unsexy industries, non-venture scale businesses, problems VCs won't see because they're outside their pattern-matching playbook. Follow threads that genuinely interest you. You'll need that conviction to survive.

For Active Founders: Don't pivot based on a VC blog post. Do understand the narrative if you're fundraising in these areas. But let your validation and market insight drive decisions, not trend reports. You're closer to the problem than any investor will ever be.

On Hype vs. Reality: Some predictions are moonshots. That's the VC job. Your job is different. Trust what you're seeing in the market over what VCs predict. Build for the world as it is, not as investors wish it would be.

A16Z's predictions show where capital flows and what narratives are gaining traction. They're not a roadmap for what you should build. The best founders build what they see needs to exist, validate relentlessly, and trust their insights over anyone else's predictions.

If you want to read the full series from a16z: Part 1, Part 2, Part 3.