Synthetic User Startups Are Venture Capital's New Frontier

Lucas: So Luna, synthetic users - the idea that you can test your product with thousands of ai generated digital personas instead of real humans - just got a massive vote of confidence from venture capitalists. Luna: You're talking about Simile, right? The startup that raised $200 million at a $2 billion valuation, just five months after a $100 million Series A. Lucas: Exactly.

That is an astonishing pace - doubling valuation in under half a year. And it signals that the market for synthetic user platforms is white-hot right now. Luna: For someone who hasn't encountered this term - what exactly is a synthetic user? Is it like a digital twin?

Lucas: Close, but more specific. Synthetic users are ai generated personas that mimic real human behavior - clicking, scrolling, filling forms, even making purchase decisions. Developers use them to test apps, websites, games, and AI agents at massive scale, without the cost or privacy concerns of recruiting real testers. Luna: So basically, instead of paying for a user testing service or running a beta, you spin up a thousand fake users and watch what they do.

That could save a lot of time and money. Lucas: Right. And the timing is interesting. Today, LinkedIn added a button to report ai generated 'slop' - that term for low-quality AI content.

There's a growing need to distinguish good AI from bad, and synthetic user platforms could help build better AI systems by stress-testing them before launch. Luna: But isn't there a risk that synthetic users don't reflect real user behavior? I mean, if you're training on synthetic data, you might get a false sense of confidence. Lucas: That's the big open question.

Simile claims their models are trained on anonymized real user sessions, so the behavior is statistically representative. But there's always a gap - especially for emotional or subjective responses. Still, VCs seem to believe the scale benefits outweigh the fidelity risks, at least for initial testing. Luna: What about the competitive landscape?

Are there other startups in this space, or is Simile the clear leader? Lucas: There are a few - Synthetic Users, UserHub, some open-source projects. But Simile's rapid fundraising suggests they've captured the most mindshare. The $200 million round was led by Sequoia and Lightspeed, so heavy hitters.

They're likely using the capital to expand their platform and hire sales teams. Luna: And the broader trend - synthetic content, ai generated everything - is clearly accelerating. Just look at the Spotify announcement today: 'User notes' for adding memories to songs. That's user-generated, but ai generated content is everywhere.

Lucas: Yeah, the line between real and synthetic is blurring fast. For VCs, investing in synthetic user startups is a bet that testing will become fully automated. If you can simulate a million users overnight, why wait for a gradual rollout? Luna: But what about the ethical concerns?

If synthetic users are used to test, say, an AI agent that makes loan decisions, could they inadvertently bake in biases? Lucas: Absolutely. If the training data for synthetic users has historical biases, those will carry over. Simile says they have bias detection built in, but it's an area where regulators will likely pay attention.

The same way the SEC looks at algorithmic trading, there might be scrutiny on synthetic testing in regulated industries. Luna: So it's a double-edged sword: faster development but potential for hidden flaws. If you're building or investing in this space, staying informed is critical. Lucas: You know, if today's episode gave you a clearer picture of synthetic user startups and why VCs are piling in - the kind of insight you can use - this show stays ad-free because of listener support.

If that's useful to you, you can help us keep going at buymeacoffee.com/fexingo. Luna: Yeah, it's a small gesture that really adds up. And we appreciate everyone who chips in.

Lucas: Now, back to Simile's valuation. At $2 billion, they're already a unicorn. The question is whether they can grow into that multiple. They'll need to prove that synthetic users aren't just a novelty but a core part of engineering workflows.

Luna: One thing that works in their favor: the shift toward AI agents. Companies are deploying more autonomous software than ever, and testing those agents requires complex scenarios. Lucas: Exactly. If you have a customer service AI, you want to simulate thousands of angry customers to see how it handles.

Synthetic users can generate those edge cases efficiently. Luna: But there's also the macro environment. Tech stocks have been mixed - META down 10% in the last five days, NVIDIA down 6% - but AI infrastructure spending isn't slowing down. So VCs are still putting money into ai enabled tools like Simile.

Lucas: Right, and that's a key point. Even as some high-flying names cool off, the demand for practical AI applications remains strong. Synthetic user testing is a classic example of a 'picks and shovels' play - you're betting on the infrastructure, not on any single product. Luna: Who are Simile's typical customers?

Are they mostly gaming companies, or is it broader? Lucas: It started in gaming - Roblox and Epic were early adopters - but now it's expanding to fintech, e-commerce, and healthcare. Any company with a digital product that needs to be tested at scale. Luna: And the pricing model?

Is it per synthetic user, per test run? Lucas: Simile charges a subscription based on the volume of synthetic interactions. So the more you test, the more you pay. It's a recurring revenue model that scales with usage - which VCs love.

Luna: Given the speed of their fundraising, I wonder if they'll go public soon. Or would they stay private longer? Lucas: At a $2 billion valuation, an IPO in the next 12 to 18 months is plausible if they can show strong revenue growth. But they might also attract acquisition interest from bigger tech companies looking to add testing capabilities - think Microsoft or Google.

Luna: Speaking of Microsoft, their stock is up nearly 20% in five days. That could give them the appetite for a deal. Lucas: Brilliant point. But for now, Simile seems focused on building independent.

The $200 million gives them plenty of runway to dominate the synthetic user space before any exit. Luna: One last thought: synthetic users might eventually become a standard part of every software development lifecycle, like unit tests. Lucas: Exactly. And if that happens, Simile could be the platform that defines the category.

But they'll have to overcome skepticism about validity and bias. It's a fascinating space to watch. Luna: Agreed. And for our listeners, if you're working on a product, might be worth giving synthetic users a try - or at least understanding how they could change the way you test.

Lucas: That's a great takeaway. Thanks, Luna. Luna: Thanks, Lucas.