Published on January 15, 2026
Watch the video of this panel on Gamma Prime’s YouTube channel:
Panel talk with Michael Terpin (Founder and CEO at Transform Group), Christian Sauer (CEO at Soonami), Catrina Wang (GP at Portal Ventures), Amit Mehra (Partner at Borderless Capital) and Anna Shakola (Head of Business Development at Cointelegraph Accelerator)
Anna Shakola (Head of Business Development at Cointelegraph Accelerator):
Okay, any names we should be aware of?
Catrina Wang (GP at Portal Ventures):
Well, to be honest, everyone will turn to AI. I mean, a year before I was doing more of an OS narrative. Now it’s an AI OS, and Polygon has an AI narrative.
Everyone wants to dovetail to what’s trendy. So it wouldn’t surprise me if just a lot of these big names will become AI, but ultimately, as an investor, it’s about parsing through the noise, right? Which company actually needs to do it.
I talked to this Prime Intellect, which is actually doing interesting things and was recently founded by a Founders Fund. I’m not sharing my own bag with someone else’s bag, but in general, I think it’s important just to figure out what’s a real AI company, authentic to their core value proposition, or what’s just riding a trend.
Anna Shakola:
Can you specify a real AI company? According to the Grab Accelerator, we have two names. One is Tesla, for companies that maybe are not very well promoted, but they have core technology.
And Edison, the company has questionable technology, but they’re really good with promotion. So from your perspective, from Portal Ventures’ perspective, how do you evaluate an AI startup? How do I—what?
Evaluate—how do you do due diligence? What are key signs that this is a good AI startup for you?
Catrina Wang:
Huh, that’s interesting. So I wrote about AI Web3 back in 2023, but actually, from then to now, we only made one investment. That’s Exibit.
That’s the one generating more than 10 million.
Anna Shakola:
This is my angel investment.
Catrina Wang:
Oh, wonderful, wonderful. The team actually has real traction, whereas a lot of the agentic narrative, or, you know, on-chain verification—okay, let me make an important distinction here.
The fact that we didn’t make the investment, or it didn’t pass our evaluation, is more because of the business model, not because the technology is not important. As an investor, and we have this trend over the past two years, there’s a romanticization of, oh, the more techie, the better. The less people can understand the tech, oh, the more high-end it must be.
But as an investor, the business model is the most important to me, and the challenge I have with the intersection of Web3 and AI is just how do they monetize this? Like, what is the willingness to pay, for example, for verification?
Verification, when it comes to the context of decentralized AI, is very important, because I need to verify if this is a deepfake. I need to verify if the output is actually coherent with the input. But then, just like privacy, everyone’s screaming, oh, super important.
But when it comes to, do you want to pay for privacy? Most people are probably saying no, right? So the reason we didn’t invest that much in AI Web3 is, like, I still struggle with this model there.
But yes, some tech needs to be built, and you need to look for what’s the Venn diagram in the middle of what’s commercial and what has product-market fit.
Anna Shakola:
Anik, do you have something to add?
Amit Mehra (Partner at Borderless Capital):
I think that also, like, a lot of other very interesting areas coming up are agentic AI and all the infrastructure that has been built. Eventually, I think the idea is, like, entire funds—for example—and the most prime area for disruption is fintech.
So how can you have an entire hedge fund? There have been several cases where there has been an agent picking tokens, investing in tokens, and generating very good returns, you know, more than actual human beings.
So I think the two more, maybe, base protocols, like MPC, what Coinbase released, X402—I think those are important. And a lot of startups that we have spoken to, so many, are actually building upon these protocols to work on agent-to-agent payments, agent-to-agent communication, and those—so those would be the other names.
Christian Sauer (CEO at Soonami):
Yeah, I like Amorphis, by the way. That’s an interesting project. I think they have funded the whole infrastructure through a fair launch model that is supported by staked Ethereum.
So we took that model and actually want to offer it to pretty much any other company out there.
So to me, this is really an interesting name.
Michael Terpin (Founder and CEO at Transform Group):
Of course, the issue with the Amorphis funding model, which worked brilliantly for them, was that the thesis with Amorphis—it was the first time it was done—then it’s been replicated by a few of their portfolio, I mean, their sort of ecosystem.
You go and stake your Ethereum on Lido, and you contribute the interest to the project, and they ended up having $400 million staked. Now, the main people behind the project were David Johnston and Eric Voorhees. Most projects don’t have iconic billionaire founders who could probably be staking that entire amount themselves, and then other people are like, oh, if they’re building this.
So when you have a newer startup, if somebody stakes a million dollars, you only get $30,000 in funding.
Yeah, when it becomes $400 million, all of a sudden you’re talking about real money, but I think it’s difficult to do that. But there are some interesting models, because again, Amorphis ended up spinning out of Venice, which is Eric Voorhees’ decentralized AI platform.
And that ended up spinning out the Venice token, VVV, which on the first day went over a billion dollars. It’s pulled back. I think it’s still a good project that will probably recover, and I’m bullish on that one because it has use cases.
You use the VVV token to actually buy inference on Venice search. You can use it as Web2, and you can also use it as Web3.
So I think there are not that many hybrids where it’s easy to come in with zero crypto knowledge, but it’s also much more profitable to utilize it with Web3. That’s what’s kind of missing right now, and I think in the next—hopefully faster—but in the next three to five years, you’re going to see that most projects in decentralized AI will need very easy onboarding for the Web2 user, or they’re just not going to be competitive.
