Is venture capital fast enough for AI scale-ups?

We are entering a 24-month window that will redefine venture capital, and most investors are still grading AI-native companies on outdated SaaS frameworks. Firms built for the last wave will be miscalibrated for this one. Here is the shift I'm watching, and the one thing about these companies I still can't price.

I don’t move fast enough. Whether it’s the world class builders we have in our Venture Studio, the talent we see coming out of universities, or AI Native Founders that have given up sleep to capitalize on the window of opportunity for AI Agent scale up companies, I am reminded every day, I am too slow. And, just like in sports, you can improve speed, but world class speed is not a matter of training or improvement, it’s a talent. 

When you're not fast, you have to leverage other strengths. For me, it’s storytelling and team building. How do we as Partners of a VC firm build the fastest and most agile team, ready for the new venture landscape? We have to build a new foundation, we have to change, we have to shift from capital allocators to true co-builders with the startups we invest into. But why do we need to become world class at speed?

We are entering a 24-month window that will redefine venture capital. While the value of traditional SaaS companies continues to lower, AI scale-up startups are presenting an unprecedented opportunity for investors, and the speed they are building with has never been seen before. 

Meanwhile, most investors are still trying to evaluate AI-native companies using outdated SaaS frameworks. The looseness of how we define these companies is starting to cost the industry, because firms built for the last wave of SaaS will be completely miscalibrated for this new era. To understand this window of opportunity, we have to look at how AI-native founders are fundamentally rewiring the organizational structure of their companies and the platforms they use to fund them.

Here’s the shift for venture capital in a world of AI scale ups. 

Traditional SaaS companies are losing their premium valuations because their moats have moved and their unit economics are different. In a traditional SaaS company, AI is just a feature bolted onto a conventional product; if you strip the AI out of the stack, the company still exists, it just runs slower and requires more people.

For AI scale-ups, the model sits at the structural core of the company. The model is load-bearing; remove it, and the entire product and operation collapses. Because the cost structure is entirely different from traditional SaaS, the diligence cannot be lifted from the old SaaS playbook.

AI-native founders are building companies with an organizational structure that looks completely wrong on a traditional SaaS spreadsheet. The ratios are off: there are more operators than engineers, and a 3-person team is now doing the work that used to require 30 people.

We have not fully worked out what this hyper-efficient structure does to compensation, equity splits, or how these companies scale past their first $10M in revenue. But one thing is clear: because these teams are smaller and heavily leveraged by AI, their rounds should probably be smaller, too.

Because AI-native founders are building companies operated by AI agents, they will not rely on the fundraising systems of the SaaS era. An AI-native founder does not need a coffee chat, a warm intro, or a twenty-slide deck to prove market fit. They will go to platforms that fit how their agent-driven companies already run, like Pitch Protocol.

What if there were an API for venture capital, one that solved the cold-start problem by letting AI agents natively pitch, match, and scale with investors in an environment built for the machine economy? A founder links their data room, and their agent packages a structured context packet and submits it in a single command. In parallel, investor agents score the pitch against a specific thesis, on stage, sector, and check size, with no human in the loop until it is time to decide on capital.

As AI agents begin to research, build, sell, and operate companies, they need a structured, agent-native way to meet the investors who understand them. That is the shift: fundraising moves from a human-only networking game to clear context, sharper matching, and faster signal, and whichever platforms get there first will define how it works. We built Pitch Protocol. AI-native founders are using it, and they like it.

Aggregators that simply forward pitch decks are obsolete. VCs must adapt to a world where founders submit structured data directly from their agents, allowing investors to evaluate on substance rather than storytelling. By receiving deals that are pre-screened and matched to their exact thesis, investors can stop reading 200 decks just to find 5 worth a meeting.

We built this foundation, fast. We are iterating and learning, fast. Our speed is matching the companies we want to fund. And those companies are rewarding us with trust, opportunity and insights. Our speed is building an unfair advantage. And we will use this to build the future of venture. 

I'll never be fast enough, and that's okay. Our team will be. This is one of the widest windows for value creation I have seen in venture, and we will move as fast as the AI-native founders chasing it. We will disrupt the old systems and networks. We will show up ready to play the same game, at the same level, and deliver wins for our investors, founders, and partners.

The new world of venture is a team sport. If you are building something that looks wrong on a SaaS spreadsheet, or you think I have the window wrong, respond to this email. I read every reply. Better yet, suit up and join us in person this summer.

TrackLight — acquires fellow Fund I company Plum Identity
Two GFV portfolio companies just became one. TrackLight, which monitors government payment programs for fraud, acquired Plum Identity's identity-verification engine to build a single anti-fraud platform covering the full lifecycle, from application to post-payment investigation. It reunites the team behind Pondera Solutions, sold to Thomson Reuters in 2020, and says it has booked over $1M in new government contracts in 90 days.

Xella Health — public launch, 130+ women's conditions
Xella launched as an AI precision-health platform built for female biology, screening more than 130 often-misdiagnosed conditions including endometriosis, perimenopause, and PCOS. It came out on $4.7M from Precursor Ventures, Capital F, and Ulu Ventures, with a 15,000-person waitlist and a $499 annual membership through its own CLIA-certified lab. The thesis: women's health is under-diagnosed for lack of female-specific data.

HuLoop — named to the 2026 AIFinTech100
FinTech Global named HuLoop to its 2026 AIFinTech100, a back-to-back pick after its 2025 nod. The recognition is for a no-code platform coordinating human, robotic, and agentic workers across lending, compliance, and fraud inside banks and credit unions. The signal for founders: the agentic-work framing is landing with the conservative institutions that write the checks, not just the analysts naming the category.