Venture Capital investor reveals her biggest fear isn’t AI companies failing — it’s their absolute success
Artificial Intelligence is already changing the rules of the game for tech investors, and anyone who hasn’t noticed yet might be sleeping on it.
Lexi Novitske, General Partner at Norrsken22, a fund focused on tech growth in Africa, published a piece in Fortune that became a hot topic in the Venture Capital world for a pretty simple reason: she said what a lot of people in the industry think but rarely have the guts to put into words.
Her biggest nightmare isn’t watching portfolio companies fail.
It’s the exact opposite.
What if AI is so successful that the very business models of these companies become irrelevant?
That question sits at the center of a discussion that’s been gaining momentum over the past few months, especially after the SaaS market started showing serious signs of pressure, with over 1 trillion dollars in market value evaporated since the beginning of the year. 📉
But the story doesn’t end there — far from it.
Lexi points to emerging markets, particularly Africa, as one of the most promising scenarios for those who want to bet on companies that AI simply won’t be able to make obsolete anytime soon, and the reasons go well beyond what you might imagine. 🌍
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The vibe coding phenomenon and the real threat to SaaS
One of the most relevant points Lexi highlights in the original article is the concept of vibe coding — AI-assisted application development. The idea is relatively new, but the impact is already being felt. Apps that used to take years of engineering work can now be replicated in weeks with the help of generative AI tools. This completely changes the value equation for startups that built their competitive edge on technical complexity and development time.
At the same time, major tech companies like Microsoft are integrating AI features directly into their product suites. Autonomous chatbots that used to command premium pricing are losing their pricing power overnight. HR management tools — those startups that automate payroll, performance reviews, and administrative HR tasks — face the same threat as those features start being offered as part of larger, cheaper bundles.
The result? Venture Capital investors face a real risk of paying inflated prices for companies that look special today but could become generic — or worse, easily replicable by AI coding tools — within a year. It’s a scenario that demands a deep shift in how investment theses are built and validated.
The real problem SaaS faces with AI
For years, the business model based on software as a service was considered practically unbreakable. Recurring revenue, low distribution costs, near-infinite scalability, and margins that would make any other industry jealous. SaaS was the darling of Venture Capital precisely because it promised predictability and compounding growth over time — two things any investor loves to see in a portfolio. Companies built on that foundation dominated the last decade of tech and shaped how digital businesses are structured to this day.
The problem is that Artificial Intelligence came along to challenge the very pillars that made this model so attractive. Generative AI tools can already automate tasks that used to require specialized software, entire teams, and hefty annual contracts. What used to cost tens of thousands of dollars a year in software licenses can, in many cases, be replaced by an AI agent running at a fraction of that cost. This isn’t science fiction — it’s happening right now, in real time, and the market is already feeling the weight of it in public company valuations and funding rounds that are taking longer to close.
The value erosion the SaaS sector faces isn’t a temporary market hiccup or the result of an economic cycle correction. It’s structural. When a technology can replicate the core function of a product, the competitive edge vanishes quickly, and what’s left is a race to the bottom on price. The so-called SaaSpocalypse that has swept through capital markets since the start of the year is concrete proof that AI agents are already capable of executing entire workflows that previously depended on multiple SaaS subscriptions. Venture Capital investors who built entire theses around vertical software companies now need to answer an uncomfortable question: which part of the portfolio survives when AI can do the same job faster, cheaper, and without needing onboarding? 😬
Lexi’s strategy: treat AI as a layer, not a vertical
Given this landscape, the approach Lexi takes with Norrsken22’s investments is very specific and structured. Instead of treating Artificial Intelligence as an isolated sector — which many funds still do — she proposes viewing AI as a technology layer that permeates everything, and focusing on the underlying structures that the technology can’t easily replace.
In practice, she looks for three characteristics in any company before investing:
- Ownership of customer trust and distribution — companies that have built real relationships with their users, something no language model can replicate from scratch.
- Integration into systems where real money flows — being embedded in critical financial and operational processes makes replacement much more difficult and risky for the customer.
- Accumulation of proprietary data that generates compounding efficiency over time — the longer the company operates, the more unique data it collects, and that data becomes a competitive advantage that grows exponentially.
This logic also changes how she allocates capital. Instead of making one big bet on a single AI-centric company, Lexi prefers to make multiple bets within the same vertical. The example she cites is the fraud, compliance, and security sector. In Norrsken22’s portfolio, Smile ID handles KYC and identity verification, while Orca works on transaction fraud monitoring. The expectation is that over time, these areas will overlap, and the company with the deepest data wins — regardless of which underlying AI model is being used.
Another area she’s watching closely is agentic platforms — tools that enter the market as generalist solutions but gradually embed themselves into the workflow of a specific industry. Companies that build for well-defined niches tend to be much harder to displace than those trying to solve everything for everyone all at once. 🎯
Why emerging markets change the game
The logic Lexi Novitske presents is both simple and powerful. In emerging markets like African countries, the problem isn’t replacing existing software with Artificial Intelligence. The problem is much more fundamental than that: a large portion of the digital infrastructure that the rest of the world built over the past few decades simply doesn’t exist there yet. There are no layers of legacy technology to dismantle, no long-term contracts with software vendors that need to be renegotiated, and no cultural resistance to adopting new tools because the old ones are already entrenched. It’s wide open territory. 🚀
The African continent is made up of 54 countries, each with its own regulatory environment, currency, and infrastructure. It’s precisely this fragmentation that caused major global AI players to overlook the region, preferring to focus on the United States and Europe. For a long time, this was seen as a disadvantage for scaling businesses. But Lexi argues that this same fragmentation works as a natural protective barrier for local companies, which can develop and mature while big tech concentrates its efforts on other geographies.
This creates a unique opportunity that Venture Capital funds focused on these territories are starting to see with much greater clarity. Companies built from the ground up in these contexts can adopt AI as their foundational infrastructure from day one, without worrying about compatibility with legacy systems or resistance from teams accustomed to established processes. The business model that emerges from this is fundamentally different from what was built in the West: leaner, more adapted to local reality, and paradoxically, more resilient to disruption because it was conceived in a world where AI already exists and operates. It’s not an adaptation — it’s a foundation.
On top of that, the specific challenges of emerging markets create demands that generic AI developed in major tech hubs will have a hard time solving on its own. Language issues, cultural context, payment infrastructure, logistics in regions with limited connectivity, financial inclusion for populations that have never had a bank account — all of this requires solutions that combine advanced technology with deep knowledge of the territory. And it’s exactly this local knowledge that represents a massive barrier to entry for any global player trying to break into these markets without partners and without roots. Artificial Intelligence can certainly automate parts of the process, but it can’t replace the contextual intelligence of someone who built the business knowing the customer up close. 🌍
The Sabi case and the power of regional data
A concrete example Lexi highlights is Sabi, a platform backed by Norrsken22 that operates in the commodities market in Africa. The company doesn’t just connect buyers and sellers — it coordinates sourcing, transportation, quality control, warehousing, and financing across borders where none of these systems communicate reliably.
AI is used to manage tracking, traceability, and price information flows throughout this entire chain. The company that solves this problem end to end becomes the default operating system for its market, and the data it accumulates in the process is extraordinarily difficult for any competitor to replicate. That data doesn’t exist anywhere else. No AI model trained on global datasets has access to this kind of granular, contextualized information.
This is the pattern Lexi is watching for: companies that don’t just sell software but actually operate the process, accumulate exclusive regional data, and build trust in markets where that asset is scarce. These are businesses that go far beyond the superficial technology layer and create value that’s hard to extract or copy.
What this means for the business model of the future
The discussion Lexi raises goes beyond geography. What she’s signaling, in practical terms, is that the next Venture Capital cycle will reward companies that build value on top of assets that Artificial Intelligence can’t easily replicate: relationships, trust, physical distribution, proprietary data generated by very specific behaviors from a very specific audience, and the ability to operate in complex regulatory and cultural environments. These elements don’t appear in any language model training dataset, and that’s exactly why they’re so valuable right now.
The business model that will thrive in the coming years isn’t necessarily the most technological or the one using the most AI in its operations. It’s the one that uses AI strategically to amplify human and local capabilities that already exist, creating a combination that any outside competitor will have a really hard time replicating. SaaS companies that understand this and start building layers of value beyond software functionality — including community, unique data, deep integrations with the local ecosystem — have a real shot at surviving and growing even as AI pressure intensifies. Those that stand still, betting only on the strength of the product itself, will feel the squeeze getting tighter and tighter. 💡
The shift in investor mindset
Lexi wraps up her argument with a reflection that sounds simple but carries serious weight. She stopped asking which AI company should I back and started asking which problems are the big players ignoring — and which companies are building solutions so deeply integrated into those problems that AI won’t make them obsolete.
This shift in perspective is significant because it reorients the entire investment logic. Instead of chasing the next AI startup with an impressive demo and a slick deck, the focus moves to the sustainability of the value created. Proprietary data that compounds over time, trust built in markets where it’s rare, and integration into real financial flows where actual money moves — these are the assets that define resilient companies in the next cycle.
For Venture Capital funds still calibrating their investment theses, the message coming from experiences like Norrsken22’s is pretty straightforward: geographic diversification is no longer just a risk strategy. It has become an intellectual survival strategy for any portfolio that wants to remain relevant in a world where Artificial Intelligence is compressing margins, accelerating obsolescence cycles, and making businesses that seemed bulletproof just two or three years ago completely irrelevant.
Emerging markets aren’t plan B. For many investors, they’re quickly becoming the main plan. In Africa, according to Lexi, these companies are everywhere. Investors just need to be willing to look where nobody else is looking. 🌐
