Historically, successful tech companies relied on moats like proprietary technology, network effects, or unique datasets in order to establish dominance over a market. But in the current AI age, one of the most common, critical and sometimes patronizing, questions that is posed to founders is “what is your moat”?
After asking this question to many founders myself, and hearing many of the companies we work with get asked this question by other investors, a different, Shakespeare-inspired, question kept coming to mind: “What’s in a moat”?
It is indeed a critical question. How should founders go about creating long-term businesses when the path towards building defensibility is blurry? And how should investors evaluate which companies have staying power in order to create true economic value in the long run?
While moats may have come in various shapes and sizes in the past, I believe that for AI companies operating at the application layer, there’s only one durable moat over the long term: brand.
Brand is trust, familiarity, emotional resonance, and habitual use all rolled into one. But brand isn’t something you start with; it’s something you earn.
The best proxy for understanding this is the AI chat interface war. While in the tech community, many are aware that a battle is underway between OpenAI, Anthropic, the hyperscalers, X and others, many in the real world equate AI with ChatGPT. ChatGPT is the Kleenex of AI. So while the actual chat application may already be commoditized, the brand of ChatGPT is, in my view, the type of impenetrable moat that startups should be trying to emulate.
So, how exactly do startups build their way toward that impenetrable moat?
The Barbell Strategy
Broadly speaking, there are two viable strategies to reach the point where you have a defensible brand.
Rapid deployment to the long tail
This strategy involves rapidly deploying a minimally viable product, often essentially a GPT wrapper, and immediately chasing broad adoption. Companies that deploy this strategy must benefit from a first-mover advantage; being first or nearly first to market provides the best shot at embedding a brand in the collective consciousness.
Due to the initial lightness of the product, the most relevant target audience for companies that employ this strategy are either consumers or nano businesses.
Initially, rapid deployment companies don’t need much capital. But here’s the catch: the moment genuine product-market fit is reached, these companies will need to aggressively raise, perhaps even more than feels necessary, in order to blitzscale their market presence. Ample capital is critical for rapidly capturing market share and solidifying brand dominance before competitors flood in.
Stealth mode deployment to true enterprises
Alternatively, companies can pursue a deep-building, stealth mode strategy. Spend 12-24 months quietly building with select enterprise design partners, ensuring that the product is enterprise grade (compliant/procurement passable) and multi-product from day one. Post eventual launch, these companies come out swinging, targeting massive contracts (high six figures or seven figures) from large enterprises right from the start of sales.
This approach requires substantial upfront capital. It’s best suited to serial entrepreneurs or founders with confidence in their ability to raise significant funds from patient, strategic investors.
Both strategies ultimately aim for the same destination: a multi-product offering that leverages user excitement around AI and cements the brand as synonymous with trust, capability, and innovation.
**Of course, all of the above analysis assumes that founders have the desire to build a multi-billion dollar company.**
The Barbell Strategy in Action
Let’s ground this theory in reality by looking at two pairs of AI startups employing these strategies:
AI Medical Scribe Companies
- Abridge has taken the stealth mode approach, spending significant time building an advanced AI medical scribe solution tailored to large healthcare systems. They’re betting big on deep integration and high-value, enterprise-grade functionality.
- Freed, on the other hand, rapidly launched a lightweight AI scribe product directly targeting individual practitioners and small clinics. By quickly iterating with real-world feedback, they’re establishing brand presence early through rapid adoption.
AI Code Gen Companies
- Cursor exemplifies the immediate GTM path. They swiftly brought a simple but powerful GPT-enhanced IDE to market, capturing developer mindshare early and leveraging self-service, bottoms-up adoption.
- Qodo, meanwhile, has quietly focused on building deep functionality aimed at enterprise-level developers and engineering teams. Their careful, patient strategy was designed for and has already secured large-scale, enterprise-wide contracts.
These examples illustrate clearly that both strategies can be successful, but execution and alignment with the founders’ strengths and market dynamics are critical.
The Mid-Market Trap
Referring back to the barbell image above, it’s worth discussing the middle of the barbell: the mid-market trap.
It’s tempting for AI startups to target the mid-market as an initial customer segment. The mid-market has larger contract sizes compared to the long tail and relatively faster deal cycles compared to enterprise clients. But the mid-market is a trap for applied AI companies because companies targeting the mid-market:
- Don’t move fast enough compared to startups chasing low-ACV, high-adoption plays
- Lack the product depth and stickiness of enterprise-focused companies
Ultimately, mid-market startups get squeezed from both sides. They’re too slow and cumbersome to blitzscale, yet too shallow and undifferentiated to fend off deeper-pocketed, more patient competitors emerging from stealth.
This makes it almost impossible to build true brand superiority.
Retention is the Real Battle
The true battleground for applied AI companies is retention. Early revenue for AI startups often feels like ACR (“annual curiosity revenue”) instead of ARR or “vibe revenue.” And that’s okay – early excitement and experimentation drive initial adoption. But customers will churn if companies don’t transition curiosity into sustained habits and deeper product engagement.
Of course, driving retention requires customer happiness and product depth and stickiness, but it’s fair to assume that in a competitive market which requires low upfront investment, multiple companies will be able to achieve those qualities.
Therefore, in my opinion, the best way to drive retention is through brand superiority. The old saying, “no one was ever fired for buying IBM”, rings true today: “no one will be fired for buying OpenAI licenses”.
Companies must relentlessly focus on building brand superiority in their category in order to drive retention.
How to Win in a Moatless World
In a landscape where traditional moats have all but disappeared, companies must avoid the temptation of the comfortable middle and choose their side of the barbell carefully:
- Move fast and blitzscale aggressively once PMF is reached
- Move deep, build patiently, and aim high from the outset
In a moatless AI world, the only moat that compounds is brand. The only question that matters is who achieves brand superiority first.
So what’s in a moat? Maybe Juliet was onto something: it’s your name.
