The Unicorn Next Door show how High-Growth AI Startups Are Changing Business Before Our Eyes

In recent years, achieving a billion-dollar valuation has become much more common amongst new businesses than when one would have appeared on the news with much shock value. These businesses are now popping up weekly. So many are appearing that the term “unicorn” has lost its impressiveness. The achievements of these companies are still very impressive. Reason for the loss of significance is that the game has dramatically shifted in the past few months with tremendous growth taking place in the high-growth AI sector. One of the things that many of these articles will not tell you. However, is that while most of these new AI unicorns are the same, they may differ significantly in their profile and structure. For example, you may read about giant “Godzilla” type AI companies that are creating supercomputers. But at the same time, there are many smaller, more innovative and “boring” companies. That they are providing some of the cleverest solutions to some of the most mundane problems within the marketplace. In this discussion, we will explore these Emerging AI unicorn profiles as they exist today, without any of the corporate jargon or hype associated with them.

Why “Unicorn” Doesn’t Mean What It Used To

In previous times, Unicorns were characterized by having a lot of employees and having many square feet of office space, allowing them to grow gradually. Traditional industries such as auto manufacturers and traditional retail banks fit this model. The advent of the internet changed this picture. Instagram had just 13 employees when it was acquired for $1 billion. WhatsApp employed 55 engineers to support 450 million users. AI has vastly increased the speed of this trend. Currently, the definition of an A.I. Unicorn is vague because the time it takes to take a new idea/startup to a valuation of $1 billion or higher has been significantly compressed. The rapid increase in start-up valuations is attributed to two main causes.

First, traditional infrastructure is now a service and can be purchased on a subscription basis instead of building your own Data Center. Thus, you can rent computer services from the large Cloud Services Providers. Secondly, the significant velocity at which Artificial Intelligence allows a 10-person (e.g., Product Development) team to build a product that would have required a minimum of 100 people to build five years ago. As a consequence of these two factors, the barriers to enter any product or service have decreased. The amount of market entrants in any given market have increased drastically. Therefore, it is difficult to differentiate between a “flash in the pan” and a legitimate company. Thus, we can only differentiate based on the product.

Emerging AI unicorn profiles: Generalists vs. Specialists

When examining patterns of emerging AI unicorns, I commonly classify them into two easily distinguishable categories. The Brain Builders (Horizontal AI) and the Workforce Fillers (Vertical AI). Brain Builders (Horizontal AI, these are your enterprise level heavyweights in the AI space are giants. As example OpenAI, Anthropic and DeepSeek who are competing to create The General-Purpose Intelligence that will have the ability to perform any task. Essentially they are building the railway system for the future of AI. They are driving billions of dollars of investment in compute to build the “Operating System” for all other future AI.

Workforce Fillers (Vertical AI) are where most of the innovation is currently occurring. They don’t try to build an AI with the knowledge to replace a Google Search; they build an AI specifically for one job. For example, developing an AI that replaces the need for a junior associate at a law firm, or replacing the need for a medical scribe in a hospital. Cognition is developing “Devin,” an AI software engineer, as an example of replacing the need for junior software engineers, as opposed to creating an AI that can write emails. Another example is Harvey, which performs legal contract review. Another is Abridge, which listens during your doctor’s visit and automatically generates notes.

These companies are winning, because they are taking a boring, expensive, human driven process (like reviewing contracts) and making them instantaneous. Network effects also come into play. When a law firm trains an AI how to write contracts the way it does, it becomes a logistical nightmare to switch to a different AI. This creates a defensive barrier or moat as it is no longer “just” a chatbot in the law firm’s product and solutions portfolio.

The Money Game: Valuations and Reality Checks

Let’s discuss the digital wallet. When you hear of a new company raising an extra $100 million in funding, you may think, “Wow, they are making money hand over fist.” The truth, however, is much different. Currently, AI company valuations are very high and growing quickly. The average value of an AI company at the seed stage ($5 million) is roughly 42% greater than its non-AI counterparts ($3.5 million). Investors believe that AI companies grow faster than non-AI companies. To put things into perspective, there are a number of issues with AI revenue models.

In the traditional SaaS world where software-based companies offer tools like CRM software, their costs are essentially fixed. They only have to spend money to develop the software once and can keep selling it at a fixed price to 1, 100, or 1,000 customers without increasing their costs. However, with AI, every time a customer interacts with an AI-enabled app (e.g., asking it a question), there is a cost associated with that interaction (i.e., inference costs). Therefore, if you have an app or service that goes viral but have a poor revenue model, you could quickly go bankrupt as your application continues to grow.

So how can startups convert usage into revenue? Usage-based models, as an example $0.10 per chatbot conversation or $0.25 per generated report. Results-based models, some AI companies are experimenting by charging customers for performance (i.e., only charging when the AI solves a customers’ problem, e.g., saves the company $$). Currently, investors are looking for companies with revenue-by-the-cloud type of businesses. Investors do not want to see companies generating $1 million in revenue and having $2 million in cloud-related expenses. Investors are looking for companies with unit economics that improve as their business grows.

The Geography Shift: Asia’s Emerging AI Unicorn Profiles

Now, let’s turn our attention home to Asia with a focus on Malaysia. When we think of Silicon Valley, we often think of how we can improve. In some areas, we are. The U.S., for example, has the largest share of AI funding. The competitive landscape of Asia is different. We don’t have to create the $852 billion brain (that’s OpenAI’s domain); we have to create the application. Malaysia is investing a lot of money into the “AI Nation” agenda, constructing data centers and sovereign cloud capabilities. But this is where the real opportunity exists for Malaysian founders: in vertical AI. There is a huge market opportunity analysis available to be conducted here. Consider our local industries: Banking, palm oil, logistics, and the Public sector. Each of these is an enormous and complex industry with unique and complex regional laws. An AI developed in the U.S. does not know how to approve a loan application in Malaysia.

This is where the geographic market expansion becomes interesting. Companies such as Eezee (Singapore/Malaysia) are using AI to improve procurement. Amity (Thailand) has raised millions to create “Agentic AI” for businesses, demonstrating that you don’t have to be in San Francisco to achieve billion-dollar valuation. Investment and venture capital activity in Sea is on the move. DBS and Granite Asia have launched a $110 million IPO fund that is focused on AI. The two companies believe that the next wave of unicorns will be generated by solving regional issues with global technology solutions.