From: mk_thisisit
Kosmos.ai: A Case Study in AI Startup Valuation and Growth
Kosmos.ai, a company founded by Miron Mironiuk in the field of artificial intelligence, has achieved a valuation of over 2 billion zlotys, equivalent to 500 million dollars [00:00:03], [00:00:12]. This valuation, like any product valuation, depends on how much an investor is willing to pay for it [00:00:08], [00:05:35]. It’s important to remember that such valuations are virtual until the company is listed on a stock exchange or sold [00:05:27].
Kosmos.ai’s Mission and Technology
Kosmos.ai’s mission is to connect the physical and online retail worlds [01:53]. While e-commerce is widely discussed, 80% of global trade (valued at $23 trillion) still occurs in the physical world [01:38], [01:49]. The company aims to combine the best aspects of both online and offline shopping experiences [02:14].
The company uses artificial intelligence in several ways:
- To accurately understand what people buy in the physical world [04:46].
- To predict future purchasing desires [04:52].
- To help users save money and time [01:07].
- To help physical store owners save time and earn money [01:11].
- By leveraging geolocation data from smartphones [01:17].
The idea for Kosmos.ai stemmed from the founder’s previous work with major brands like McDonald’s and Unilever, where most products were sold physically, yet available tools focused on online purchases [03:07]. Marketers faced challenges like the inability to “retarget” customers who visited a physical store but didn’t buy [03:47]. The goal was to provide similar capabilities for offline retail, leading to huge savings for marketers and a better, personalized experience for buyers [04:12].
Strategies for Startup Growth
Market Selection and Global Perspective
The decision to establish Kosmos.ai in Asia, specifically Hong Kong and later Singapore, was strategic [02:21], [03:00], [11:07]. The founder initially considered the United States, but an analysis revealed key advantages in Asia:
- Market Size: China alone has 1.4 billion people, providing access to a vast user base and a billion phones [02:32], [02:55].
- Smartphone Penetration: While the US had only 25% smartphone consumption at the time, Asia offered higher penetration [09:17].
- Effort vs. Market Size: The effort required to enter a market is similar regardless of its size, making larger markets more appealing for greater potential users, customers, and data for AI algorithms [11:21].
- Data for AI: Artificial intelligence algorithms are only as good as the data they are built on, and Asia offers an “unlimited resource of data” [11:14], [11:51].
The selection process involved an Excel table comparing population size, smartphone penetration, retail market size, and advertising expenditure per capita [09:48]. This led to Hong Kong, Singapore, and Australia as top contenders, with Australia being ruled out due to its smaller population and distance [10:07].
The Role of Polish Programmers
Poland is recognized as a global leader in programming talent [00:21]. In international programming competitions over the last decade, Poland ranks fourth globally, just behind Russia, China, and the United States [12:44]. When considering medals per capita, Poland is the world leader [12:57]. Kosmos.ai employs a team of 80 people, with about 30 programmers, nearly all from Poland, many of whom are medalists from international programming Olympics [14:42], [14:55], [15:05]. This access to top-tier talent is considered a significant factor in the company’s success and is viewed as a “natural resource” more valuable than diamond deposits in the 21st century [13:03], [13:12].
Luck and Determination
While the founder attributes much of his path to luck, he emphasizes that “luck has to be helped” through stubbornness and determination [14:33], [14:38]. For Polish startups, the advice is to think beyond Poland, or even Europe and the United States, and consider the wider global market [27:21]. To seriously pursue international expansion, founders must be willing to go abroad themselves [27:34].
The Broader Landscape of AI and Its Future Implications
AI’s Current State and Impact
The development of artificial intelligence is considered a “breakthrough” moment in history [18:59], [19:32]. While AI has existed for years in applications like phone maps, email spam filters, and search engines [18:24], the recent public availability of interfaces like ChatGPT has made AI accessible and popular [17:36]. The success of OpenAI and ChatGPT is attributed not to a fundamental mathematical problem solved, but to a business genius in securing funding and providing an accessible, conversational interface [16:28], [17:10].
This era of artificial intelligence, alongside quantum computing and automation, presents a unique opportunity for progress [19:41]. It is compared to the Second Industrial Revolution in the 1920s, which allowed the United States to surpass Great Britain economically [19:50]. Artificial intelligence is seen as the greatest revolution since the beginning of the 20th century, potentially shifting the global economic puzzle and affecting the prosperity of nations more profoundly than the internet, personal computers, or social media [20:02], [20:54].
Potential Economic Impact of Artificial Intelligence
It is predicted that artificial intelligence could replace 30% to 50% of jobs by 2030, particularly repetitive and routine positions [21:50]. These jobs are often found in “third-world countries” [22:01]. This could lead to a situation where richer countries become even richer, and poorer countries become poorer, especially as global populations increase in regions like Africa and Asia while work becomes scarcer [22:06], [22:18]. This imbalance could lead to instability if not addressed [22:32].
Regulation and Control of AI Technologies
The speaker asserts that artificial intelligence is a tool that can be controlled by users and countries [23:38]. The key to controlling AI lies in controlling data [22:53], [22:59]. While individuals can limit access to their data, it’s not always easy, and regulations like GDPR are seen as having good intentions but insufficient to stop malicious actors [24:34].
Banning AI, as Italy has done, is considered unwise, as progress cannot be stopped unless all countries agree [26:08]. For businesses, a sustainable model is built on users being aware of and willing to transmit their data, rather than exploiting ignorance [25:18].
For Poland and the European Union, the opportunity exists to lead the artificial intelligence revolution on their own terms [24:44]. This requires wise regulations regarding data, taxes, blockchain technology, and education [29:34]. Such regulations should be agreed upon by citizens and entrepreneurs, not politicized, and should focus on enabling Poles to better utilize this technological miracle [29:44], [30:09].