Lovable, Legora, and Tandem Health have impressed with strong growth and valuations in the billion-class, as reported by major media. However, for many AI sensations like these, about half of gross revenues go to purchasing the underlying AI power from actors such as OpenAI and Anthropic, according to multiple reports. According to Dagens Industri, Jarek Kutylowski, founder of German Deepl, described a risk of an AI bubble because American companies are subsidizing the application layer.
The exact revenue percentages or financial details behind the claim that half of gross revenues go to AI power purchases remain unclear, and it is uncertain whether the risk of an AI bubble is imminent or speculative, or what evidence supports it. Billions are rolling in to the Swedish AI successes Lovable and Legora, multiple reports indicate, and everything seems to be pointing upward for these firms. Yet large risks are now visible for Lovable and Legora, and they do not come from competitors but from significantly closer and more sensitive quarters, according to major media.
Of course there is a risk of an AI bubble because the American companies are subsidizing the application layer.
What specific 'closer and sensitive' risks Svenska Dagbladet refers to for Lovable and Legora has not been detailed. How the valuations of Lovable, Legora, and Tandem Health compare to industry benchmarks or sustainability metrics is also unknown, as is the competitive landscape and how these Swedish AI companies differentiate themselves from global players like OpenAI and Anthropic.
