They’ve Raised Billions. But Can AI Start-Ups Make Money?

Estimated reading time: 0 minutes

Expensive Models

AI companies have promised to revolutionize the business world. Now it’s time for them to show their receipts.

For start-ups, the problem is that building and running AI models requires vast amounts of energy, and is very, very expensive. OpenAI CEO Sam Altman said its latest publicly available model, GPT-4, cost more than $100 million to train. Anthropic CEO Dario Amodei said the industry could see a $10 billion model by 2025.

And while it’s cool that ChatGPT can summarize a boring research paper or write you a love letter in the style of Warner Herzog, those use cases are not going to drive revenue.

Monetizing AI

OpenAI makes money in several ways. First, it sells subscriptions to individuals, who can pay $20 a month to use ChatGPT Plus. It also provides services to developers and businesses, both directly and through its partnership with Microsoft (MSFT).

Google (GOOGL) has a similar business model, charging users $20 monthly for subscription access to Gemini Advanced and licensing its technology to enterprise users. It has also integrated AI into its own products.

For start-ups without the resources of a tech giant, it can be harder to provide a timely return on investment.

Under Pressure

Anthropic, which makes ChatGPT competitor Claude 3, is trying to fill a $1.8 billion gap between sales and expenses, according to the New York Times. At least it has investors with deep pockets, including Amazon (AMZN) and Google, who have poured billions into the company.

In March, Microsoft acquired most of the staff of Inflection AI. Another once-hot start-up, Stability AI, laid off employees last month after its CEO resigned.

AI start-ups need businesses to adopt their technology to be profitable, but “hallucinations,” when AI presents false information, and copyright and security concerns can make enterprise customers wary. Implementing AI can also be tremendously expensive and especially so if a return on investment can’t be proven.Last year, the average round of funding for AI start-ups was $51 million, far above the tech industry average of $8 million, according to CB Insights. And the cost of training competitive AI models is only going to increase. If more enterprise clients don’t begin to see the technology as worth the investment, start-ups could have a progressively harder time competing with Big AI.

Disclaimer
SoFi Securities (Hong Kong) Limited and its affiliates (SoFi HK) may post or share information and materials from time to time. They should not be regarded as an offer, solicitation, invitation, investment advice, recommendation to buy, sell or otherwise deal with any investment instrument or product in any jurisdictions. Keep in mind that investing involves risk, and past performance of an asset never guarantees future results or returns. It’s important for investors to consider their specific financial needs, goals, and risk profile before making an investment decision.
SoFi HK does not make any warranties about the completeness, reliability and accuracy of this information and will not be liable for any losses and/or damages in connection with the use of this information.
The information and materials may contain hyperlinks to other websites, we are not responsible for the content of any linked sites. The information and analysis provided through hyperlinks to third party websites, while believed to be accurate, cannot be guaranteed by SoFi HK. These links are provided for informational purposes and should not be viewed as an endorsement. The risk involved in using such hyperlinks shall be borne by the visitor and subject to any Terms of Use applicable to such access and use.
Any product, logos, brands, and other trademarks or images featured are the property of their respective trademark holders. These trademark holders are not affiliated with SoFi HK or its Affiliates. These trademark holders do not sponsor or endorse SoFi HK or any of its articles.
Without prior written approval of SoFi HK, the information/materials shall not be amended, duplicated, photocopied, transmitted, circulated, distributed or published in any manner, or be used for commercial or public purposes.