Analysts at RBC Capital Markets have provided an early assessment this week, and it is not particularly positive. In a research note, they stated that "due to GenAI, long-term software gross margins will be structurally lower. " According to RBC, the shift from on-premise software, where companies ran it on their own computers, to the cloud resulted in a decrease in gross profit margins from 90% to 75%. The transition from cloud computing to generative AI is expected to push software margins even lower, estimated at around 60%. Gross margin is a straightforward measure of profitability that takes revenue and subtracts the cost of goods sold. Traditionally, in the software industry, gross profit margins have typically hovered around 90%. While this may sound high, it is why this sector is appealing to investors and why software companies have such high valuations. Developing new software requires a significant upfront investment. However, once it is created, the cost of producing new versions and distributing them to customers is nearly negligible. Consequently, every time software is sold, profits increase substantially, hence the emphasis on "scale" by tech investors. So why might the software business become less profitable in the AI era?According to the RBC analysts, achieving similar levels of efficiency on the profit and loss statements with GenAI may prove challenging.

Generative AI is expensive, both in terms of development and ongoing operation. AI model training, for instance, necessitates the purchase of expensive GPUs from Nvidia. These AI chips are then placed in servers requiring specialized cooling and networking within large data centers. These facilities consume substantial amounts of electricity, leading to high costs and costly upgrades. Furthermore, the cost of acquiring the data required for AI model training cannot be ignored. Although big tech companies and startups attempt to minimize these expenses, data collection and cleaning remain significant costs. Once the AI models are trained, they need to be deployed. This inference step, where the models process new data or requests, also requires expensive chips and incurs ongoing expenses. In contrast to the on-premise software business, where each sale resulted in almost 100% profit, with GenAI, every time an AI customer utilizes a service, the provider incurs numerous costs. For instance, OpenAI reportedly spends $700, 000 per day to operate ChatGPT, as estimated by industry analyst Dylan Patel last year. Despite these challenges, the RBC analysts are not entirely pessimistic. They anticipate that GenAI will revolutionize the market to such an extent that customers will significantly increase their spending on new AI-powered software, potentially doubling or even tripling current revenue levels. Considering the expanded software market, there may also be more "profit dollars" available, even with lower profit margins. "Profit dollars" refer to the absolute profit generated by a company and can serve as a measure of financial success. To illustrate, let's consider a hypothetical company with $100 million in revenue and 10% profit margins, resulting in $10 million in absolute profit. If this company experiences a revenue surge to $300 million but sees the margin decline to 8%, the income would still amount to $24 million—more profit than before. Consequently, the RBC analysts conclude that while GenAI is expected to impact margins, long-term gross profit dollars are projected to be higher in a post-GenAI world.