BlockBeats News, July 20th: Market analyst qinbafrank stated in a post that merely releasing more large models, disclosing higher Token usage, or continuing to increase capital expenditure is no longer sufficient to reverse the current market's concerns about AI investment returns. What the market truly anticipates is a fundamental evidence chain that can mutually confirm, including continuously expanding AI revenue sources, increasing number of enterprise customers, production-level workloads, paid seats, renewal rates, and synchronized growth in single customer spending. Even after excluding a few large customers such as OpenAI and Anthropic, overall demand is expected to remain robust. Additionally, AI's incremental gross profit must continue to outpace the growth of additional costs such as depreciation, energy, network, and personnel, while the unit cost of inference decreases, allowing the total gross profit to continue to rise.

Furthermore, the market will also focus on whether AI order backlogs can accelerate the conversion into revenue for the next 12 to 24 months, rather than just signing long-term contracts; the utilization rate after the launch of new computing power, whether customer consumption exceeds the initial commitments, and whether in-house chip and model optimization can bring verifiable economic benefits will also become key indicators. In comparison to merely claiming that self-developed chips such as Trainium, TPU, or MAI have lower costs, investors would rather see an increase in the proportion of self-developed chips, a continuous decrease in cost per token or per task, an increase in gross profit per accelerator-hour, and synchronized growth in customer consumption and platform add-on revenue. Additionally, enterprise customers need to demonstrate the true return on investment (ROI) of AI, including revenue growth, improved conversion rates, reduced labor costs and processing time, and the smooth transition of projects from pilot phases to large-scale production deployment.

As for the upcoming second-quarter earnings reports of major tech companies, the market is most looking forward to the "Blonde Girl" combination: Cloud and AI revenue exceeding expectations, maintaining a relatively stable profit margin, Backlog starting to accelerate conversion into revenue, capital expenditures not showing unexpected expansion, or new investments being contractually supported by demand, while free cash flow deterioration has stopped. On the contrary, if Token usage and model calls continue to grow, but single customer spending, gross margin, and free cash flow show no improvement, remaining deliverables are still reliant on a few strategic customers, and capital expenditures are substantially increased again while revenue and margin guidance are revised downwards, then the market's concerns about AI investment returns may further intensify.