Nvidia's Nov. 17 earnings may shift AI spotlight to cost-control software

There is a contrast between the headline-grabbing infrastructure buildout and the less visible software layer that manages its costs. Nvidia's Nov. 17 earnings and the Anthropic IPO are presented as the events that will reveal whether the current pace of AI spending holds, but the piece argues the more interesting action may be in companies that sit where bandwidth, latency and cloud bills collide.

Fastly is offered as one example: an edge cloud platform that processes, secures and delivers high-intensity digital traffic closer to users to improve performance and manage delivery costs. Its edge network sits exactly at that intersection, and an investor narrative describes edge computing as processing data in localised servers rather than sending it to a central location.

TXT e-solutions is the second example, with Smart Solutions and analytics platforms that touch performance, efficiency and monitoring in complex, data-heavy environments — described as where AI infrastructure costs and model behavior become hardest to control. The piece also notes its revenue mix: about €258 million from Software Engineering, €99 million from Smart Solutions and €76 million from Digital Advisory services. If the Nvidia and Anthropic catalysts raise doubts about the buildout, attention could pivot to these quieter names, though it is not asserted that it will.