Shawn:
You know, I’ve been thinking about the arguments made by that experienced Wall Street tech equity research professional regarding AI infrastructure investments. He’s very skeptical about the $1 trillion in capex expected in the coming years, especially the idea that AI will solve complex enough problems to justify that kind of spend.
Emma:
I remember that analysis. He compared AI to the early internet, where low-cost solutions quickly displaced expensive incumbents. His concern was that AI doesn’t have the same immediate, cost-effective impact. It’s an interesting comparison, but I think it might be overlooking some long-term value.
Shawn:
Exactly. I think his timeframe is too short. AI infrastructure isn’t like the early internet or e-commerce, where the value became clear quickly. With AI, we’re talking about long-term infrastructure that continues to provide value well beyond a 3-5 year depreciation period. We’ve already seen advances like NVIDIA’s structured sparsity, which optimizes AI inference by pruning unnecessary parameters, lowering energy costs, and increasing throughput.
Emma:
That’s a good point. These aren’t one-off investments. AI is going to keep evolving, and the infrastructure we build now will serve enterprises for years. The idea that generative AI doesn’t offer cost advantages over human labor also feels off. AI’s scalability, elasticity, and its ability to operate 24/7 without fatigue makes it fundamentally different. Techniques like model quantization are already reducing computational requirements and improving efficiency.
Shawn:
Right, and let’s not forget that AI’s strength lies in its ability to process massive amounts of data quickly, something that humans simply can’t do. So yes, it’s expensive now, but the cost is coming down as the technology becomes more efficient.
Emma:
Agreed. In fact, Marc Benioff from Salesforce recently said something along those lines. He acknowledged that many CEOs haven’t yet seen the full value of their AI investments, but tools like Einstein and Agentforce are already delivering results. Einstein is handling over a trillion predictive and generative transactions per week, and Agentforce has a 90-95% accuracy rate in resolving customer service issues. That’s not insignificant.
Shawn:
That’s huge. I mean, those are the types of real-world outcomes that the experienced Wall Street professional argued were missing. But Benioff is showing that AI is already solving complex business problems. Companies like Salesforce have integrated AI across customer touchpoints, marketing, and analytics—generating both cost savings and new revenue streams.
Emma:
Exactly. And then there’s Bill McDermott over at ServiceNow. He made a similar point. McDermott mentioned that generative AI is moving beyond infrastructure into real-world applications. ServiceNow’s AI is deflecting 90% of customer service cases, which drastically reduces the time and cost needed to resolve issues. That’s exactly the kind of efficiency boost enterprises need.
Shawn:
It’s a clear counterpoint to the idea that AI lacks cost-effectiveness. McDermott’s example shows that AI isn’t just a flashy tool—it’s fundamentally changing how businesses operate, making them leaner and more productive.
Emma:
And it’s not just about cost savings either. AI is enabling companies to scale faster without having to significantly increase their workforce. AI is automating low-level tasks, freeing up human employees to focus on higher-value activities. That’s something the Wall Street professional might be underestimating.
Shawn:
Agreed. And it’s not just Salesforce and ServiceNow—Microsoft and Amazon have similar takes. Microsoft’s Kevin Scott compares AI to the early days of cloud computing, where upfront costs were high, but the long-term platform it created was invaluable. He sees AI evolving in the same way, as a foundation for future innovation.
Emma:
That’s a great analogy. Cloud computing wasn’t cheap at first either, but look where we are now. Every major business relies on it. AI will likely follow that same path, becoming a critical platform that businesses build on. Microsoft’s tools like Copilot are already integrating AI into workflows to make specific tasks more efficient, lowering costs while improving accuracy.
Shawn:
And then there’s Amazon. Matt Garman mentioned how AI is driving efficiencies in sectors like transportation, where predictive models are reducing downtime for Japan’s bullet trains. AI is literally keeping things running smoothly. It’s already solving real-world problems today, not just in the future.
Emma:
That’s a perfect example. AI’s ability to analyze vast amounts of data and predict issues before they occur is a huge cost saver, especially in industries like transportation. It’s hard to argue against those kinds of concrete results.
Shawn:
Right, and while the Wall Street professional worried about AI’s commoditization, I think Microsoft and Amazon actually see that as an advantage. Kevin Scott pointed out that as AI becomes commoditized, it’ll spur more innovation. It won’t erode AI’s value, it’ll democratize it, making it accessible to more companies who can then build on top of it.
Emma:
Exactly. Commoditization will fuel the creation of new applications, just like with the cloud. Matt Garman emphasized that as AI infrastructure becomes more widespread, it enables companies to innovate faster, creating new use cases and driving long-term growth.
Shawn:
So the idea that commoditization diminishes AI’s potential feels shortsighted. Instead, it’s going to unlock more possibilities. And honestly, we’ve only scratched the surface of what AI can do. As it continues to advance, the problems AI can solve will only get more complex and valuable.
Emma:
I agree. The experienced Wall Street professional’s concerns about AI’s high cost and commoditization are valid in the short term, but the long-term view shows a different picture. Companies like Salesforce, ServiceNow, Microsoft, and Amazon are already proving that AI is delivering real value. The infrastructure being built now is laying the foundation for future innovations that will far surpass the initial investment.
Shawn:
Absolutely. AI isn’t just a short-term expense—it’s a transformative investment. We’re already seeing tangible results, and as the technology matures, it’s going to become even more integral to business strategy. The companies investing in AI infrastructure today are positioning themselves for long-term success.
Shawn:
In the end, while that experienced Wall Street professional’s skepticism focuses on the short-term costs and ROI of AI, the combined perspectives from Salesforce, ServiceNow, Microsoft, and Amazon paint a much more optimistic picture. AI is already delivering tangible results, from automating customer service to optimizing workflows. As AI continues to evolve and become more cost-efficient, it will unlock even greater potential for businesses to innovate, scale, and grow.
Emma:
Right. AI infrastructure isn’t just an expense, it’s an investment in the future. The companies that understand this are the ones that will lead the next wave of innovation.