Once You Go OpEx with Cloud Computing, It’s Hard to Go Back

ETR Insights hosted a panel discussion between technology executives to reflect on the January 2025 Technology Spending Intentions Survey (TSIS) findings. These executives anticipate rising IT spending in 2025, driven by investments in AI, efficiency improvements, and supply chain restructuring. These industry veterans warn that failing to invest in emerging technologies like AI could leave businesses behind. Amid regulatory pressures, some are exploring private cloud solutions. Tools like Copilot and ChatGPT are improving communications and automating tasks, though difficulties handling AI hallucinations persist. RPA remains the gold standard for repetitive tasks, but AI is beginning to extend into more complex processes that require “thinking.” CrowdStrike faces mixed reactions following service outages, with newer security players like Wiz and Darktrace gaining traction. Read on to learn more about hyper- converged infrastructure, AI-driven operational efficiency, cloud repatriation, micro-segmentation in cybersecurity, and more. Plus, listen to a clip from the panel discussion.

Vendors Mentioned: Abnormal Security / Amazon (AWS) / Appian / Arctic Wolf / Automation Anywhere / CrowdStrike / Cyera / Darktrace / Datadog / ExtraHop / Google (Gemini, Google Cloud Platform) / Grammarly / Jasper / Kaseya / Microsoft (Azure, Copilot, Power Automate) / New Relic / Okta / OpenAI (ChatGPT) / Oracle (Oracle Cloud Infrastructure) / Salesforce (MuleSoft) / SentinelOne / Snyk / SolarWinds / UiPath / Wiz

Panel Highlights

IT spending is on the rise for 2025 as companies prioritize artificial intelligence and efficiency improvements; executives from various industries report anticipated spending increases of up to 10%, reversing the last few years of cautious budgets. “I think that at first companies and people are a bit reserved and apprehensive about their spending, in our political climate and current world events,” says one CIO at a cybersecurity startup. “However, I think that's the wrong way to approach the challenge, and I think whoever doesn't pivot or increase spending will be left behind.” A VP from a large tech enterprise also imagines that political resolution will likely unleash spend. “It's a little clearer the direction we're going in. That’s caused many businesses to open up their wallets and increase their spend on IT.” Manufacturing companies, facing potential tariff impacts, are restructuring supply chains by shifting production away from China and increasing local inventory.

AI will be critical to efficiency and long-term competitiveness, and industry leaders say the pressure to invest in AI has become too significant to ignore. “Streamlining processes and becoming a more efficient company, enterprise, or entity will [necessarily] involve some sort of investment into the AI realm.” These executives are focusing on automation, operational efficiency, and customer engagement, overriding any previous hesitancy to boost spending without clear returns. “A lot of organizations were basically stuck with the Hobson's choice of, if you want to invest more in this, you're going to have to decelerate your spending somewhere else. [However], I think there's been a realization that this is a game changer, and that investment is required.” Cloud computing, operational consolidation, and ongoing investment in security round out current trends; these executives are now fully committed to advancing technology initiatives with long-term benefits.

Cloud computing and possible repatriation. The three leading cloud providers—AWS, Microsoft Azure, and Google Cloud Platform—recently reported earnings misses, citing surging demand for AI-related usage that outstrips current supply. In response, these firms plan to increase capital expenditures. Our guests, however, wonder about possible cooling demand for public cloud services. One VP mentions that at their organization, they are actually ramping up spend in private cloud. “We’re actually making the capital investments ourselves.” Another points to increased regulatory pressure within industries like finance – “You don't have to worry about telling regulators that it's floating around the cloud and you can't explain where it is” – and in “red box countries,” those with stringent data localization requirements and high-risk data protection and privacy laws. “We’re also seeing hyper-convergence [infrastructure], where you can run kind of your own cloud version of Azure, or even an Azure stack or something like that.” Some companies may have rushed to the cloud without refactoring applications, which is now proving prohibitively expensive. However, despite recent media attention, some view cloud repatriation as overhyped; one Director of IT on the panel dismissed the notion outright. “We've already convinced the business that the OPEX model is what's working. Once you go to the OPEX model, I don’t think you’re going back.” Another CIO theorizes that these earnings misses reflect a market recalibration rather than a downturn. “It’s more of a healthy pause in growth before another leap.”

ETR Data: According to the January 2025 Macro Views Survey, the majority of respondents (72%) had no plans to move public cloud workloads back on-premises, with only 15% indicating they planned to do so.

AI and automation. These executives are using tools like Copilot, Gemini, and ChatGPT to enhance communications, automate tasks, and streamline operations. “AI can’t drive a forklift, and AI can’t pack and unpack a container, but AI can count what's in the container, which saves an individual a lot of time, or how many SKUs I have on the shelf or things like that. I think there are real use cases, and we see real value with that.” While these applications have reportedly improved organizational quality and modestly reduced headcount, formalizing ROI remains difficult, and ongoing hallucinations limit deployment. “Copilot's really great for simple things like writing an employee review, but when it comes to the large-scale stuff, we've kind of underestimated how much the hallucinations count.” Customer- facing applications can demonstrate clear a ROI via new revenue, but measuring internal efficiency is more complex. “It's fairly easy to measure the output of a developer who's using this technology versus a developer who's not using the technology. But it's a lot harder to measure the productivity of someone who's using Microsoft Copilot to create emails and PowerPoint presentations.” A Director of IT in industrials and manufacturing agrees: “Real-world environments are what matter. You have to balance the real use cases with sort of the fluff.”
 
Despite AI’s growing capabilities, RPA remains entrenched in many business operations. “67% of the people that either call in or do chat are looking for tracking information, so you really want to automate a number of those things,” says one director. “The problem is, when we're talking about advanced functions and problem-solving with AI, those language models have to be designed, and they're very different for my organization compared to someone who sells handbags.” AI conversations can feel unnatural, and voice recognition still leaves much to be desired. “I think people are leery to move into new solutions until they're actually vetted as being better than what they currently have. I think you have a lot of solutions that are maybe marginally better, but until we get to a point where it's more conversational and more natural-flowing for natural language, there won't be a shift in the spending.” Until AI can consistently deliver natural, human-like interactions, RPA tools like UiPath and Power Automate remain the more practical choice.
Another executive with “dozens and dozens of bots in production” handling recurring, straightforward processes, acknowledged the potential for AI to expand RPA’s reach into more complex decision-making tasks. “With the AI enhancements that UiPath is rolling into its product, we might be able to move up the chain and start to automate processes that are a little more complex and require choices or ‘thinking,’ which actually may increase our usage of RPA because it allows it to expand it into more complex use cases.”

ETR Data: In the January 2025 Technology Spending Intentions Survey (TSIS), Microsoft Power Automate continues to lead the Robotic Process Automation sector in Net Score by a considerable margin. Appian and UiPath stand in second and third place in Net Score, and both have seen growing spending momentum in the past quarter.

The state of security. CrowdStrike is navigating mixed reactions from users following a recent outage that disrupted many businesses. While some of our panelists are reevaluating their relationship, others remain steadfast in their support. “We did stay up all night and get our Windows servers back up, so that it did not impact our customers or our employees. But it's one of those vendors, I call them ‘anchor tenants.’ I have a few vendors that I really trust and I buy multiple products from. CrowdStrike is one of them. I'm willing to take their word that they've improved their update vetting process.” Others are reluctant to abandon a well-established platform, weighing the costs of migration against the benefits of staying. “There's a hard cost, and then there's a soft cost. I can switch to a SentinelOne, I can switch to an Arctic Wolf, or whatever that may be, but how much money do I have to spend to get there? I have to remove this off of systems, I have to vet it, and I have to train it so our minor nuances for different things that we do don't affect production.” That said, discontent with CrowdStrike’s initial response to the outage is evident. “They left a lot to be desired there. Our customers are definitely trying to get rid of it. We've seen a lot of growth in Datadog and SentinelOne. We use even ExtraHop. It's kind of like one of these things, too, where they’re kind of the latest newsworthy thing. It’s kind of like a love/hate.” Newer security vendors are gaining traction; Wiz is an industry standout. “We were an early Wiz customer, and it's been a fantastic journey with them. The product is just super easy to implement and manage.” One CIO particularly likes Darktrace’s proactive responses: “They prevented definitely a few incidents, let’s put it that way.”

Key priorities for 2025. Our panelists signed off with three key priorities: accelerating cloud migration, investing in AI tools and training, and modernizing infrastructure to replace end-of-life systems. Hyper-converged infrastructure, multi-tenant cloud environments, and advanced AI implementations lead IT spending agendas. “We're getting to a point we’ve spent capital expenditures on what's left in our hybrid environment, and we can go one of two ways - we make that level of investment, or we could spend the millions of dollars it would take even to refactor or replace applications, and the amount of time that it would take. It’s a push-pull from the lines of business to IT; we still need to make sure that we run the business and that we have a reliable source of storage and replication.” Strengthening cybersecurity—particularly through micro-segmentation and API security—remains equally top of mind, along with agentic AI implementations. Much more to come from ETR on both!