I’m tracking the warnings about chip and memory shortages. A global shortage of RAM is expected to persist into 2027, resulting in rising prices for smartphones and PCs. The International Data Corporation reports that major memory manufacturers, including Micron, Samsung, and SK Hynix, are prioritizing supply for artificial intelligence companies, impacting pricing across various consumer electronics. According to the IDC, smartphone prices could increase by 3 to 5 percent, or even 6 to 8 percent under a pessimistic scenario, as manufacturers like Xiaomi and Honor pass on higher costs to consumers. Meanwhile, PC makers such as Dell and HP are also preparing for price hikes, with average PC prices projected to rise by 4 to 6 percent, or 6 to 8 percent in a worse-case scenario, amidst a shift towards memory-intensive AI technology.
Global spending on cloud infrastructure services reached $102.6 billion in the third quarter of 2025, reflecting a 25% year-on-year growth, according to research from Omdia. This marks the fifth consecutive quarter of growth exceeding 20%, driven by increasing enterprise demand for artificial intelligence, which is transitioning from early experimentation to large-scale deployment. Amazon Web Services, Microsoft Azure, and Google Cloud maintained their positions as the top three cloud providers, collectively accounting for 66% of global spending. AWS experienced a 20% year-on-year growth, while Microsoft Azure and Google Cloud reported over 35% growth each. As enterprises seek to implement AI applications more broadly, hyperscalers are shifting their strategies to support multi-model deployment and enhance platform capabilities, emphasizing collaboration across the ecosystem to meet the evolving needs of businesses.
In 2026, the landscape of software pricing is shifting from traditional models to outcome-based frameworks, where companies may pay only for the actual results delivered by the software. According to insights from West Monroe, this transition could lead to the end of per-seat licensing as artificial intelligence increasingly drives software usage and purchasing decisions.. With 12% to 15% of enterprise IT budgets now allocated to AI, companies that develop AI capabilities enhancing customer experiences are expected to see improved renewal rates and profit margins.
Why do we care?
Memory prices are going up because AI companies are willing to pay more. That’s not a supply accident—it’s a market choice. SMBs don’t get a vote in that. They just see more expensive laptops, slower refresh cycles, and systems that struggle under heavier workloads.
Now add cloud. Hyperscalers are growing fast, and they’re doing it by encouraging more consumption, more models, more services. Every one of those layers adds cost sensitivity. When something spikes, the MSP is the one explaining why the bill changed.
Outcome-based pricing sounds like the safety net. It isn’t. It shifts the argument from “how many seats” to “did this work,” and that argument almost always drags the service provider into the middle. If the AI didn’t deliver the outcome, was it the data, the model, the workflow, or the infrastructure? Guess who gets asked to sort that out. If that question isn’t answered contractually in advance, the default assumption is that it’s the MSP’s problem.
This matters now because if MSPs don’t reframe their role—from implementer to risk manager—they’ll be underwriting decisions they didn’t make. The harm isn’t theoretical. It shows up as margin erosion, contract disputes, and customers who think technology failed when the real issue was economics.
Ignoring that connection is how good operators get trapped in bad deals.