The global IT sector is poised for a 9% increase in spending in 2025, largely fueled by investments in artificial intelligence and cloud computing, according to a report by S&P Global Ratings. While major cloud providers such as AWS, Microsoft (NASDAQ:MSFT) Azure, and Google (NASDAQ:GOOGL) Cloud continue pouring billions into AI development, businesses remain hesitant about large-scale AI implementation. Many enterprises are still in the early stages of determining practical applications for AI, balancing potential returns with concerns about security, legal risks, and explainability. The latest findings indicate that although spending on AI infrastructure is rising, companies are taking a cautious approach to deployment.
A similar trend was observed in previous years, with IT spending increasing by 8.3% in 2024. However, the current focus on AI and cloud services has intensified compared to past reports, where emphasis was more evenly distributed across various technology sectors. One key difference in 2025 is the emergence of cost-effective AI models, such as DeepSeek’s R1, which challenges the traditionally expensive AI training methods used by U.S. firms. While previous predictions claimed AI adoption would accelerate, this year’s data suggests a more measured and selective approach by enterprises. Additionally, trade policies and tariffs are now playing a bigger role in shaping IT spending behavior than in prior years.
Are Businesses Ready to Fully Embrace AI?
Despite rising AI investments by cloud providers, many enterprises remain cautious about integrating AI into their operations. According to S&P’s Technology Director Christian Frank, most organizations are still testing AI in proof-of-concept stages rather than rolling out large-scale implementations.
“Enterprises are still figuring out use cases that can deliver good ROI. Explainability, security, and legal risks remain key challenges,”
Frank explained. While broader AI adoption is expected in 2025, businesses are proceeding carefully to avoid unforeseen risks.
What Does DeepSeek’s Model Mean for AI Costs?
Chinese AI startup DeepSeek has introduced an AI foundation model that costs significantly less to train compared to its U.S. counterparts. The company spent just $5.6 million on training, utilizing 2,048 slower Nvidia AI chips while achieving performance levels comparable to leading American models. Unlike most U.S. firms, DeepSeek has opted for an open-source approach, allowing free use and modification of its model. While its cost-efficient strategy could lower the financial barriers to AI adoption, Frank remains skeptical about its immediate impact, stating,
“We need additional confirmation points to validate this hypothesis and its potential impact on global IT spending patterns and competitive dynamics.”
Rather than reducing overall investments, cost-efficient AI models could actually drive higher spending by cloud giants. The concept aligns with the Jevons paradox, which suggests that increased efficiency results in greater consumption. Companies such as Microsoft believe that lower AI costs could lead to wider generative AI adoption. IBM CEO Arvind Krishna also expressed confidence in this trend, saying,
“AI adoption will explode as costs drop.”
OpenAI, in collaboration with SoftBank, Oracle, and MGX, has already announced massive infrastructure investments between $100 billion and $500 billion to expand AI capabilities.
Trade policies and tariffs could also influence IT spending patterns. A new 10% tariff on Chinese imports took effect in early February, while higher tariffs on Mexican and Canadian goods were temporarily delayed following agreements with the U.S. S&P notes that enterprises may accelerate hardware purchases, such as PCs, to avoid potential price hikes. PC sales are projected to rise by 3% in 2025, up from 1% in 2024, partly due to aging systems from the COVID-19 era and Microsoft’s discontinuation of Windows 10 support.
Although AI features are being incorporated into PCs and smartphones, they have not led to a significant boost in hardware sales. Consumer interest in AI-enabled devices remains subdued, with enterprises and individuals still evaluating the benefits before making large-scale investments. Frank commented on this trend, stating,
“Despite the introduction of new AI features on many flagship smartphones, consumer demand has been relatively subdued.”
Manufacturers continue to introduce AI-powered products, but widespread adoption is yet to be seen.
The projected 9% increase in IT spending for 2025 highlights the growing financial commitment to AI and cloud services. However, caution persists among enterprises as they weigh the long-term benefits and risks of AI adoption. While cost-efficient models like DeepSeek’s could reshape AI economics, it remains uncertain whether they will significantly influence global spending trends. Additionally, geopolitical factors, including tariffs and trade policies, add another layer of complexity to IT investment decisions. Businesses will likely continue navigating these challenges carefully, balancing innovation with financial and regulatory considerations.
