Global spending on generative AI (GenAI) is set to hit a massive $644 billion in 2025, a 76% jump from last year, according to Gartner’s latest forecast. That’s despite some early letdowns, like proof-of-concept projects that didn’t quite deliver. But investment isn’t slowing down.

In fact, it’s ramping up, thanks to better foundational models and the growing presence of AI in everyday tech.

Interestingly, the report shows a big shift in how companies are approaching GenAI. Instead of building their own solutions from scratch, many CIOs are now leaning on commercial software that already includes GenAI features.

And with hardware making up about 80% of total GenAI spending, devices like AI-ready smartphones, laptops, and servers are doing most of the heavy lifting.

Why It Matters: This forecast signals a significant transformation in IT and consumer markets, with GenAI poised to become an embedded and standard part of technology ecosystems. It also underscores the growing reliance on off-the-shelf AI capabilities over custom development, a trend likely to reshape how enterprises adopt and scale AI.

• Hardware is where most of the money’s going: Nearly 80% of all GenAI spending in 2025 will go toward hardware. That means companies and consumers alike will be buying more AI-powered devices, whether they’re actively looking for AI features or not. Smartphones, PCs, and servers with embedded AI capabilities are quickly becoming the default. Even if consumers aren’t asking for GenAI features directly, they’ll be getting them built-in as standard, kind of like how every new phone eventually just came with a camera.

• CIOs are pivoting away from DIY GenAI builds: Many companies that launched ambitious GenAI projects in 2024 are now stepping back to reevaluate. After running into issues like underwhelming results or costly failed pilots, CIOs are turning their focus to off-the-shelf GenAI tools that are already integrated into commercial software platforms. The idea is to reduce complexity, avoid risk, and deliver real business value faster, without having to reinvent the wheel.

• Model quality is improving fast, even if users aren’t wowed yet: Foundational model providers are still pouring billions into making GenAI systems more powerful, reliable, and scalable. That’s happening behind the scenes, even as users express mixed feelings about GenAI’s performance today. Gartner says we’re in a “paradox” where expectations might be cooling, but investment and capability are still growing fast. This gap will likely persist into 2026.

• Consumer demand isn’t necessarily driving the trend, but manufacturers are: Gartner notes that consumers aren’t specifically asking for GenAI in their devices. But they’ll end up with it anyway, as hardware makers make AI capabilities standard. This is less about chasing a flashy new feature and more about manufacturers reshaping the baseline of what “normal” devices can do in the AI era.

• Every major GenAI market segment is growing quickly: GenAI spending isn’t just going up in one area, it’s expanding across the board. Services will grow by over 160%, software by 94%, and devices will nearly double in 2025 alone. Even server spending, which is already huge, is set to grow by 33%. It all points to a broad and deep investment wave across the entire tech ecosystem.

Go Deeper -> Gartner Forecasts Worldwide GenAI Spending to Reach $644 Billion in 2025 – Gartner