AI’s Progress Shows Signs of Both Breakthroughs and Barriers

The latest artificial intelligence models are delivering mixed signals about the technology’s pace of advancement, with some showing dramatic leaps forward while others hit unexpected roadblocks.

The complex picture emerges at a crucial moment for businesses navigating AI investment decisions, as researchers debate whether recent achievements — from advances in scientific reasoning to struggles with basic reliability — reveal an innovation curve that defies simple characterization as either accelerating or stalling.

“AI isn’t hitting the brakes; it’s just outgrowing the public data pool,” Eugina Jordan, chief marketing officer of the TIP Telecom Infra Project, told PYMNTS. “While we’ve pretty much wrung every drop of insight from external data (and sometimes maybe without data owners knowing), the real magic is sitting inside private organizational data. Businesses can fine-tune AI models with their own treasure troves, making tools smarter and hyper-relevant.”

Google Chief Sees Steeper Climb for AI Innovation

Sundar Pichai, CEO of Google and Alphabet, recently acknowledged at The New York Times DealBook summit that while AI development isn’t hitting a wall, progress will likely face increased challenges in 2025 as the industry moves beyond “low-hanging fruit.”

While disputing claims of a performance plateau and maintaining that Google’s AI models are superior to Microsoft’s, Pichai said future breakthroughs will require more substantial advances in reasoning and reliable action sequences. Despite the billions invested in AI, Pichai noted that the technology has yet to generate significant profits for corporate users and will require more profound technical innovations to achieve real progress.

Kelwin Fernandes, co-founder and CEO of NILG.AI, told PYMNTS that progress in artificial intelligence is accelerating and not slowing.

“The devil is in the details,” he said. “Those small percentage points on performance we’ve been getting in the last months enable a vast number of new use cases. We’re now tackling more modalities than a year or two ago. Plus, the development of UI-interacting AI agents could potentially transform the whole software market.”

Fernandes pointed to a development that highlights a quickening pace in artificial intelligence. Google’s Gemini 2.0, a next-generation AI model, introduces “agents” said to be capable of handling tasks like real-time research and online shopping with minimal human help.

Meanwhile, OpenAI has unveiled Sora, a text-to-video creation tool, though its current capabilities lean more toward refining video editing workflows than fully autonomous production. Apple, not to be left behind, has added AI features to its latest iOS update, including tools for creating custom emojis and AI-driven illustrations.

These innovations are part of a broader race among tech giants to develop smarter, more practical artificial intelligence tools. While the technology is far from flawless, its rapid evolution suggests a push toward making AI more accessible and integrated into everyday life. The competition is heating up, and companies are pouring resources into delivering faster, more capable systems.

AI Evolution Creates Wait-and-See Mindset 

Given the rapid pace of AI evolution, some retailers are concerned about early investment in AI, Marios Savvides, chairman and CTO at UltronAI, told PYMNTS.

“Yet those that approach AI with clear expectations and a pragmatic plan will gain not only immediate benefits but also knowledge and expertise that will continue to position them ahead of the curve as the technology advances,” he said. “These early adopters will build a foundation for long-term success, leveraging AI to continuously refine operations and enhance customer experiences.”

As companies prove their AI tools can deliver real results, they can charge premium prices, Savvides said.

“Yet over time, AI products are becoming more similar and widely available, putting pressure on vendors to lower their prices,” he said. “This tension between value-based pricing and market competition is reshaping how much businesses pay for AI solutions.”