The future of AI: 23 industry leaders on what's next in AI for startups

“AI is transforming every organization around the world and represents an unprecedented opportunity to solve complex problems, drive growth, create efficiencies, and open up new business opportunities. This is particularly true for startups, who are moving very quickly to address new market opportunities with AI” - Thomas Kurian, CEO of Google Cloud.

Startups are leading the generative AI charge by experimenting with new applications across industries, pushing the boundaries of what's possible, and advancing the technology's potential. They are the engine of gen AI innovation in a rapidly evolving landscape.

To gain a deeper understanding of just where gen AI might be headed — and the business world along with it — we've gathered perspectives from 23 industry leaders and investors on AI's trajectory and implications for startups in 2025 and beyond. We’re publishing our findings today in Google Cloud’s latest report, the "Future of AI: Perspectives for Startups 2025."

The evolving role of AI agents

Among the trends discussed, AI agents were the leading topic of conversation. The growing innovation of AI leaders with agents comes as no surprise, but their collective vision of how these agentic systems will evolve reveals an intriguing perspective on the increasingly sophisticated interactions we’ll soon encounter and their business implications.

David Thacker, VP of product at Google DeepMind, points out that “ Gemini can also natively use tools like Google Search to access real-time information, and DeepMind's Project Mariner has demonstrated that agents built with the Gemini model can complete tasks using a web browser. Conversational experiences can now be built with the Gemini Multimodal Live API, which accepts audio and video streaming input. The combination of these capabilities enables a new class of agentic experiences and we’re excited to see what startups build with Gemini in 2025.”

The opportunities seem quite expansive. For example, Matthieu Rouif, co-founder and CEO at Photoroom, emphasizes AI becoming increasingly adept at recognizing and responding to human emotions; this could enable AI systems to tailor content and experiences to individual emotional responses in ways that create more meaningful connections. Jia Li, co-founder, president, and chief AI officer at LiveX AI, sees a similar evolution enabling AI agents to move beyond simple task completion to truly understand customer intent and emotional states, offering personalized guidance that feels more human than mechanical.

“To truly leverage agentic systems," says Harrison Chase, CEO and co-founder of LangChain, "I want them to scale beyond just what I can ask them to do — to be ‘ambient agents,’ running in the background, always on, monitoring streams of events, and alerting me only when something interesting happens or when they need my help.” Such intelligence and deep personalization could transform customer experience and business relationships, marking a significant evolution from today's task-oriented AI to tomorrow's empathetic digital companions.

Rethinking AI infrastructure

Yet for all the amazing leaps AI is offering, almost none of it would be possible without the right infrastructure, which is another major focus topic.

"Tight synchronization and massive compute requirements push infrastructure to never-seen-before levels of compute density and capability,” said Amin Vahdat, VP & GM for Systems and Cloud AI at Google Cloud.

The future infrastructure landscape will likely be characterized by modular architectures that combine smaller, specialized models across different data modalities, supported by sophisticated orchestration and observability layers.

“Integrations and infrastructure for observable workflow management will become even more critical parts of the AI stack," Mayada Gonimah, CTO and co-founder at Thread AI, said.
"Companies will need interfaces that let them insert AI in their existing workflows and test them in parallel to existing processes. I think being more intentional around where you embed AI and having the tooling around observability is going to be key moving forward as we build these production-worthy AI-native workflows.”

Infrastructure efficiency will become a critical competitive differentiator, with organizations that achieve twice the efficiency gaining significant market advantages. Arvind Jain, CEO of Glean, emphasizes the importance of designing your infrastructure to be more agnostic to model and tooling advancements. Predict what’s coming and enable plug-and-play where feasible to take advantage of updates without massive overhauls or disruptions.

Cloud providers will continue to play a central role, but successful organizations will build flexible systems that can seamlessly integrate new models and technologies. For a deep dive around into impactful trends like multimodal AI, customer experience, enterprise search, and more technical advancements, check out the full report.

Priorities for investors in AI startups in 2025

The AI investors in the report are zeroing in on startups that deliver concrete solutions to real-world challenges, moving beyond the initial AI hype. Salim Teja, a partner at Radical Ventures, is focused on startups that use AI to address real problems such as improving health, fighting diseases, combating climate change, and addressing the affordable housing crisis through robotics in construction.

Having data alone isn't enough — investors want to see strategic use of high-quality, secure data that enhances AI performance in specific applications. Venture capitalists are particularly drawn to solutions that boost productivity and transform business operations, especially in developer tools.

The winning formula? A clear competitive edge, path to profitability, and deep integration into user workflows. According to Crystal Huang, a general partner at GV, “If your product is easy to implement, it’s just as easy to uninstall. Products need to be stickier to create lasting value, which means being both indispensable and deeply integrated into the user’s workflow.”

Leapfrog the competition

In today's competitive AI landscape, building a defensible market position requires a sophisticated strategy that goes beyond simply implementing the latest models.

Yoav Shoham, co-founder of AI21 Labs, emphasizes the importance of "Product-Algo Fit" - understanding and leveraging current AI capabilities rather than banking on future potential. Rather than waiting for perfect models, organizations must invest continuously in evolution and adaptation, integrating AI solutions that solve real business problems and deliver measurable returns.

David Friedberg, CEO of Ohalo Genetics, put it this way: “If you're just an LLM wrapper, it's going to be hard to build a sustainable business — you're likely going to get commoditized away. Businesses need an engine of value creation that persists with an initial advantaged offering with continuous improvements. This will typically come from proprietary data generation, which is used to continuously improve model performance, or network effects that lock-in access to data or customers.”

Looking ahead

The breakneck pace of change being driven by generative AI means that startups are facing unprecedented challenges. At Google Cloud, we’re collaborating with researchers, founders, startups, investors, enterprises, partners, and public sector agencies to think critically about how our responsible AI solutions can meet the needs of employees, customers, patients, and citizens.

This includes providing world-class infrastructure and full-stack capabilities at the forefront of innovation, engaging deeply with inventors on data, agents, and applications to help bring new outcomes to life, and partnering with enterprises to evaluate how AI advances can modernize experiences both inside and outside organizations.

Our hope is that the Future of AI: Perspectives for Startups 2025 report will shed more light on one of the fastest-moving technologies our industry has ever seen — and in particular, provide exciting guidance to founders who are imagining the next wave of innovative AI startups.

No matter where you are with AI adoption, we’re here to help: Book your generative AI consultation today, get up to $350,000 in cloud credits with the Google for Startups Cloud Program, or contact our Startup sales team.