AWS report: Generative AI surpasses security in global tech spending for 2025

According to a study, global IT leaders will prioritize generative AI tools over cybersecurity in the years ahead. Today, a comprehensive new study was released by Amazon Web ServicesThe Amazon Web Services

AWS Generative AI Adoption Index (), which surveyed 3,739 senior IT executives across nine countries, revealed that 45% of companies plan to prioritize generative AI over traditional IT investments such as security tools (30%). This is a significant change in corporate technology strategy, as businesses race to capitalize AI’s transformative power.

In an exclusive interview with VentureBeat, Rahul Pathak said, “I don’t believe it’s a cause for concern.” “The way I interpret it is that customers’ safety remains a massive prioritiy. We’re seeing that AI is a top priority from a budget perspective because customers see so many uses for AI. It’s the need to accelerate AI adoption that’s driving this particular outcome.”

This extensive survey, conducted in the United States, Brazil and Canada, France, Germany India, Japan South Korea and the United Kingdom shows that generative AI has reached a critical point of adoption, with 90% of organizations deploying these technologies. It is also worth noting that 44% of organizations have already moved from the experimental phase to production deployment.

As AI initiatives spread across organizations, new leadership models are emerging to manage complexity. The report found 60% of organizations had already appointed an AI executive such as a Chief AI Office (CAIO), and another 26% plan to do so by the year 2026.

The executive commitment reflects the growing recognition of AI’s strategic importance. However, the study notes that by 2026, nearly a quarter of organizations will not have formal AI transformation strategies, indicating potential challenges in change-management. The report stresses that “a thoughtful change management strategy is critical.” The ideal strategy should address data management practices, scaling strategies, talent pipelines and operating model changes.

The production gap challenge: Companies conduct an average of 45 AI tests, but only 20 reach users by 2025. This highlights the persistent implementation challenges. Pathak said that he was surprised to see that over 40% of the AI experiments were put into production. “I think this is a pretty rapid and high adoption rate,” he said. “That said, customers are definitely using AI in production at a scale, and I believe we want to see that continue to increase.”

According to the report, talent shortages were the primary barrier for transitioning experiments into the production phase. 55% of respondents cited the lack of a trained generative AI workforce as the biggest challenge.

Pathak told VentureBeat that customers should work backwards to understand what business objectives they are trying to achieve, and then understand how AI will interact with their data. It’s when you combine AI with the unique insights that you have about your company and your customers, you can achieve a differentiated outcome.

In 2025, 92% of organizations plan to hire AI talent while 75% will implement training in order to bridge the skills gap.

Organizations are using both internal training and external recruiting strategies to address the skills gaps. The survey revealed that 56% have already developed generative AI plans, and another 19% plan to do so by 2025. Pathak, speaking about the talent shortage, said that “it’s clear to me that it’s on customers’ minds.” It’s about how we can bring our teams and employees along to a point where they can maximize the opportunity.

Instead of specific technical skills, Pathak stressed adaptability: “I believe it’s more about whether you can commit to sorting out how to use AI to build them into your daily workflow and maintain that agility?” I think mental agility will be very important for us all.

The talent drive extends beyond training and aggressive hiring. 92% of organizations plan to recruit for roles that require generative AI expertise by 2025. In a quarter, at least half of the new positions in an organization will require these skills.

Only 25% of financial services companies are building solutions from scratch.

A hybrid approach appears to be the answer to the long-running debate about whether to build proprietary AI models or leverage existing ones. Only 25% of organizations intend to deploy solutions created in-house, from scratch. 58% plan to build custom applications based on pre-existing model and 55% to develop applications based on fine-tuned model. This is a significant shift for industries that are traditionally known for custom-developed software. The report found 44% of financial service firms plan to use off-the-shelf solutions, a departure from the traditional preference for proprietary systems.

Pathak explained that “many select customers still build their own models.” “I think that the core foundation models are excellent starting points because they have so much capability and investment. We’ve worked hard to ensure that customers can feel confident that their data will be protected. Nothing leaks into models. They can do fine-tuning, customization or model distillation without compromising their IP.

India leads the world in AI adoption with 64%. South Korea is second at 54%. This outpaces Western markets

. While generative AI investments are a global trend but regional variations were revealed by the study. The U.S. had 44% of organizations prioritizing investments in generative AI, which is aligned with the global average rate of 45%. However, India (64%)and South Korea (54%) showed significantly higher rates.

Pathak noted that “we are seeing massive adoption all over the world.” “I found it interesting that there seemed to be a high level of consistency across the globe. I think that if we squint, we can see India slightly ahead of the average, other parts slightly below, and the U.S. in the middle.

65% of organizations plan to rely on third party vendors to accelerate AI implementation by 2025

Organizations are increasingly relying on external expertise as they navigate the complex AI environment. The report found that in 2025 65% of organizations plan to rely on third-party vendors in some capacity. 15% will rely on vendors exclusively, and 50% will adopt a mixed strategy combining internal teams and external partners.

Pathak, AWS’s director of customer support, said that the company’s approach to custom and pre-built solution support is very much a “and” type of relationship. “We want customers to be met where they are. We have a large partner ecosystem in which we have invested from the perspective of a model provider, so Anthropic, Meta, Stability and Cohere. We have a large partner ecosystem of ISVs. We have a large partner ecosystem of system integrators and service providers.”

Act now or risk being left in the dust

Pathak issued a warning to organizations that are still hesitant to embrace AI generative. “I think customers should lean in, or else they risk being left out by their peers.” The benefits that AI can bring are real and significant.

Pathak emphasized the rapid pace of innovation: “The rate at which AI technology is improving and changing, and the rate at which costs like inference are being reduced are significant and will continue be rapid.” In three to six months, things that seem impossible now will be old news.

The widespread adoption of AI across sectors echoes this sentiment. Pathak said, “We are seeing such a rapid and mass adoption.” “Regulated Industries, Financial Services, Healthcare, we see Governments, large enterprises, startups. The current crop is almost exclusively AI driven.”

The Business-First Approach to AI Success

AWS report reveals the rapid evolution of generative AI from a cutting-edge experiment into a fundamental business infrastructure. The data shows that we’ve reached an important tipping point for enterprise AI adoption as organizations shift budget priorities and restructure their leadership teams.

Despite the technological gold rush the most successful implementations are likely to come from organizations who maintain a relentless emphasis on business outcomes instead of technological novelty. Pathak said, “AI is an extremely powerful tool, but it’s important to start with the business goal.” What is your organization trying to achieve?

The companies that will thrive are not necessarily those with the largest AI budgets or most advanced models. They are those that can most effectively use AI to solve real problems with their unique data. In the new competitive landscape, it is not a question of whether or not to adopt AI but rather how quickly companies can turn AI experiments into tangible advantages before their competitors.