Agentic AI innovation is unprecedented, yet most companies are still in experimentation mode
To meet the expected AI demand by 2030, $2 trillion in annual revenue is crucial to fund the necessary computing power. However, despite the savings generated from AI initiatives, there remains an $800 billion gap to keep pace with this demand, according to new research from Bain & Company.
Bain’s sixth annual Global Technology Report, released today, indicates that by 2030, global incremental AI compute requirements could reach 200 gigawatts, with the U.S. responsible for half of this power. Even if companies in the U.S. redirected all of their on-premise IT budgets to cloud computing and reinvested the savings from AI applications in sales, marketing, customer support, and R&D into capital spending for new data centers, the total would still fall short of the revenue needed to support the full investment. This is due to AI’s compute demand growing at more than twice the rate of Moore’s Law, Bain notes.
“If the current scaling laws hold, AI will increasingly strain supply chains globally,” stated David Crawford, chairman of Bain’s Global Technology Practice. “By 2030, technology executives will be faced with the challenge of deploying about $500 billion in capital expenditures and finding about $2 trillion in new revenue to profitably meet demand. Meanwhile, because AI compute demand is outpacing semiconductor efficiency, the trends indicate a need for dramatic increases in power supply on grids that have not added capacity for decades. Add to this the arms race dynamic between nations and leading providers, and the potential for overbuild and underbuild has never been more challenging to navigate. Working through the potential for innovation, infrastructure, supply shortages, and algorithmic gains is critical to navigate the next few years.”
Scaling AI for profitability
As computational demand rises, leading companies have transitioned from piloting AI capabilities to realizing profits from AI as organizations scale the technology across core workflows, yielding earnings before interest, taxes, depreciation, and amortization (EBITDA) gains of 10 percent to 25 percent over the past two years. Yet, the report concludes that most companies today remain mired in AI experimentation mode, content with modest productivity improvements.
Tech-forward enterprises are actively pursuing agentic AI capabilities, resulting in an unprecedented pace of innovation, according to the report. In the next three to five years, 5 percent to 10 percent of technology expenditure could be allocated toward building foundational AI capabilities, including agent platforms, communication protocols, and real-time data access and discoverability for agents. Bain estimates that as much as half of overall technology spending by companies could be directed toward AI agents operating across the enterprise.
Four levels of AI maturity
As AI progresses, the report illustrates that leaders will further distance themselves from laggards in four levels of maturity: (1) large language model (LLM)-powered information retrieval agents, (2) single-task agentic workflows, (3) cross-system agentic workflow orchestration, and (4) multi-agent constellations. Levels 2 and 3 are where capital, innovation, and deployment velocity converge. With AI innovation advancing rapidly, leaders are compounding their advantages, while followers lag behind.
Meanwhile, enterprise IT architectures are struggling to realize the vision of contextually-informed, secure agents collaborating seamlessly across multiple applications and databases to automate the diverse tasks typically performed by humans. Bain identifies a north star architecture as critical, but profit motives and security demands will lead to uneven progress toward the four levels of agentic maturity over the coming years.
Disruption in the SaaS landscape
SaaS providers are encountering disruption from the emergence of generative and agentic AI, yet this does not equate to obsolescence. In many instances, this can be total addressable market (TAM)-additive for SaaS providers. Companies should assess two independent characteristics as they plan to leverage AI in their industry: the potential for AI to (1) automate SaaS user tasks and (2) penetrate SaaS workflows. SaaS incumbents are well-positioned to lead, but this will necessitate high-stakes strategic decisions—such as selective open-sourcing or a shift in the monetization model—and will result in a unique, durable industry influence position. To maintain a competitive edge, providers must own the data, lead on standards, and price for outcomes rather than log-ons in an AI-first landscape.
“In the Middle East, SaaS providers are already beginning to see the impact of AI across critical workflows,” remarked Brahim Laaidi, partner at Bain & Company Middle East. “The opportunity is to use AI not just to optimize tasks, but to reimagine how value is created for customers. Providers that act boldly, by adapting monetization models, investing in data ownership, and embedding AI-first standards, will secure long-term leadership in a rapidly evolving market.”
Sovereign AI as a strategic advantage
Tariffs, export controls, and the global push for sovereign AI are accelerating the fragmentation of technology supply chains, Bain finds. Cutting-edge domains like AI are no longer merely catalysts for economic growth; they have become conduits for countries’ political power and national security. As semiconductor supply chains fragment, the U.S. and China remain at the forefront of the decoupling movement, with China accounting for roughly 20 percent of global chip manufacturing capacity this year, the report indicates.
“Sovereign AI capabilities are increasingly viewed as a strategic advantage on par with economic and military strength,” said Anne Hoecker, head of Bain’s Global Technology practice. “While sovereign AI is a global priority, the objectives of individual countries vary. Therefore, for most countries, achieving full-stack independence is not feasible, at least not today. Given these differences, global AI standards are unlikely to converge.
“To succeed, multinational firms will need to localize not just compliance, but also their technology architecture. Businesses must make decisions with optionality, moving boldly where confidence is high and prioritizing flexibility where uncertainty prevails.”
Unlocking market value potential
Alongside the rapid acceleration of AI initiatives, two distinct technological phenomena are emerging: advancements in quantum computing and the rise of humanoid robots.
Quantum computing has the potential to unlock up to $250 billion in market value across industries such as pharmaceuticals, finance, logistics, and materials science, presenting opportunities that will unfold gradually—though not guaranteed, Bain estimates. While the potential of quantum computing is vast, achieving full market potential will require a fully capable and fault-tolerant computer at scale, which remains years away, the report concludes.
In parallel, interest in humanoid robots has surged from viral videos to billion-dollar valuations. As humanoid robots become more common, commercial success will depend on ecosystem readiness, and companies that pilot these technologies early will be best positioned to lead in a new era of growth. While robots attract headlines and investment, most deployments are still in early stages and heavily reliant on human supervision, according to Bain.
“Technologies like quantum computing and robotics may seem distant, but their potential applications are highly relevant for industries in the Middle East,” stated Dr. Houssem Jemili, senior partner at Bain & Company Middle East. “From energy and logistics to healthcare and smart cities, these innovations could fundamentally reshape efficiency and competitiveness. Regional leaders that explore practical pilots today will be best positioned to capture value as these technologies scale globally.”
Quest for new growth opportunities
The era of easy technology private equity (PE) deals is fading as deal momentum driven by software sales slows. According to Bain’s 2025 analysis of PE technology deals in North America, deal-making has faced challenges. Tariff-related uncertainties and geopolitical tensions have impacted the market. This is true despite an uptick in tech deals during the first half of this year, the report noted. While software spending continues to outpace overall GDP growth, its penetration is reaching its limits across major sectors, such as manufacturing and retail, presenting new challenges for tech investors who must work harder to uncover new sources of top-tier growth. Despite the slowdown in deal processes, investors remain optimistic, as technology has outperformed most other sectors in terms of deal-making, Bain’s analysis reveals.
Global AI infrastructure investment surge
Beyond Bain’s report, recent research estimates that global AI infrastructure investment could reach $6.7 trillion by 2030, driven by surging demand for AI-powered data centers and cloud capacity. AI workloads are projected to account for about 70 percent of data center demand by 2030. This shift will require substantial upgrades in compute hardware and energy supply. These projections align with Bain’s warnings about infrastructure constraints and supply chain challenges. AI chip markets are also forecast to exceed $400 billion by 2030, highlighting intense competition and innovation in specialized processors for scalable AI.
Cutting-edge AI innovations, such as agentic AI, involve autonomous software agents performing complex workflows. These innovations are expected to account for 5 percent to 10 percent of technology spending in the near term. Advances in neuro-symbolic AI combine deep learning with logical reasoning. AI systems with long-term memory are also being developed. These advancements enable more reliable and contextually aware applications across industries.
Sustainability remains a critical factor. Energy-efficient AI models are increasingly prioritized. Green data centers powered by renewable energy are also becoming a focus as the sector faces soaring power demands. Quantum computing is still in its early stages. However, it promises to unlock new AI capabilities and market value across pharma, finance, and logistics in the coming decade. This aligns with Bain’s emphasis on emerging technologies reshaping AI’s future.