In their latest "2026 Big Ideas" report, three partners who have long invested in US industry, financial services, and enterprise software share their key observations for 2026. The overall discussion focuses on three main areas: the accelerated formation of stacked architectures in the electrical industry; the transformation of core architectures in the financial and insurance sectors using AI systems; and the rise of dynamic AI agent layers in enterprise software, which are beginning to shake the dominance of traditional system cores.
The primary focus: the stacking of electrical industries, driving a new industrial revolution.
The electrical industry is stacked and formed, and the industrial revolution is entering the interior of machines.
Ryan McEntush, a partner at Dynamic Investments in the United States, points out that the key change in 2026 is that the “electrical industry stack” will begin to take shape, driving the next industrial revolution.
He stated that industrial progress is no longer confined to factories, but extends to the equipment and machines themselves. Electric vehicles, drones, data centers, and modern manufacturing all rely on a common set of electrical and electronic technologies, including semiconductor components such as batteries and power electronics, as well as core components such as computing power and motors.
He pointed out that the United States is not lagging behind in engineering and key technologies; the real challenge lies not in the technology itself, but in how to industrialize and scale up the technology and make it cost-competitive. In contrast, China's advantage comes from its complete supply chain and supporting systems, which can quickly support enterprise expansion.
Ecosystems and supply chains determine long-term competitiveness.
McEntush, citing SpaceX as an example, points out that high vertical integration is often not a strategic choice, but rather a result of insufficient supply ecosystem. In China, a complete supply chain has been established, but in the United States, it still needs time to be completed.
Therefore, to build an electrical industry stack, it is necessary to advance technology, supply chain, and systems simultaneously, rather than simply transferring bottlenecks. Regarding talent, he believes that software culture and traditional industrial experience must be combined, accelerating progress through close collaboration between engineering and manufacturing, and instilling a strong sense of mission to attract top talent.
As software and AI continue to penetrate deeper into industrial and military applications and gain control of key supply chains, they will influence the global economic and military power distribution in the coming decades.
The second major trend: The financial and insurance industry is moving away from its old core focus, with AI-native platforms becoming the new mainstream.
When the system transformation reaches its critical point, the risks of the old architecture become even higher.
Angela Strange, a partner at the AI Applications Fund, points out that 2026 will be a key turning point for the financial and insurance industry. For a long time, the industry has generally believed that replacing core systems is too risky, but this perception is changing.
She observed that an increasing number of large institutions are choosing to let their contracts expire and switch to AI-native platforms, because the risks of not upgrading outweigh the transformation itself. She pointed out that the core of next-generation infrastructure is not "adding AI to old systems," but rather integrating information scattered across existing core systems, external systems, and unstructured data to rebuild the data core, enabling financial institutions to expand and truly realize the benefits of AI.
Process restructuring and scaling up have widened the gap between pioneers.
Strange stated that the new platform brings three structural changes: process parallelization, integration of risk and compliance data, and software addressing manpower gaps, which has amplified the market size.
She also explained that the timing of this transformation is related to the fact that legacy mainframe systems are nearing their limits, the revenue opportunities brought by AI are becoming tangible, and the emergence of AI-native startups that truly understand the industry. Strange stated that banks and insurance companies that have already completed their system transformation have seen a significant increase in the profitability of some of their businesses, while the gap between them and their slower-transforming peers is not just a few months, but is widening on a yearly basis.
The third major direction: the formation of dynamic AI agent layers, and the entry of enterprise software into a period of structural transformation.
The rise of dynamic proxy layers has weakened the core status of the system.
Sarah Wang, a partner at a16z's growth investment team, focuses on the structural changes in enterprise software. She points out that as AI agents can move directly from "user intent" to "actual execution," the core of traditional systems, which are primarily based on passive recording, begins to lose its original rationale.
She frankly admitted that her long-term investment in core systems such as ERP was precisely because of their data stickiness, but this is the first time that a technological condition has emerged that could truly shake their position. Taking IT service management as an example, the field, previously dominated by enterprise software company ServiceNow, is being rapidly rewritten by a new generation of AI agents. A senior IT executive even stated that this is the first time in his 20-year career that he believes IT support will undergo fundamental changes within the next 5 years.
Being close to users and iterating rapidly become the key to success.
Wang pointed out that the disruptive nature of AI agents lies in their ability to instantly understand needs, classify requests, correspond to processes, and complete execution, significantly shortening the originally lengthy application and processing procedures.
She believes that while the basic model layer will still exist in the future, the real long-term value will be accumulated in the agent layer, which is closest to the user. This layer will continuously collect user preferences and behavioral data, forming new competitive advantages. At the same time, the speed of product evolution becomes crucial; if the agent is not accurate and reliable enough, it will be difficult to gain user trust. She also observed that even agents built on large platforms are being replaced by startups like AI SRE, indicating that the enterprise software market is undergoing rapid reshuffling.
(Note: AI SRE companies refer to a type of startup that focuses on applying AI to website/system reliability engineering, enabling AI to automatically monitor, detect, diagnose, and even repair problems in IT infrastructure or software systems, rather than just passively issuing alerts.)
This article, "Looking Ahead to 2026: Electrical Industry, Financial Infrastructure, and AI Agent Layers: The Pillars of the AI Era," first appeared on ABMedia, a ABMedia .
