VCs Bet on Enterprise AI Boom in 2024
▼ Summary
– Despite significant investment, most enterprises (95% in one survey) are not yet seeing a meaningful return on their AI investments.
– A survey of enterprise VCs indicates that 2026 is widely expected to be the year businesses start to meaningfully adopt AI and see value from it.
– VCs predict key 2026 trends will include a shift from general LLMs to custom models and vertical solutions, a focus on AI in the physical world, and the rise of voice AI.
– For AI startups to succeed, VCs emphasize the need for a defensible “moat” built on proprietary data, deep workflow integration, and clear enterprise traction, not just model performance.
– Enterprise AI budgets in 2026 are expected to increase but become more concentrated, funding only proven solutions while cutting experimental tools and vendor sprawl.
Three years after the initial wave of excitement sparked by ChatGPT, the enterprise software landscape is still grappling with the practical realities of artificial intelligence. While investment has poured into a multitude of startups, a significant gap remains between promise and measurable payoff. A recent survey highlighted that a vast majority of enterprises have yet to see a meaningful return on their AI investments. This has led industry observers to look further ahead, with many venture capitalists pinpointing 2026 as the pivotal year when businesses will begin to meaningfully adopt AI, see clear value, and significantly increase their budgets for the technology.
The question remains whether this prediction will hold true after years of similar forecasts. To understand the path forward, insights from two dozen enterprise-focused VCs reveal the trends, investment theses, and benchmarks that will define the coming era.
Several key trends are expected to gain serious traction by 2026. Kirby Winfield of Ascend notes a shift in enterprise thinking, moving away from viewing large language models as universal solutions. The focus will turn to custom models, fine-tuning, and robust data governance. Molly Alter of Northzone anticipates a strategic pivot for some AI companies, evolving from product-centric businesses into AI consulting and implementation partners as they accumulate deep customer workflow knowledge.
Other areas drawing significant attention include voice AI, which Marcie Vu of Greycroft sees as a more natural interface poised for reinvention, and the physical world, where Alexa von Tobel of Inspired Capital predicts AI will transform infrastructure and manufacturing through predictive systems. Lonne Jaffe of Insight Partners is watching how leading AI labs may move beyond model training to deploy turnkey applications directly in sectors like finance and healthcare.
Regarding investment focus, themes are coalescing around specific challenges and opportunities. Emily Zhao of Salesforce Ventures is targeting AI’s integration into the physical world and next-generation model research. The immense energy demands of AI infrastructure are a major concern, leading Aaron Jacobson of NEA to seek out breakthroughs in performance per watt, from efficient chips to advanced cooling solutions. Jonathan Lehr of Work-Bench emphasizes vertical enterprise software where proprietary data and workflows create strong defensibility.
This concept of a defensible “moat” is critical for AI startups. Investors widely agree that advantages based solely on model performance are fleeting. Rob Biederman of Asymmetric Capital Partners argues that true moats are built on economic integration and switching costs, not the model itself. Molly Alter finds it easier to build durable advantages in vertical categories where consistent data creates a “data moat.” Harsha Kapre of Snowflake Ventures looks for startups that excel at transforming an enterprise’s existing data into actionable insights without creating new data silos.
On the central question of value realization, opinions vary but trend toward cautious optimism. Scott Beechuk of Norwest Venture Partners believes 2026 will be the year the application layer begins to turn infrastructure investment into tangible returns as specialized models mature. Jennifer Li of Andreessen Horowitz contends that value is already being gained, citing the indispensable nature of AI coding tools for engineers, and expects this to multiply across organizations.
Budget forecasts suggest a year of concentration and rationalization. Rajeev Dham of Sapphire expects budgets to increase through a shift in labor spend or strong ROI, while Rob Biederman predicts a bifurcation in spending: sharp increases for products that deliver clear results and declines for others. Andrew Ferguson of Databricks Ventures foresees CIOs pushing back on vendor sprawl, cutting experimental budgets, and doubling down on proven technologies.
For startups seeking Series A funding in 2026, the bar is clearly defined. Jake Flomenberg of Wing Venture Capital states that founders must combine a compelling “why now” narrative with concrete proof of enterprise adoption, where customers view the product as mission-critical, not just a nice-to-have. Jonathan Lehr stresses that startups must demonstrably save time, reduce costs, or increase output in a way that withstands rigorous enterprise procurement reviews.
The role of AI agents by the end of 2026 is another area of speculation. Nnamdi Okike of 645 Ventures believes agents will still be in early adoption, facing technical and compliance hurdles. In contrast, Rajeev Dham envisions a move away from siloed agents toward a single, universal agent with shared context across organizational functions. Aaron Jacobson offers a more personal prediction: the majority of knowledge workers will have at least one agentic co-worker they know by name.
Currently, the strongest growth is seen in companies addressing new challenges born from AI adoption itself. Jake Flomenberg points to cybersecurity tools for LLM data security and new marketing paradigms like Answer Engine Optimization. Andrew Ferguson highlights success for companies that start with a narrow, focused use case and expand from that solid foundation.
Strong customer retention, a key health metric, follows distinct patterns. It is highest for companies that become foundational infrastructure, as noted by Jonathan Lehr, or that solve problems that intensify with greater AI deployment. Tom Henriksson of OpenOcean observes exceptional retention in serious enterprise software providers that deeply transform customer operations and build up irreplaceable proprietary knowledge.
The collective view from the venture capital community suggests that 2026 will be a year of maturation, scrutiny, and selective scaling for enterprise AI. The era of broad experimentation is giving way to a demand for proven value, deep integration, and sustainable competitive advantages.
(Source: TechCrunch)
