From AI agents to enterprise budgets, 20 VCs share their predictions on enterprise tech in 2025

While AI is lauded by some as the biggest technological breakthrough since the industrial revolution, enterprises — arguably the tech’s biggest potential customer base — have been slow to adopt AI.

While some investors predicted that 2024 would be the year we’d start to see more AI adoption by enterprises, that didn’t play out as budgets remained constrained and AI tech often remained in the “experimental” category.

Will that all start to change in 2025? Depends on who you ask.

TechCrunch talked to 20 venture capitalists who back startups looking to sell to enterprises about their predictions for 2025. They told us what they anticipate regarding enterprise budgets, trends worth following, and what it will take to raise a Series A in 2025, among other things. Here’s what they said.

SC Moatti, managing partner, Mighty Capital: I’m really looking into this theme — AI adoption hinges on better data. As enterprises transition from AI experiments to large-scale deployment, the demand for high-quality data intensifies.

Aaron Jacobson, partner, NEA: Code agents for app development modernization are underhyped. Expect to see AI being used to re-platform mainframe apps to the cloud and upgrade older codebases.

Molly Alter, partner, Northzone: A key focus of mine is on spaces that were historically untouchable by venture funds because their business models demanded high COGS or OpEx. We’re seeing AI automate so much behind-the-scenes work that sectors like accounting services, or revenue cycle management, or white-glove legal services can now command software-like margins.

Marell Evans, founder and general partner, Exceptional Capital: Understanding trends in enterprise sales cycles — what is the duration certain organizations are trialing tools for before making decisions about internal adoption? In addition, understanding the different pricing models of AI [in relation to] traditional SaaS, consumption-based and/or outcome-based.

Mike Hayes, managing director, Insight Partners: An unappreciated metric and something that I think will gain traction in 2025 is TTFV, or time-to-first-value. I see this as a proxy for ease-of-implementation, so faster TTFV solutions should have a bigger advantage going into [the] new year.

What areas are you looking to invest in?

Liran Grinberg, co-founder and managing partner, Team8: Enterprise resilience, whether in front of operational faults or malicious insider or outsider threats. The CrowdStrike software update incident demonstrated how fragile our digital world is, not only due to cyberattackers but also just mistakes.

Jonathan Lehr, co-founder and general partner, Work-Bench: Data sovereignty as a service. Organizations are increasingly investing in data sovereignty solutions driven by regulatory requirements and geopolitical concerns. We are exploring startup opportunities that enable companies to maintain complete control over their data’s location, storage, processing, and governance while ensuring compliance with local regulatory frameworks.

Mark Rostick, vice president and senior managing director, Intel Capital: One area we’re looking at is companies that focus on task-specific models. While the foundational models are well established, I find models that excel at specific functions particularly intriguing, especially when combined with agents built on top of them. In addition, we are closely monitoring the development of alternatives to transformers and any possible solutions to reduce the need for the huge amount of computing capacity now required to train LLMs and use them in production.

Mike Hayes, managing director, Insight Partners: Enterprises have historically thought of technology as either driving revenue or reducing cost, but that is quickly changing in favor of technology that drives enterprise value while simultaneously reducing business friction. I look for solutions that solve unique, orthogonal challenges for enterprises — areas where traditional solutions have fallen short; this includes vertical and persona-specific workflows reimagined with GenAI or agentic automation and security innovations that not only identify and alert, but also remediate.

Jason Mendel, venture investor, Battery Ventures: A few interesting areas where I think AI can add significant value, and which I’m excited about, include observability / incident response, IT service management, demand generation and sales engagement, offensive security, software development, and the SOC workflow.

Ed Sim, founder and general partner, Boldstart Ventures: We think in second order effects. So if we assume that in the future, meaning the next two to three years, we could live in a world where each of us has dozens or hundreds of agents doing work for us, we need to think about all of the infrastructure that needs to be built to support these new digital employees. Who’s going to provide the security infra to provide access control? Who’s going to manage these? Is there a platform to manage disparate agents and secure them? What about a runtime system for Claude’s MCP, which feels like a dockerized, secure sandbox for agents to do work.

What technologies, sectors, companies, etc., are you finding interesting that aren’t AI?

Liran Grinberg, co-founder and managing partner, Team8: Quantum computing is still promising. Cybersecurity isn’t going anywhere as well, with attackers leveraging AI and an increased complexity in protecting our digital infrastructure.

Nina Achadjian, partner, Index Ventures: We’ve seen a resurgence in fintech, SaaS, and e-commerce, which were hot sectors that saw a slowdown in the last couple of years. Beyond that, we expect cyber and gaming to continue to be interesting this year, with cyber accelerating further as the IPO market opens up and regulations and disclosure rules around security increase.

Aaron Jacobson, partner, NEA: There is a ton of hype around securing AI, but the bigger opportunity is helping enterprises apply ”Cybersecurity 101” at scale in a way that doesn’t impede user productivity. Key areas of particular interest are enforcing least privilege access, maintaining a secure data posture, and preventing ransomware. I’m also excited to invest in technology that facilitates multi-cloud deployments for enterprises.

Molly Alter, partner, Northzone: I’m really excited about companies addressing the public sector. The fiscal environment for government contracting is flush; total federal agency contracts reached $774 billion in 2023. Technology adoption and modernization are key to driving the efficiencies that the new administration is committing to, and there is a growing ecosystem of companies that are tackling this head-on.

Andrew Ferguson, vice president, Databricks Ventures: We’re spending a significant amount of time with our system integrator partner ecosystem. These companies are doing the hard work of helping enterprises take their data and AI strategies and turn them into real-world implementations.

Janelle Teng, vice president, Bessemer Venture Partners: We are moving beyond the modern data stack. The data infrastructure landscape is undergoing a massive transformation, fueled by various factors, including the rise of lakehouse architecture and convergence toward specific open table format standards.

Raviraj Jain, partner, Lightspeed Venture Partners: Energy is a huge sector to invest in given increasing demand for energy for data centers and the challenges with grid failures across the country. We’ll see continued interest in nuclear — both fusion and fission.

When it comes to AI, how are you determining that a company has a moat?

Cathy Gao, partner, Sapphire Ventures: I think about it in a “5D framework”: design, data, domain expertise, distribution, and dynamism. Since early this year, we at Sapphire have used this framework to evaluate companies building applications with AI.

SC Moatti, managing partner, Mighty Capital: An AI moat is built on proprietary data, cutting-edge algorithms, and scalable infrastructure, enabling unique and superior solutions.

Scott Beechuk, partner, Norwest Venture Partners: The deepest moats will be created by large proprietary datasets. The companies with the greatest long-term potential are those building their own unique datasets to excel in their particular, verticalized channels — often by either training or fine-tuning their own models.

Jonathan Lehr, co-founder and general partner, Work-Bench: As a pureplay seed fund, we’re focusing most of our energy in vertical AI opportunities tackling business-specific workflows that require deep domain expertise and where AI is mainly an enabler of acquiring previously inaccessible (or highly expensive to acquire) data and cleaning it in a way that would’ve taken hundreds or thousands of man-hours.

Raviraj Jain, partner, Lightspeed Venture Partners: Question to ask is, As models become better, does this company get threatened or strengthened?

What does it take to raise a Series A as an enterprise startup in 2025?

Liran Grinberg, co-founder and managing partner, Team8: With a strong founder-market fit, and an ambitious vision to build a big company, one can raise a solid $15 [million to] $25 million Series A round with only a few $100Ks in ARR.

Molly Alter, partner, Northzone: Successful Series A enterprise startups will show strong topline traction (>100% YoY) with low burn multiples; gone are the days of 2021 when it was all about growth at all costs. More importantly, these businesses will show a clear long-term differentiation strategy that will set them apart from the host of other offerings attempting to raise money and sell into the same enterprise customer base.

Kirby Winfield, founding general partner, Ascend: Go from zero to $1 million in two quarters with an A-plus team in a massive market with a differentiated solution having created overwhelming demand.

Andrew Ferguson, vice president, Databricks Ventures: If you’re building an AI-first product, an all-star technical team and early product market traction ($2 [million to] $5 million ARR) may be the Series A expectation. The time from product launch to $5 million ARR is materially faster in the AI era than it was in the traditional SaaS era. I expect that the Series B bar will be much higher — and it remains to be seen if this early ARR is high-quality and durable.

Jonathan Lehr, co-founder and general partner, Work-Bench: We’re hearing from downstream peers that the bar is around $1.5 million with the ability to grow 3x from there sequentially to raise a stellar Series A.

Jason Mendel, venture investor, Battery Ventures: Repeatability. Startups that are solving a real pain point in a large market where there is clear urgency from a buyer/user perspective should be well-positioned to raise a Series A in 2025.

Do you predict enterprises will increase their tech budgets for 2025? Will they decrease them?

Aaron Jacobson, partner, NEA: Within AI, we’ll see budget allocated away from “chatbots” to agents. Enterprises will move beyond the low-hanging fruit of “GPT wrappers” to deploy digital workers that can reason and take action to make a real business impact.

Scott Beechuk, partner, Norwest Venture Partners: Tech budgets across many industries will increase in 2025, driven by leaders’ desire to achieve two goals — which will sometimes be at odds with each other. The first goal is consolidation. The second is increasing top-line growth and improving operational efficiency, both of which are achievable with AI-based software applications. Buyers will purchase startup solutions in this category despite their desire to consolidate.

Kathleen Estreich, partner, Pear VC: In 2024, we expected to see more enterprise adoption of AI. But that hasn’t panned out, primarily because we haven’t yet figured out use cases that are tightly scoped enough and the tools to reduce hallucinations and validate outputs have not gotten robust enough. In 2025 I expect to see more enterprise adoption as the model providers extend their stack upward. Every enterprise will need an AI tech strategy. If you don’t adopt, you won’t keep up. This will also create a lot of false signals on the revenue side for AI startups as experimental budgets will be high, but true product-market fit will be harder to see at first glance.

Kirby Winfield, founding general partner, Ascend: Enterprises will increase AI budgets in 2025. The question isn’t whether they’ll invest but how they’ll tackle pricing, testing, and data security. Companies like Salesforce and Smartsheet have already committed to AI adoption and will push harder to leverage their data assets to stay competitive.

Susan Liu, partner, Uncork Capital: Probably the same for the first half, and then as the economy improves and revenue/profits improve, we’ll see an increase in tech budgets in the second half.

Mike Hayes, managing director, Insight Partners: Based on what I’m hearing from our enterprise partners, they’re likely to marginally increase their tech budgets in 2025, with a focus on areas that deliver measurable ROI and clear KPIs. I expect pressure from boards and CXOs to put AI use cases into production to increase and receive discretionary budget. I also expect continued enterprise investment in cybersecurity and cloud optimization. Said differently, the right emerging technologies should not have trouble landing due to tech budgets.

Jason Mendel, venture investor, Battery Ventures: I’m optimistic about 2025 and expect to see companies increase their IT budgets with a strong focus on emerging technologies. Heading into the 2025 budgeting season, we at Battery Ventures polled 100 CXOs, collectively representing over $35 billion in annual technology spend, and 74% of them expected to increase their technology spend in 2025.

Will there be more AI adoption?

Paul Drews, managing partner, Salesforce Ventures: Yes, essentially all enterprise workflows can be optimized with AI — especially agentic AI. We’re seeing real demand for AI and ML tools that can make underlying models 50% more efficient while delivering improved results. AI is experiencing froth, but from a larger market perspective (not just Silicon Valley), AI is still new and everyone is trying to figure out how to use it, price it, and purchase it.

Mark Rostick, vice president and senior managing director, Intel Capital: For the moment, it is clearly easier to adopt AI through application vendors than trying to build your own platform given that the market for enterprise platform tools is still very, very fragmented. I do think there is pent-up demand for some sort of platform solution, so I believe we’ll see many founders trying to address that problem this coming year.

Raviraj Jain, partner, Lightspeed Venture Partners: It’s a consensus view but AI adoption will continue to accelerate in 2025 as (1) model capabilities improve, (2) enabling infrastructure is built out, and (3) stronger AI-first products come to market.

What kinds of companies in your portfolio are seeing the strongest growth? Do you predict that will change in 2025?

Marell Evans, founder and general partner, Exceptional Capital: Urgent pain points for AI-ready customers are producing shorter enterprise sales and procurement cycles and therefore faster traction and scale. As we see AI adoption more broadly, we may see enterprises will have greater appetite to try not just solving for the urgent problems but also planning ahead to maintain competitive edge with “nice to have” or more future-forward and strategic solutions.

Kathleen Estreich, partner, Pear VC: We are seeing great traction in vertical agents with a clear understanding of the unique needs of their customers. I think vertical SaaS is a huge opportunity in 2025 to own the end-to-end workflows with custom-built agents for the tasks to be done.

Janelle Teng, vice president, Bessemer Venture Partners: Many of Bessemer’s AI defense tech companies experienced tremendous growth this year. One of our observations earlier in the year is that the defense community is not sitting idly by as the AI revolution sweeps the consumer and commercial industries by storm. The [Department of Defense] mapped and released its formal AI adoption strategy last year, and we predicted that advancements and applications of ML will be embraced as essential for the national agenda and the defense community’s day-to-day work. This prediction proved prescient as the year continued.

Mark Rostick, vice president and senior managing director, Intel Capital: Another strong segment of the portfolio focuses on the infrastructure layer of software and services companies. Anyscale is a fantastic example. With their software, developers can build, run, and scale AI applications instantly. There’s also RunPod, a virtual cloud service provider (CSP) for inference. It can bridge the gap between hardware and software stacks, which allows for seamless operation across various server providers, addressing a current challenge in the AI space.

Ed Sim, founder and general partner, Boldstart Ventures: No. This is one of the greatest platform shifts I’ve seen in 29 years of being a venture capitalist and IMO this will only accelerate.

What are your predictions for the exit environment next year?

Cathy Gao, partner, Sapphire Ventures: I predict M&A activity will increase as large companies seek to acquire AI expertise. Strategic acquirers will focus on startups with domain-specific AI capabilities or high data moats. The IPO market will remain cautious, but high-growth companies with profitability metrics might test the waters.

Nina Achadjian, partner, Index Ventures: I anticipate more liquidity in 2025, both for M&As and the public markets.

Aaron Jacobson, partner, NEA: With the change of administration, I expect the return of mega M&A deals. We are going to see a multi-billion and even decacorn M&A outcome for a leading AI company.

Marell Evans, founder and general partner, Exceptional Capital: We expect exits to pick up slightly next year, possibly more acquisitions and IPOs. Although, given the latest fed meeting, exit volume might be slower than we expected.

Kirby Winfield, founding general partner, Ascend: I predict new FTC leadership under the incoming administration will make hyperscalers more acquisition-friendly for tech and talent. But the IPO market will likely remain sluggish, given the frothy valuations some companies can command from the private market.

Andrew Ferguson, vice president, Databricks Ventures: 2025 may finally be the year that we see an uptick in tech M&A activity, as more favorable macro and (potentially) less onerous regulatory oversight make larger companies less skittish about M&A. Most strategic M&A will be focused around amazing technical founders and technology, rather than on scaled business, especially ones that matured during the ZIRP era where the growth/profitability metrics may still not pencil out for strategic acquirers. It’s possible that private equity or growth equity investors make a play to consolidate that class of assets into broader platforms.

Paul Drews, managing partner, Salesforce Ventures: The likely emphasis on government efficiency and lower regulation will spur growth, investments, and exits. The public markets are soaring, but there continues to be hesitation around the IPO process from a private company perspective. We’ve seen glimmers of hope in the IPO markets, which pre-IPO businesses should take as a good sign, but there is still some disconnect between the last private valuation and where the public market will price businesses.

Mike Hayes, managing director, Insight Partners: I think enterprises will look to strengthen their inorganic growth through acquisition more in 2025 than in 2024. As far as the IPO market, I do think that enterprises focusing on mission-critical solutions with predictable revenue will have opportunities in 2025. I’m optimistic and energized for 2025.