By Eugene Lee and Julianna Vitolo
As we move into 2025, we wanted to share some of our predictions about where the market is headed next year – particularly as it pertains to our area of investment focus. Some of these are going to be fairly obvious, but just because there’s consensus around a topic doesn’t mean there’s no upside opportunity, even when it comes to venture-scale returns.
Overall, 2024 was a year of major upheaval and significant structural changes in enterprise software, especially at the foundation layer. We think 2025 could be even more transformative – and we anticipate that the volume and character of activity in B2B software venture deals will reflect that.
Here’s a rundown of our top predictions for 2025, and some of the resulting implications:
Macroeconomics & Policy Changes
Interest rates will continue to decline, albeit gradually, which will continue to prop up the venture fundraising environment and increase capital deployment rates. While pacing was conservative in 2024, we expect it to quicken and drive more competition between VCs for “hot” deals
Agents
Agents will continue their starring role in the AI enthusiasm cycle, with most startups focused on automating tasks and actions for enterprises
OpenAI’s planned launch of Operator in Jan 2025 will further fuel this trend
We will continue to see applications of agents customized and trained to complete a specific function or use case. We have seen this for customer support, for example.
We’ll likely continue to see an explosion of AI agents for very specific use cases, rather than general purpose agents (for now)
The end state will be horizontal (orchestrator) and vertical (including function-specific). Horizontals will come as soon as verticals prove their worth and accuracy
AI: Most Disruptive Impacts
Areas where AI will be most disruptive:
AI will begin to replace the Big 4 professional services firms – potentially by buying individual practices and replacing them with software. Any more rules-based divisions like tax, audit, or recruiting services will likely be the first to get disrupted, but eventually any consulting services could be impacted.
Compliance: Monitoring of adherence to rules, instant updates based on rule changes, auto-checks on outgoing and external comms/reporting becomes much more prevalent, and a default for tech-forward enterprises
AI replaces middle management layer at some large-scale tech companies looking to get back to “founder mode”. As IC’s become supercharged with AI capabilities and productivity soars, we envision there will be less of a need for a layer between them and the upper level management. Companies will be incentivized to implement scrappy scaling tactics and to keep teams lean
Code: Improvements will result in the ‘democratization of the 10x engineer,’ which will mean funding imbalances at early stages have less impact on startups’ ability to compete with one another
We will see the beginnings of a fully AI-native ERP by end of next year
AI: Slower Burn Transformations
AI is machine communication and learning — so it’s specifically designed for machine tasks
During the first Industrial Revolution, we learned to operate machines and they were designed for the optimal interaction of humans and machines
During the current Machine Intelligence Revolution, we’re removing ourselves from the equation — machines will be designed for optimal interaction with other machines (i.e. the multi-agent systems of the future will likely have different UI than the human-machine interfaces of present)
Companies will need to figure out their proprietary data moat if they're building at the application layer. Everyone will integrate with everyone else, and throw some automated agent or workflow on top, but what’s unique to you and what you’re building going forward will be in the new data being captured or created
The definition of core vs. non-core will change: with embedded products, more businesses will use internal capabilities to build AI-specific internal tools either for their own use or into their products. As a result, tools designed for non-AI engineers will see tremendous growth
We will see increased trust & safety demand as AI proliferates since it’s the main blocker to enterprise adoption
Pricing models in the near-term will be volume and outcome-based. We have already begun to see this practice take off within the customer support agent use case. Our expectation is that it will proliferate across other use cases in the next year as companies toy with pricing to encourage greater AI adoption
Build vs. buy
Most large enterprises will be grappling with whether they should build new AI technology in-house or buy from third parties. We believe the main thinking around this will be to only build what you can’t currently buy, as long as you have the engineering talent to support it.
If you have the technical prowess to build it, it will become a core competency for the business, and it won’t take too much time and energy away from the business’ own objectives then building it might make sense.
An important caveat is that if it’s possible to buy it, then building it is likely not technically challenging enough and won’t create a deep enough competitive advantage
Combining a build/buy approach will create the most unique advantage
Financial Markets (M&A and IPO)
M&A Markets
Corporate development teams finally figure what they want to buy, primarily driven by a top down AI strategy or acqui-hires to pick up AI and ML talent to build internally
IPO Markets
Based on the positive reception of the ServiceTitan public offering, IPO hopefuls like Stripe, Chime, Klarna, Coreweave, StubHub, and Cerebras may get the green light
We expect to see the number of software IPOs hit double digits in 2025
The one catch will be in post-IPO performance – public markets will be less tolerant of cash burning public stocks and will expect a tighter timeline to hit profitability – likely two-to-four quarters rather than the previous four-to-six
So far, ServiceTitan has performed quite well, jumping over 50% in the first week of trading after its official public debut at $71 per share. This bodes well for the rest of the market and gives us greater confidence we’ll see some movement in the IPO backlog in 2025
As we approach 2025, these emerging patterns represent just a snapshot of what we're monitoring. What developments do you anticipate shaping the coming year? Reach out and let us know!
