Welcome to the final installment of Term Sheet’s “Crystal Ball” series, where we present predictions and insights, collected from our readers, about what’s in store for the private markets and tech. Today, we focus on artificial intelligence.
2024 sometimes felt like one long conversation about AI—and if the surfeit of comments from Term Sheet readers are any indication, it looks like that conversation (and deluge of investment) will continue well into 2025. Which industries and incumbents will be the most impacted by AI? Will LLMs hit or break through the proverbial wall? And which direction will valuations go?
I asked ChatGPT to write a joke about what will happen with AI in 2025. Here’s what it served up:
Why did the AI in 2025 go broke?
Because it kept investing in “cloud” storage and forgot to check the forecast!
Color me unimpressed. For now, humans clearly still have the edge over AI when it comes to humor. As for predictions, here’s what Term Sheet’s human readers see on the horizon.
Note: Many answers have been edited for clarity and/or brevity.
What’s next: Everything, it would seem
The impact of generative AI on financial services is going to be dramatic. Financial services is full of areas where it’s human beings reading and writing and processing documents. Generative AI is very good at those tasks and we see opportunities for generative AI to be disruptive to the way things are done. —Matt Harris, partner, Bain Capital Ventures
In 2025, the traditional search engine will begin its slow decline as the primary gateway to the internet. AI-powered, interactive interfaces will emerge as the dominant force in how consumers discover and engage with content. —Dave Fink, CEO and cofounder, Postie
Voice AI will pick up the phone. Voice tech will have a transformative effect on call operations in legacy industries, particularly in insurance and healthcare across revenue collection, patient scheduling, plan enrollment and explanations, and clinical trial data collection. —Jordan Segall, partner, Redpoint Ventures
Generative AI is unlocking voice as the most important modality in personal computing over the next several years. State-of-the-art speech models will fundamentally change how we access information, make decisions, and take actions. Over the next 2-3 years with voice AI agents and speech models improving at such a rapid pace, most customer support, booking, healthcare, creative entertainment, social media, and education/learning will be based around the primary modality of voice AI. —Steve Jang, founder and managing partner, Kindred Ventures
In 2025, we’re likely to see more application companies emerge—these are companies that are leveraging AI capabilities to create new ways of doing things in age-old industries like financial services or insurance. You can compare this to the mobile wave—companies like Uber and Airbnb popped up, not as mobile companies, but as companies that wouldn’t exist without mobile. Just like they became application companies for mobile, we’ll see companies do the same with AI and these will be the winners of this next wave. —Wesley Chan, cofounder and managing partner, FPV Ventures
Competition for data will increase; while use cases become more complex: With generative AI and agentic applications on the rise, the competition for data will heat up, driven by synthetic data generation, crowdsourced content, and content studio partnerships. —Lauri Moore, partner, Bessemer Venture Partners
Consumer adoption of AI will soar beyond chat GPT. Everyday people will adopt personalized applications that solve real problems. Happy and frequent use will be driven by how helpful and valuable they are, not because the backend is powered by AI (although it will be). —Sara Deshpande, general partner, Maven Ventures
If 2024 was the year of LLMs, 2025 will be the year of Vertical AI. Industries like healthcare, financial services, logistics, construction, and home services are going to be transformed faster with AI than what’s being done to generalized business work by ChatGPT, MSFT Copilot, etc. —Karthik Ramakrishnan, partner, IVP
In 2025, we will see the impact of AI on proptech, specifically the mortgage industry. We’ll see AI-powered agent assistants capable of reviewing and evaluating mortgage applications, interacting with borrowers on behalf of loan officers. —Diane Yu, cofounder, TidalWave
2025 will be the year of true AI-driven personalization. —Meera Clark, principal, Redpoint Ventures
Amazon will launch AI-powered search. Generative AI will continue to revolutionize online search through ChatGPT and Microsoft Bing. This will be especially true for one of the biggest Internet search engines—Amazon.com, which is the fifth highest-traffic website in the U.S. and arguably the default search for eCommerce in the U.S. Next year, we expect Amazon to integrate generative AI into its search bar soon, allowing shoppers to visit Amazon not only for shopping but also for inspiration and answers for the day-to-day. —Bob Ma, partner, WIND Ventures
We’ll finally cross the uncanny valley in video AI. A major Hollywood studio will embed AI video in a feature film. —Amy Wu, partner, Menlo Ventures
In 2025 and beyond, voice AI will become a true and new kind of personal assistant never seen before, meaning it will rapidly evolve from being task-driven to becoming a contextually aware companion. This transition is only possible as voice AI assistants will now [be able] to anticipate needs based on tone, sentiment, and daily routines, adjusting communications styles dynamically (e.g., offering sympathy when you’re upset or being concise when you’re in a hurry), and also collaborate with other AIs and the internet at-large for holistic personal assistance. These new voice driven assistants (and wow do we need a better word) will also be able to detect stress, depression, or anxiety from vocal tones and suggest immediate coping mechanisms, proactively recommend self-care activities and, perhaps most importantly, will develop a deep “emotional profile” to tailor themselves to your true personality. —Andy Weissman, managing partner, Union Square Ventures
AI in enterprises: Reckoning
2025 will mark the year when AI must prove its ROI. This transition will end the era of “AI theater.” —Megh Gautam, chief product officer, Crunchbase
2025 will see AI adoption in financial services surge, with intelligent automation transforming workflows. —Rob Stone, Senior V.P., General Manager of Intelligent Automation & Analytics, SS&C Technologies
AI data centers and climate clean energy will become a peanut butter & chocolate collision: AI hunger for firm energy will push green generation, storage, and transmission technologies down the learning curve towards cost parity. A good thing. The peace dividend of the AI wars. —Reed Sturtevant, general partner, Engine Ventures
As AI handles more technical tasks, empathy and human connection will emerge as critical differentiators for businesses. In 2025, organizations will prioritize understanding customer needs and fostering collaborative team environments. —Nick Durkin, field CTO, Harness
We will begin to see unprecedented levels of automation at the largest Fortune 500 companies. —Rudina Seseri, founder and managing partner, Glasswing Ventures
(Most) customer support reps log off. The majority of customer support inquiries will be handled by AI, leading to a decline in headcount. —Patrick Chase, partner, Redpoint Ventures
In the last year and a half, many have spent a large portion of their budgets on GenAI, and they may have put other important areas of the IT footprint and data on the back burner and under-invested. So next year, we will see many organizations calibrating the budget better to do more. —Molham Aref, CEO, RelationalAI
Businesses will grow increasingly skeptical of AI offerings. This skepticism will be fueled by the proliferation of rushed AI products that create more challenges than they solve, leading to stalled adoption and investment among wary buyers. —Jim Palmer, Chief AI Officer, Dialpad
Amidst the excitement AI-driven transformation has brought across industries, I predict a reset in 2025 is approaching. The past two years were full of fast-paced AI experiments within companies, but 2025 will be about putting AI budgets to the test and proving its return on investment. With one in three enterprises planning to invest at least $10 million in AI next year, AI budgets are poised to surge, and companies will need support on how to best spend and utilize it effectively. —Richard Robinson, CEO and cofounder, Robin AI
We anticipate the appetite for acquisitions of AI companies by strategics and sponsors will grow in 2025 as the technology becomes more mainstream and every large software solutions provider needs to have a solid offering to stay competitive. This should lead to acquisition prices remaining healthy even for targets that may have not yet achieved traditional SaaS business metrics. —Jon Shalowitz, managing director, Union Square Advisors
In 2025, robust funding for defense-tech is likely to continue, driven by geopolitical tensions, rising defense budgets, and the Pentagon’s push for technological innovation. Investments in key areas like AI/ML, unmanned systems, and space will persist. —Veronica Daigle, partner and president, national security, Red Cell Partners
Valuations: Hitting the wall, maybe
Ethical AI startups will seek big valuations. As AI becomes increasingly integrated into everyday business and personal life, a new industry is going to emerge focused on the ethical and secure usage of AI. SaaS startups focusing on AI integrity, security, governance, privacy, and transparency will generate significant interest in 2025, and attract capital as governments and policymakers introduce stricter AI-related regulations. We are going to see continued VC investments into this field over the year ahead, with newcomers that solve these issues garnering big valuations. —Ivan Nikkhoo, founder, Navigate Ventures
Valuations will remain a tale of two cities: New companies deploying AI in an interesting way will continue to raise at sky-high valuations, yet those with several years of operating history and modest revenues will experience challenges fundraising and maybe even see down rounds for the first time. —Tom Biegala, cofounder, Bison Ventures
The unreasonable AI valuation bonanza will be over in 2025, valuations will be based on financial metrics, customer traction and revenues again. —Michael Hammer, CFO, BMW i Ventures
A leading foundational model AI startup that’s raised $100M+ goes under. Once considered hot, they lose investor confidence as training costs skyrocket, market share proves elusive, and any model not at the cutting edge depreciates to near zero value within a year. —Richard Dulude, cofounder and partner, Underscore VC
Agents: On the move
Time magazine will put an AI agent on its cover, just like it did with the PC in 1983. 2025 will be the break-out year for AI Agents. —Joff Redfern, partner, Menlo Ventures
While the valuations are still soaring, most AI products are big promises with low quality. Real reasoning models, like o1, and more agentic workflows will be the transformation next year—and create the productivity gains everyone is eagerly awaiting. I predict that will change in 2025 as agentic workflows transform the game. —Jeremy Epling, chief product officer, Vanta
AI agents will operate within your organization just like humans. They’ll even begin to interact with other AI agents to accomplish their job. This means AI agents are going to look, feel, and act just like humans do in an organization. They’ll be added to HR systems, have their own permissions and access privileges, and will also need to be on-boarded and off-boarded from systems just like regular human users. —Alex Bovee, cofounder and CEO, ConductorOne
AI agents will become autonomous. Thanks to recent 10x speed boosts in LLMs and specialized hardware, we expect AI to plan, reason, and remain aware of real-time information. Agents will use many kinds of input to execute complex, multi-step projects on your behalf and with minimal need for your intervention. The growth in the use of Agents quickly makes inference the dominant AI workload. —Rodrigo Liang, CEO, SambaNova
Barriers ahead: Costs, LLM limitations, and more
In 2025, there will be a resurgence of old fashioned selling (phone calls still work!) and the emergence of B2B social selling with influencers dominating. Human opinions (or the appearance of them) will influence purchasing decisions more than ever before. And while social selling will largely trump AI tactics, watch for AI-powered voice as the wave of the future—with AI chatting up unsuspecting prospects on the phone and booking meetings at an unprecedented clip. —Stevie Case, chief revenue officer, Vanta
We are going to see some of the AI startups lose momentum in the market as their growth turns to churn if they don’t look at their books closely. They need to understand now what is truly sustaining revenue and what’s not, and adjust their business model accordingly. —Bhaskar Roy, Chief of AI Products & Solutions, Workato
We are all waiting for 2025 to bring down the cost of AI. By then, data center construction will increase, AI chip technology will advance, and competition in chip production will grow. —Tracy Young, CEO and cofounder, TigerEye
LLM-based AI hits a wall. There are early signs that scaling of transformer-based LLM is hitting a wall. That doesn’t mean AI isn’t getting better anymore; we have years of substantial incremental improvements ahead of us, but it does mean that the application layer above is settling in, and the most immediate wins are coming from innovation in user experience built on top of AI. —Malte Ubl, CTO, Vercel
By 2025, companies that view AI as a simple replacement for human workers will face significant challenges. Instead of achieving greater efficiency, these companies will likely experience a loss of their most valuable employees. These talented individuals won’t want to work in an environment where they feel undervalued or replaced by machines. —Frank Roe, CEO, SmartBear
Deepfakes will become an increasingly large problem in the age of AI. New tools will emerge to help us separate the truth from the AI-generated content we will be bombarded with. —Erica Brescia, managing director, Redpoint
The commercialization and increasing popularity of AI have caused energy demand at data centers to skyrocket to untenable levels. Over the next 12 to 36 months, as a greater number of companies seek out energy-saving alternatives. —Grant Verstandig, founder, Red Cell Partners
See you Monday,
Joey Abrams curated the deals section of today’s newsletter. Subscribe here.
- Swave Photonics, a Leuven, Belgium-based developer of holographic technology for augmented reality glasses, raised €27 million ($28.3 million) in Series A funding. imec.xpand and SFPIM Relaunch led the round and were joined by EIC Fund, IAG Capital Partners, Murata Electronics North America, Inc. and existing investors Qbic Fund, PMV, and Luminate.
- Rembrand, a Palo Alto, Calif.-based developer of AI-powered product placement technology, raised $23 million in Series A funding. super{set} led the round and was joined by The Trade Desk, Naver D2SF, and existing investors.
PRIVATE EQUITY
- Highlander Partners acquired Ergobaby, a Los Angeles, Calif.-based developer of baby carriers. Financial terms were not disclosed.
- Funds advised by Permira agreed to acquire a majority stake in Westbridge Advisory, a Frankfurt, Germany-based energy and sustainability firm for real estate companies. Financial terms were not disclosed.
OTHER
- Thomson Reuters acquired SafeSend, an Ann Arbor, Mich.-based developer of technology for tax and accounting professionals, for $600 million in cash.
