CB Insights

CoreWeave’s IPO is already oversubscribed — but success post-listing isn’t guaranteed. According to ChatCBI, which combines our research and data with genAI, CoreWeave faces four key risks: lack of client diversification, competition, integration issues, and market volatility. With AI infrastructure spend set to top $1T, CoreWeave is well-positioned — but it’s not the only player racing to capture market share. See how we’ve mapped the full AI data center value chain at the link in comments.

The future of the AI race will be determined by whoever owns the hardware. SoftBank is making aggressive moves across the AI stack — over 50% of its deals in the last two years have been in AI and the hardware to support it. With the Ampere acquisition, it’s doubling down on the infrastructure layer. You can explore the broader AI hardware landscape in our latest Market Map. Link in comments.

AI agents are moving toward greater specialization. Nearly half the private AI agent landscape is targeting horizontal applications like customer support, coding, and sales — securing $3.5B across 149 deals since 2020. But differentiation is becoming critical. To stay ahead in 2025, companies need more than basic AI features — they have to integrate deeper into customer workflows and make the most of proprietary data. Players like Sierra ($20M ARR), Cursor ($100M ARR), and Glean ($100M ARR) are gaining traction by tightly integrating with enterprise systems, making their AI agents harder to replace. Get the lay of the land below:

Investment patterns in clinical development AI have shifted significantly, with annual equity funding falling from $1.9B in 2021 to $0.7B last year. But despite reduced capital availability, two signals indicate activity in the AI-driven clinical drug development sector isn’t slowing down. Companies are expanding, with headcount growing 14% on average last year. At the same time, new company formation is experiencing a rebound, driven by growing pharma demand for AI tools that speed up and streamline clinical trials. See the data in the image below.

AI is rapidly reshaping insurance, from underwriting to disaster response. Based on CB Insights data, we expect 6 trends to redefine the insurance space in 2025. Here are 3 of them: ??France becomes the insurtech capital of the EU: Dealmaking to France-based insurtechs, combined with the country’s broader generative AI innovation, will outpace other EU nations. ??Auto insurers drive more telematics consolidation: Insurers will embrace tech M&A to strengthen connected vehicle capabilities and prepare for autonomous driving. ??Breakthroughs in AI weather prediction reduce catastrophe losses: Advanced forecasting models will enable insurers to better anticipate and mitigate losses from extreme weather events. Explore all 6 of our predictions in the image below.

As hospitals look to streamline workflows and reduce administrative burdens, AI-powered documentation is becoming essential. Clinical documentation tools — which generate real-time medical notes to reduce paperwork and clinician burnout — secured 4 of healthcare’s 5 largest generative AI deals last year while gaining significant commercial traction. That momentum has carried into 2025. For one, Abridge raised the largest deal in 2024 and followed it up with a $250M Series D just last month. Meanwhile, it forged business relationships with 14 health systems last year and has already worked with several more in Q1’25. See other top deals in the space below.

Small teams are getting big payouts. So far in 2025, tech companies acquired for $100M+ had just 100 employees at the median, according to our M&A transaction and headcount data. Many of these companies are young, with an average age of seven years, and have raised less than $20M in equity funding. Join us on March 18 for our Tech M&A Predictions for 2025, where we’ll break down: ?? The most active acquirers in tech ?? Which companies are likely acquisition targets ?? Key trends shaping the next wave of deals Follow the link in the comments to save your spot for the briefing.

From clinical documentation to drug discovery, AI is tackling two of the biggest challenges in healthcare & life sciences: soaring costs and staffing shortages. Generative AI is streamlining workflows, automating high-volume tasks, and alleviating strain on overburdened staff. Nearly two-thirds (65%) of healthcare organizations are already considering or implementing genAI solutions. We mapped 85+ startups bringing genAI to clinical operations, revenue cycle management, and drug development.

AI agents are climbing the autonomy scale. Right now, most operate with guardrails, relying on constrained architectures like decision trees to complete tasks. Fully autonomous agents remain limited due to challenges in reliability, reasoning, and access — but that’s changing. As the core mechanisms — reasoning, memory, tool use, and planning — improve, AI agents will evolve beyond simple copilots, adapting to complex workflows with minimal human intervention. See where AI agents stand today in the framework below, and get our full outlook on the future of AI agents in the comments.

AI agents need trust to thrive. Reliability and security are the top concerns in AI agent deployment. Nearly half (47%) of organizations we recently surveyed — from SMEs to large enterprises — cite these as key challenges. Based on our briefings with 20+ AI agent companies in Q1’25, vendors are using five key strategies to build trust: ?? Transparency ?? Human oversight ??? Safeguards ?? Security & compliance ?? Continuous improvement See how startups are addressing these challenges in the chart below.