Deals downstream.

Innovaccer's $275 raise.

FTC vs. Uber.

 

View in browser

April 22, 2025

Vital signs

Hi there, 

 

Digital health came roaring back in Q1, with venture funding, M&A, and unicorn births all climbing to multi-year highs. 

 

AI is playing an outsized role in the sector’s rebound.


AI companies grabbed 60% of all digital health equity funding in the quarter — a new record — to drive total funding up 47% QoQ to $5.3B.

ai startups pulled in 60% of q125 digital health funding

This isn’t just broad market enthusiasm for AI — it’s a sign that AI is evolving from horizontal to vertical applications that can move the needle in industries like healthcare. 

 

For instance, specialized AI models — trained on clinical reasoning and domain-specific knowledge — demonstrate higher accuracy in healthcare workflows. Healthcare tech players can use these to extract more accurate insights across their datasets and increase providers’ operational efficiency.

 

Below, we break down how AI is reshaping the digital health landscape using CB Insights data.

 

Digital health is maturing

 

The sector’s momentum is shifting downstream, with investors concentrating capital in later-stage startups that are scaling AI-powered solutions.

 

Early-stage deals have declined from 60% of total activity in full-year 2024 to 51% in 2025 so far, while mid- and late-stage deals have seen their shares increase by several percentage points.

 

The median size of late-stage deals has nearly doubled — from $37M in 2024 to $73M this year — as investors prioritize companies with regulatory milestones and clear commercial traction.

 

The biggest late-stage deal of Q1 went to Innovaccer, which raised a $275M Series F in January from investors including Kaiser Permanente, Banner Health, and Microsoft’s M12.

 

The healthcare data platform now serves 6 of the US’ top 10 healthcare systems. Following its Series F, it rolled out a suite of AI agents in February that aim to automate processes like intake, referral, and prior authorization. 


We recently featured Innovaccer in the healthcare category of our AI agent market map.

the ai agent market map

For more companies like Innovaccer, customers can see 370+ high-potential AI startups in digital health on the CB Insights platform.

 

Unicorns are back

 

AI is also driving a resurgence in digital health unicorn births.

 

6 new digital health unicorns were minted in Q1’25 — the most in a single quarter since Q2’22 — and half of these are applying AI to provider workflows: 

  • Hippocratic AI (patient follow-up — also featured in the AI agents market map above)
  • Abridge (clinical documentation)
  • OpenEvidence (healthcare decision-making)

Today's digital health unicorns are also leaner, according to CB Insights headcount data. 

 

The 6 new unicorns in Q1 have an average of just 176 employees, compared to 423 for digital health unicorns born during the 2021-2022 boom. OpenEvidence reached the milestone in February with merely 21 team members.

more unicorns are born in q125 than in all of 2024

Get the data behind digital health’s healthy quarter in our just-released State of Digital Health Q1’25 Report.

 

CBI customers can go deeper with our report series on AI in drug R&D. We break down how the tech is transforming each stage of the clinical pipeline:

  • AI in drug discovery
  • AI in preclinical development
  • AI in clinical development

TLDR — Tech loves drama, right?

Here's a roundup of recent tech drama:

  • Zuck on the stand: Meta CEO Mark Zuckerberg began testifying in a high-stakes FTC antitrust trial last week, defending the company’s acquisitions of Instagram and WhatsApp. The FTC argues the deals were meant to squash emerging rivals and maintain a social media monopoly. If the court agrees, Meta could be forced to spin off both platforms — a move that would upend its $160B ad business. Zuckerberg said the deals were aboveboard and long approved, while the FTC pointed to internal emails showing Meta feared Instagram’s rise. Testimony also touched on Meta’s market power, user dominance, and shift from “friends” to content discovery.
     
  • Palantir pushback: Palantir’s $30M contract with ICE to build “ImmigrationOS” — a tool for identifying and tracking deportation targets — reignited scrutiny over the company’s role in government surveillance. After Y Combinator founder Paul Graham slammed it as “police state infrastructure,” Palantir exec Ted Mabrey fired back, citing national security work like Operation Fallen Hero and calling Graham’s critique a “courtroom parlor trick.” Mabrey doubled down on Palantir’s mission — and its effort to recruit “true believers” willing to weather political attacks.

  • Name drop: ChatGPT has started greeting users by name — even when they haven’t shared it or enabled memory. The unexpected shift has creeped out users, who describe the behavior as “fake,” “invasive,” and “uncanny.” OpenAI hasn’t confirmed whether it’s a glitch or an early rollout of its personalization features, but the backlash highlights the risks of humanizing bots too much, too fast. Experts say names signal intimacy — and when an AI gets familiar without warning, it can feel less like personalization and more like a privacy violation.
simon willison tweet about chatgpt name use

Source: X

  • Cancel culture: The FTC is suing Uber, accusing it of sneakily signing customers up for its Uber One subscription without consent — and then making cancellation a maze of 23 screens and 32 taps. The agency says promised savings were misleading and Uber charged some users mid-trial. Uber denies wrongdoing, claiming its sign-up and cancel processes are now streamlined. The lawsuit lands just weeks before the FTC’s new “click to cancel” rule kicks in. With Uber One at 30M members and $1B+ in revenue last year, the case could dent one of Uber’s fastest-growing products.
     
  • DKIM deception: A sophisticated phishing campaign exploited a gap in Google’s email system to send spoofed messages from “no-reply@google.com” that passed DKIM authentication. The attack tricked users with a fake Google security alert and a nearly identical support portal hosted on sites.google.com. Ethereum Name Service developer Nick Johnson uncovered the scam, which leveraged Google’s OAuth infrastructure to make the message appear legitimate. Despite early pushback, Google now acknowledges the risk and is working on a fix. A similar DKIM exploit was previously used in PayPal phishing schemes.

    I love you.

     

    Anand

    @asanwal 

    Co-Founder & Exec Chair

     

    P.S. Tired of reading? Hear directly from our analysts on the AI opportunity in pharma with this live briefing. 

    Get started with CB Insights

    Start your free trial

    CB Insights' emerging technology insights platform provides all the

    analysis and data from this newsletter. Our data is the easiest way to discover and respond to emerging tech. 

    Was this email forwarded to you? Sign up here

    X
    LinkedIn
    CB-Insights-Icon-Light

    Copyright © 2025 CB Insights, All rights reserved.

    498 7th Avenue, NY, CB Insights, New York,10018

    About Us | Update Preferences | Research | Newsletter