November 11, 2021: Correction: The original version of this story reported incorrectly that the venture fund Venrock invested in Akili, Woebot and Lyra Health; in fact the company is only invested in Lyra. (You can report errors in stories here and find corrections on our Corrections page.)

In 2013, Megan Jones-Bell, an associate professor of psychology at Stanford, founded Lantern, a digital behavioral health company aimed at addressing anorexia and binge eating disorders as well as anxiety and depression. She had spent years researching eating disorder treatments – an illness she suffered as a teen – and made use of the science underpinning cognitive behavioral therapy (CBT), a decades-old approach that uses mood logs, mindfulness, and behavior modification and is widely accepted as effective against multiple conditions. What intrigued her was that CBT had recently been found to be at least as effective when delivered digitally, suggesting a more cost-effective and accessible therapy. 

But financing Lantern was brutal. “All the VCs said, ‘An app to address mental health – that’s ridiculous,’” she recounted in a recent interview. She eked out funds to launch but struggled with follow-on financing and traction. Companies balked at her pitch to buy Lantern for their employees. “I had so many meetings with benefits directors. They didn’t think it was their job to offer anything beyond the health insurance benefit,” she said.  I interviewed Jones-Bell back then for a 2017 story on the nascent field of mental health apps. The next year, Lantern folded. “There was just no market,” she said. 

In the four years since, digital behavioral health has exploded. Today, there are upwards of 10,000 apps and 1,000 start-ups, and a swift business of employers buying phone and computer-based therapies for employees. Concepts range from asynchronous text therapy and group meditation to on-command ambiance therapy – “Alexa, tell Soaak to boost my mood” – and even mail-order ketamine plus virtual coach for your at-home trip. The top players are selling to employers and providing care via a stepped delivery model, which means deploying coaches and escalating to therapists and psychiatrists when needed.

Investment is at “frenzy” level, says Amanda Eilian of Able Partners, a backer of Spring Health, Ksana and Ellipsis. Venture investors, often with no background in healthcare, are in a mad brawl to beat each other into bigger and bigger funding rounds. “There’s a lot of dumb money rushing in,” says What If Ventures founder Stephen Hays, whose fund focuses exclusively on mental health.

The average deal size for later-stage digital behavioral health companies nearly tripled to $64.2 million between 2016 and 2020, according to Rock Health, and Investors poured $2.4 billion into this market in 2020. The 2021 figure is tracking to surpass that, according to Crunchbase. Since June, Cerebral has raised $127 million for medication management, NOCD $33 million for virtual exposure and response therapy, Woebot $90 million for AI-powered chatbot therapy, Sondermind $150M for video telehealth, and QuitGenius $64 million to fight addiction. Mental health is outraising all other health categories in attracting VC investment.

Venture capitalists have incentive to dream big: Since 2019, seven mental health unicorns have been created, five in 2021 alone. The farthest along is Lyra Health, which doubled its customer base to 2 million in 2020 and raised $200 million – with a $4.6 billion valuation – in June 2021. Modern Health, with a $1.2 billion valuation, became the fastest ever female-founded company to hit unicorn status. Also in the club: BetterUp (fronting Prince Harry as its “Chief Impact Officer”); Pear Therapeutics; Calm; Talkspace, which went public via a gimmicky financing vehicle; Spring Health; and Ginger, which merged with Headspace and in January reported a 335% uptick in demand over pre-COVID 2019 averages.

Photo: Shutterstock

The pandemic effect 

The pandemic supercharged the category. As isolation, economic volatility and uncertainty spiked, America’s mental health tanked. By January 2021, 41 percent of adults reported symptoms of depression or anxiety, a four-fold increase from 2019, according to NHIS and the U.S. Census Bureau. One quarter of 18-to-24-year-olds reported having suicidal thoughts in the past 30 days in 2020, the CDC reported. 

The added burden exposed a system in duress. Sixty percent of US counties do not have a single psychiatrist. For OCD sufferers, just a few hundred therapists deliver exposure and response prevention, the gold-standard treatment. The average wait time for a mental health visit is 25 days, according to a 2015 study in Psychiatric Services. Meanwhile, the US suicide rate has jumped 33 percent since the ‘90s, the CDC reports. Add to this mix that just a small minority of therapists accept health insurance, and you get a massive supply-demand imbalance. It’s not surprising, then, that over half of U.S. adults with a mental health condition received no treatment in the previous year, National Institute for Health data shows. The deficit rises to 90% in under-resourced communities

Amidst widespread distress, employers had to toss out a safety net – and a mental health benefit became an expectation. “Every workplace needs to offer health, dental, vision— and now behavioral. It’s the fourth pillar,” says Venrock partner Bryan Roberts, whose firm invested in Lyra Health. It’s not just coastal tech companies providing this benefit. Industries based in the US heartland – airlines, accounting firms – now offer Lyra, Ginger, Modern Health, and more to their workers. When COVID hit, Reno, Nevada Mayor Hillary Schieve offered Talkspace to her entire 200,000-person town. 

With life migrating onto Zoom, virtual health solutions became a necessity. Suddenly seeing a psychologist via iPhone, or having Prozac prescribed in a video visit, felt normal. Private and public insurers began covering telehealth, including teletherapy. Telehealth claims received by Blue Cross Blue Shield of Massachusetts jumped from 200 to 38,000 in the three months after the March 2020 lockdown. 

Safety net clinics also pivoted to offer counseling, therapy and support groups online for their low-income patients. For many, especially those in far-flung rural areas or struggling with recovery from opioids, virtual therapy and support groups were a lifeline. For patients, a cell phone or computer meant they could attend sessions anywhere, from a tent in a homeless encampment to a car or even the only private room in some houses: a bathroom. The New York Times captured the cultural shift in this headline: “In Psychotherapy, the Toilet Has Become the New Couch.” 

A widening digital divide

While VCs talk about access for all, low-income people are often left behind. Few work in the companies providing these therapies as a benefit, and few of the upstarts target people on Medicaid, the federal health insurance program for people with low incomes or disabilities. “Most of the behavioral health companies we talk to — they are not prioritizing Medicaid,” says Melissa Buckley, director of California Health Care Foundation’s Health Innovation Fund, which funds digital solutions for the safety-net population. “They say let us figure out the other (markets) first,” says Buckley.

Although almost 76 million people are enrolled in Medicaid, “the financial incentives are not there,” says Stephen Schueller, associate professor of psychological science and informatics at UC Irvine, who researches the care gap. 

Communities of color suffer from mental illness at higher rates and are less likely to get care — the proportion of African Americans getting needed mental health treatment is only half that of whites. Virtual therapies offer new promise, but also erect new barriers. While smartphones are increasingly ubiquitous—58 percent of homeless people have one, a 2017 study found—access is still spotty. “They might have a smartphone, but not a data plan, or minutes,” says Buckley. 

App creators often don’t include these populations in their development process, so they too often assume an inappropriate level of digital literacy, health literacy, or English proficiency. Efficacy studies are performed overwhelmingly on older white women. “The therapeutic system we have was designed for middle class white families,” says Hurdle founder Kevin Dedner, who heads one of a handful of companies working to offer an exception. Hurdle is providing care to the Black community using “culturally intentional teletherapy.” He says the “[mental health] revolution won’t have success unless we deal with the approach to care and (unless) therapists are prepared to serve people with diverse backgrounds.” 

Relaxation apps “going after what is easy” 

Another kind of digital divide is also growing. Since almost everyone feels stressed and isolated, the stigma connected with having a little anxiety or depression has faded, fueling the demand for wellness apps aimed at people with subclinical, mild and moderate diagnoses. Companies like Ginger offers services “like a gym, to stay in shape,” says Khosla Ventures general partner Samir Kaul, an early Ginger investor. Sail sells its service – primarily virtual support groups for men – as “Peloton for the mind.”

But what about the one in 20 U.S. adults who contend with serious mental illness (SMI) — conditions like bipolar, major depression and schizophrenia? For the most part, there’s no room at the wellness-app inn. Just look at the disclaimers posted by most of these companies. “ does not provide medical treatment, advice or diagnosis and is not a healthcare provider,” says that company’s terms of service. Woebot Health calls itself a “pure self-help” program whose services “do not include the provision of medical care, mental health services or other professional services.” 

“What isn’t happening is somebody saying how do we find people with schizophrenia and keep them from going to jail, or how do we keep those with bipolar out of crisis,” says Tom Insel, former director of the National Institute for Mental Health and an entrepreneur and advisor in mental health start-ups. (Insel is also a donor to MindSite News and chair of its editorial advisory board.) 

Jones-Bell, who was plucked up by Headspace after Lantern shuttered, is disillusioned by “apps that are essentially relaxation being marketed as impacting depression and anxiety.  What’s happening to those people with real issues?” Most recently, she joined Google to head up the tech giant’s own digital mental health effort as clinical director of consumer and mental health. 

Ironically, people with serious disorders represent an outsized investment opportunity. Their treatment is expensive and much of it is taxpayer funded, with an average lifetime cost of $1.85 million, a study in Health Affairs found. But effectively treating their complex needs requires integrated care tailored to individual needs. That makes it difficult to bring solutions to scale – and scale is how VCs capture value. Many people with serious mental illness are on Medicaid, so reimbursement requires tangling with the government and anemic reimbursement rates.

A lot of companies and investors “are going to go after what is easy,” says What If Venture’s Hays. “It’s not just coming up with the right app or pill or better mousetrap. We have to figure out the business model.” Still, he says, even helping people incrementally saves a lot of money. “If they go to the emergency room one less time, it’s a significant cost-savings.” 

A ‘therapeutic alliance’ with Woebots and digital penguins? 

The digital revolution has also brought us AI-powered bots represented by Wysa, a pudgy blue penguin, and Woebot, a block-headed robot, as friendly therapist avatars. These apps are based on well-tested CBT techniques but the interaction happens on a screen, without the involvement of another human, through a feat of natural language processing and programming. Woebot claims that users achieve a robust “therapeutic alliance” with the bot that “appeared to be non-inferior to the bond created between human therapists and patients.” These claims to connection were determined with no control group, based on reaction to phrases including “Woebot and I respect each other” and “I believe Woebot cares about me even when I do things it does not approve of.” 

The bot-therapist is in: Wysa introduces themself to the author…

When I downloaded the app and texted Woebot that I felt anxious, it offered calming exercises. In one, Woebot’s belly expanded and contracted, and I was to match my breathing to the rhythm. Another time, Woebot had me list out negative thoughts and imagine they had a voice. When I texted I was low on talent, Woebot asked me to reframe. The approach may be rooted in CBT, but with therapy chopped into bite-size nibbles instead of 50-minute sessions, it’s hard not to wonder: Is this still CBT?

Woebot Health founder Allison Darcy argues that 10-minute sessions, five times a week, mirror the traditional approach. “We’re trying to meaningfully translate tools that have been the unique domain of professionals and make it more usable so people can take care of their mental health on a daily basis,” she says. However, a 2021 review of 98 CBT-based apps in the Journal of Medical Internet Research found just one-third actually contained comprehensive CBT programs. “It’s an

Woebot spots distorted thinking in our author’s responses…

exciting idea. But I do wonder really if people are going to get the therapy they need if it’s done while stuck in traffic or waiting in line for a sandwich,” says Jesse Wright, MD, who directs the University of Louisville Depression Center. 

It’s not enough to download the app 

In July, on her podcast Sway, Kara Swisher pressed Talkspace founder Oren Frank about Talkspace’s claim to have 55 million people covered by their employer, yet only 60,000 active users — an uptake rate of .01%. His response: “That’s exactly why we went public and why we want to scale.” 

Uptake remains the Achilles heel. Novel solutions will cure no one unless they are used. “It’s not enough to sell me a subscription. Are you selling me one that I’m taking, and that is making me better?” says John Torous, MD, director of the digital psychiatry division at Beth Israel Deaconess Medical Center. One quarter of downloaded health apps get just a single use, and 75% achieve ten uses maximum, a Localytics study found. Four months after Mayor Schieve made Talkspace available to Reno’s 200,000 residents, just 1,357 had logged on

...and praises our author’s strong insights!

Poor compliance isn’t just a feature of digital therapy. Only half of patients with chronic diseases follow the therapies or take the medications they are given. Pokemon Go was briefly able to excite users enough to boost their daily steps but after six weeks, the walkers had all slipped back to their previous sedentary habits. 

Even more unsettling is the dearth of proven efficacy. I asked Insel to name some companies where the funding is out ahead of the science. “All of them,” he said. 

Among 293 apps for anxiety and/or depression studied in a 2021 World Psychiatry meta-analysis, just 6 percent had published evidence citing their efficacy. “There are thousands and thousands of mental health apps that have no basis in science,” Hays says. Too often start-ups “take a small pilot study and extrapolate that it can work for huge populations,” Torous concurs. 

Able Venture’s Eilian concedes that the field is “in the early innings,” but is convinced the new attention will advance the science. “Now that there is venture money behind it, we can see more progress,” she says.

Loose proof has long been a central problem for mental health treatments, one further magnified in the digital space. Outcomes often go unmeasured. Users are rarely asked if they feel better, and whether the therapy was responsible. “It’s a friggin’ disaster,” Insel says. “Something like 80 percent of providers don’t measure what they do” – a metric he applies to both digital and in-person care. 

One-click platform for antidepressant prescribing

The digital nature of services can also magnify other flaws. Cerebral – which launched last year, is now valued at $1.2 billion and just signed Simone Biles as a spokesperson – is essentially a one-click platform for prescribing antidepressants with the potential to further accelerate their prescribing. Since Prozac debuted in 1987, the percentage of Americans who took at least one antidepressant in the prior 30 days has increased by nearly 500%. Yet Prozac and other antidepressants only slightly outperform placebo and the prevalence of depression has not dropped, according to a 2021 study in Behavior Research and Therapy. Yet hot, VC-backed Cerebral makes money by writing more prescriptions. “How is that not a perverse incentive?” huffs Able Venture’s Eilian.  

After visiting the Cerebral website, an email pinged into my inbox: “TODAY ONLY: Get 70% OFF.” Is a full-court press toward drugs for anxiety and depression societal progress? Is a monthly check-in with a “care counselor,” who may or may not be a licensed practitioner, adequate for overseeing medication management? 

I emailed Steven Hollon, the Gertrude Conaway Vanderbilt Professor of Psychology at Vanderbilt University and an expert in the treatment of adult depression. He characterized services like Cerebral as “a double-edged sword.” He’s all for increasing access but notes that “we probably have half the population walking around incorrectly medicated. It’s a little scary.” 

Graphical user interface, text, application

Description automatically generated with medium confidence

Talkspace, a newly public company repped by Olympic gold-medalist Michael Phelps, provides coaches for daily texting, and therapists and psychiatrists when needed. The company website’s research page lists eight studies as evidence for the effectiveness of its approach, but only one directly studied TalkSpace itself. Referring to that study, the site states in large letters: “EFFECTIVENESS 80%.” The study itself had no control group and its principal finding was that 46% of the 57 participants experienced “clinically significant symptom remission.” 

I reached out to researcher Thomas Hull, formerly VP of Clinical R&D at Talkspace and now an independent consultant, about the results. He told me to ignore the linked study and instead look to a large study in BMC Psychiatry, also funded by the company but oddly not on their research page. Here, 48% and 53% of 10,718 users saw clinically significant improvements using two symptom scales, respectively. A far cry from 80% effectiveness.

The Talkspace home page also boasts a colorful list of benefits in bullet points. What no bullet point claims is that Talkspace will cure depression, reduce anxiety, or keep them both at bay. The words of John Torous come to mind: “There’s an implicit assumption that the app is better than nothing. But what if it isn’t?”

Aditi Malhotra contributed reporting for this story.

MindSite News’ coverage of mental health issues in California is supported by the California Health Care Foundation.

Type of work:



Correction: The original version of this story reported incorrectly that the venture fund Venrock invested in Akili, Woebot and Lyra Health; in fact the company is only invested in Lyra. (You can find our Corrections page here and our best practices in our Mission, Ethics and Guidelines statement here.)

Diana Kapp is a journalist and the author of Girls Who Run the World: 31 CEOs Who Mean Business (Random House 2019). Her writing work has taken her inside San Quentin prison and to deepest Afghanistan....