If the Covid era marked a boom time for digital health companies, 2024 was the reckoning.

In a year that saw the Nasdaq jump 32%, surpassing 20,000 for the first time this month, health tech providers largely suffered. Of 39 public digital health companies analyzed by CNBC, roughly two-thirds are down for the year. Others are now out of business.

There were some breakout stars, like Hims & Hers Health, which was buoyed by the success of its popular new weight loss offering and its position in the GLP-1 craze. But that was an exception.

While there were some company-specific challenges in the industry, overall it was a “year of inflection,” according to Scott Schoenhaus, an analyst at KeyBanc Capital Markets covering health-care IT companies. Business models that appeared poised to break out during the pandemic haven’t all worked as planned, and companies have had to refocus on profitability and a more muted growth environment.

“The pandemic was a huge pull forward in demand, and we’re facing those tough, challenging comps,” Schoenhaus told CNBC in an interview. “Growth clearly slowed for most of my names, and I think employers, payers, providers and even pharma are more selective and more discerning on digital health companies that they partnered with.” 

In 2021, digital health startups raised $29.1 billion, blowing past all previous funding records, according to a report from Rock Health. Almost two dozen digital health companies went public through an initial public offering or special purpose acquisition company, or SPAC, that year, up from the previous record of eight in 2020. Money was pouring into themes that played into remote work and remote health as investors looked for growth with interest rates stuck near zero.

But as the worst waves of the pandemic subsided, so did the insatiable demand for new digital health tools. It’s been a rude awakening for the sector.  

“What we’re still going through is an understanding of the best ways to address digital health needs and capabilities, and the push and pull of the current business models and how successful they may be,” Michael Cherny, an analyst at Leerink Partners, told CNBC. “We’re in a settling out period post Covid.”

Progyny, which offers benefits solutions for fertility and family planning, is down more than 60% year to date. Teladoc Health, which once dominated the virtual-care space, has dropped 58% and is 96% off its 2021 high.

When Teladoc acquired Livongo in 2020, the companies had a combined enterprise value of $37 billion. Teladoc’s market cap now sits at under $1.6 billion.

GoodRx, which offers price transparency tools for medications, is down 33% year to date. 

Schoenhaus says many companies’ estimates were too high this year.

Progyny cut its full-year revenue guidance in every earnings report in 2024. In February, Progyny was predicting $1.29 billion to $1.32 billion in annual revenue. By November, the range was down to $1.14 billion to $1.15 billion.

GoodRx also repeatedly slashed its full-year guidance for 2024. What was $800 million to $810 million in May shrank to $794 million by the November.

In Teladoc’s first-quarter report, the company said it expected full-year revenue of $2.64 billion to $2.74 billion. The company withdrew its outlook in its second quarter, and reported consecutive year-over year declines.

“This has been a year of coming to terms with the growth outlook for many of my companies, and so I think we can finally look at 2025 as maybe a better year in terms of the setups,” Schoenhaus said.  

While overzealous forecasting tells part of the digital health story this year, there were some notable stumbles at particular companies. 

Dexcom, which makes devices for diabetes and glucose management, is down more than 35% year to date. The stock tumbled more than 40% in July – its steepest decline ever – after the company reported disappointing second-quarter results and issued weak full-year guidance. 

CEO Kevin Sayer attributed the challenges to a restructuring of the sales team, fewer new customers than expected and lower revenue per user. Following the report, JPMorgan Chase analysts marveled at “the magnitude of the downside” and the fact that it “appears to mostly be self-inflicted.” 

Genetic testing company 23andMe had a particularly rough year. The company went public via a SPAC in 2021, valuing the business at $3.5 billion, after its at-home DNA testing kits skyrocketed in popularity. The company is now worth less than $100 million and CEO Anne Wojcicki is trying to keep it afloat.

In September, all seven independent directors resigned from 23andMe’s board, citing disagreements with Wojcicki about the “strategic direction for the company.” Two months later, 23andMe said it planned to cut 40% of its workforce and shutter its therapeutics business as part of a restructuring plan. 

Wojcicki has repeatedly said she intends to take 23andMe private. The stock is down more than 80% year to date. 

Digital health’s bright spots

Investors in Hims & Hers had a much better year.

Shares of the direct-to-consumer marketplace are up more than 200% year to date, pushing the company’s market cap to $6 billion, thanks to soaring demand for GLP-1s. 

Hims & Hers began prescribing compounded semaglutide through its platform in May after launching a new weight loss program late last year. Semaglutide is the active ingredient in Novo Nordisk‘s blockbuster medications Ozempic and Wegovy, which can cost around $1,000 a month without insurance. Compounded semaglutide is a cheaper, custom-made alternative to the brand drugs and can be produced when the brand-name treatments are in shortage.

Hims & Hers will likely have to contend with dynamic supply and regulatory environments next year, but even before adding compounded GLP-1s to its portfolio, the company said in its February earnings call that it expects its weight loss program to bring in more than $100 million in revenue by the end of 2025. 

Doximity, a digital platform for medical professionals, also had a strong 2024, with its stock price more than doubling. The company’s platform, which for years has been likened to a LinkedIn for doctors, allows clinicians to stay current on medical news, manage paperwork, find referrals and carry out telehealth appointments with patients. 

Doximity primarily generates revenue through its hiring solutions, telehealth tools and marketing offerings for clients like pharmaceutical companies.

Leerink’s Cherny said Doximity’s success can be attributed to its lean operating model, as well as the “differentiated mousetrap” it’s created because of its reach into the physician network. 

“DOCS is a rare company in healthcare IT as it is already profitable, generates strong incremental margins, and is a steady grower,” Leerink analysts, including Cherny, wrote in a November note. The firm raised its price target on the stock to $60 from $35. 

Another standout this year was Oscar Health, the tech-enabled insurance company co-founded by Thrive Capital Management’s Joshua Kushner. Its shares are up nearly 50% year to date. The company supports roughly 1.65 million members and plans to expand to around 4 million by 2027. 

Oscar showed strong revenue growth in its third-quarter report in November. Sales climbed 68% from a year earlier to $2.4 billion.

Additionally, two digital health companies, Waystar and Tempus AI, took the leap and went public in 2024. 

The IPO market has been largely dormant since late 2021, when soaring inflation and rising interest rates pushed investors out of risk. Few technology companies have gone public since then, and no digital health companies held IPOs in 2023, according to a report from Rock Health. 

Waystar, a health-care payment software vendor, has seen its stock jump to $36.93 from its IPO price of $21.50 in June. Tempus, a precision medicine company, hasn’t fared as well. It’s stock has slipped to $34.91 from its IPO price of $37, also in June.

“Hopefully, the valuations are more supportive of opportunities for other companies that have been lingering in the background as private companies for the last several years.” Schoenhaus said. 

Out with the old

Several digital health companies exited the public markets entirely this year. 

Cue Health, which made Covid tests and counted Google as an early customer, and Better Therapeutics, which used digital therapeutics to treat cardiometabolic conditions, both shuttered operations and delisted from the Nasdaq. 

Revenue cycle management company R1 RCM was acquired by TowerBrook Capital Partners and Clayton, Dubilier & Rice in an $8.9 billion deal. Similarly, Altaris bought Sharecare, which runs a virtual health platform, for roughly $540 million.

Commure, a private company that offers tools for simplifying clinicians’ workflows, acquired medical AI scribing company Augmedix for about $139 million.

“There was a lot of competition that entered the marketplace during the pandemic years, and we’ve seen some of that being flushed out of the markets, which is a good thing,” Schoenhaus said.

Cherny said the sector is adjusting to a post-pandemic period, and digital health companies are figuring out their role.

“We’re still cycling through what could be almost termed digital health 1.1 business models,” he said. “It’s great to say we do things digitally, but it only matters if it has some approach toward impacting the ‘triple aim’ of health care: better care, more convenient, lower cost.”

President-elect Donald Trump is tapping tech heavyweights to join his new administration, continuing a trend of Silicon Valley’s growing influence in a second Trump White House.

Trump said Sunday he would nominate Scott Kupor, a managing partner at Andreessen Horowitz, to be director of the Office of Personnel Management, which coordinates recruitment and provides resources for government employees.

Kupor thanked Trump in a post on X and said the opportunity would allow him to work with Elon Musk and Vivek Ramaswamy in their leadership of the Department of Government Efficiency, or DOGE, a nascent commission aimed at cutting government spending and regulation.

Trump also picked Sriram Krishnan as senior policy advisor for artificial intelligence at the White House Office of Science and Technology Policy. Krishnan, who most recently served as a general partner at Andreessen Horowitz, has had a long career in tech, with roles at Microsoft, Meta, Twitter, Snap and Yahoo. He has previous ties to Musk, helping him “temporarily” run the social media service X after Musk acquired the platform, formerly known as Twitter, for $44 billion in 2022.

Musk, a tech billionaire who was one of Trump’s top donors and most vocal supporters during his campaign, has emerged as one of the president-elect’s closest advisors. His outsized influence over Trump has led to growing consternation among Democrats, foreign leaders and business executives, some of whom compete with Musk’s companies. Along with X, Musk runs vehicle maker Tesla, defense contractor SpaceX and brain tech startup Neuralink.

Krishnan will likely work closely with David Sacks, another tech executive who has a long history with Musk. Trump earlier this month named Sacks — a venture capitalist, former PayPal COO and popular podcaster — as “czar” of crypto and AI.

Trump on Sunday also tapped Ken Howery, a co-founder of PayPal and Founders Fund, as his pick for U.S. ambassador to the Kingdom of Denmark. And he appointed Michael Kratsios, who was most recently a managing director at tech startup Scale AI, as the director of the White House Office of Science and Technology Policy. Kratsios served as chief technology officer during Trump’s first administration.

In addition, Trump named former Uber executive Emil Michael as undersecretary for research and engineering.

Tech business leaders cheered the choices in social media posts. Former Meta executive David Marcus called Trump’s selections “remarkable picks,” while Box CEO Aaron Levie said the choices were “very strong.”

Since Trump’s election victory, a slew of tech companies have thrown their support behind the president-elect — a significant departure from his first term, when the industry at large maintained a tense relationship with Trump.

Amazon, Meta and OpenAI Sam Altman have announced donations of $1 million each to Trump’s inaugural committee. And in recent weeks, Silicon Valley executives have made pilgrimages to Trump’s residence Mar-a-Lago in Palm Beach, Florida.

“Aut Zuck Aut Nihil” spanned the front of Mark Zuckerberg’s loose-fitting black shirt during his keynote at Meta’s Connect event in September. 

The words, donned in all-caps and gray font, were a play on the Latin phrase “Aut Caesar Aut Nihil,” which translates to “Either Caesar or nothing” or rather “All or nothing.” It was a fitting phrase for a company that in 2024 put the full weight of its resources behind its artificial intelligence strategy. 

Meta in April said it would raise its spending levels in 2024 by as much as $10 billion to support infrastructure investments for its AI efforts. Although the announcement sent shares plunging as much as 19% that evening, investors have come around to the company’s costly AI ambitions. Meta’s stock price hit a record on Dec. 11, and it’s up nearly 70% year to date as of the market’s close on Friday.

“It’s clear that there are a lot of new opportunities to use new AI advances to accelerate our core business that should have strong ROI over the next few years, so I think we should invest more there,” Zuckerberg said on a call with analysts in October.

He noted AI’s “positive impact on nearly all aspects of our work,” highlighting how the technology was key to rebuilding the company’s online advertising business that took a lashing from Apple‘s iOS privacy update in 2021. Additionally, he said AI underpins Meta’s more nascent projects, such as its Ray-Ban Meta smart glasses and experimental Orion augmented reality headset that Zuckerberg believes could represent “the next computing platform.”

Zuckerberg’s comments about AI underscore how the technology has become Meta’s top priority, directly impacting the company’s business and potentially paving the way for future revenue opportunities. Unlike the company’s more conventional services, like Instagram and Facebook, AI is an infrastructure technology that Zuckerberg wants hardwired into its various products, particularly as competitors like OpenAI continue to make inroads with consumers.

While OpenAI’s GPT family of AI models help power apps like ChatGPT, Meta’s family of Llama AI models feeds the company’s newer generative AI features like the Meta AI digital assistant. That chatbot represents Zuckerberg’s primary way to introduce generative AI technologies to its billions of users.

“Meta AI is on track to being the most used AI assistant in the world by the end of this year,” Zuckerberg said at Connect. 

The company has been increasingly releasing new generative AI features for advertisers to continue improving the efficiency of its online advertising platform. And with the hiring last month of Clara Shih, who had been Salesforce’s CEO of AI, to lead a new business AI group, Meta aims to build a more enterprise-focused unit in the new year.

Meta’s all-encompassing approach to AI has led analysts to predict that Meta is positioned for more success in 2025.

Analysts at Jefferies chose Meta as one of generative AI’s “winners” heading into 2025, writing in a Dec. 15 note that the company’s massive user base represents “one of the richest surfaces to introduce Gen AI tools.” Truist Securities analysts said in a note last week that the Meta AI digital assistant could challenge Google’s search as “an answer engine for all kinds of queries” and that the social media company is likely to outperform in 2025, potentially benefiting from offering businesses more advanced customer service chatbots. 

“We believe META has a unique opportunity to introduce Gen AI tools to the almost 4B users & >200M businesses across its family of apps,” the Jefferies analysts wrote.

Meta declined to comment for this article, but pointed to previous statistics and executive comments about AI.

When it comes to monthly active users, the most popular of these generative AI chatbot apps is ChatGPT by a wide margin, followed by Google Gemini, Microsoft Copilot, Claude and Perplexity, Curry said.

The Meta AI “standalone website gets less than 10 million views per month, putting it far below the major services (ChatGPT, Gemini, etc) and even lower than some mid-range players like Anthropic,” Curry said, based on data he accessed via the Similarweb web-tracking service. 

Meta’s finance chief, Susan Li, told analysts in July that India has become the company’s “largest market for Meta AI usage.” That usage has coincided with promising signs of retention and engagement on WhatsApp, Li added. 

Among those users is Sonny Ravan, a music producer in Pune, India. Ravan said he finds Meta AI, which he uses through WhatsApp, helpful for learning about the history of songs that he enjoys. He also uses it as a tool to learn about people in the music industry who he plans to work with or meet, describing it as great for preparation.

Sathish Thiyagarajan, 30, a technical support engineer for marketing tech firm GoX.AI, said he’s increasingly using Meta AI as a search tool via WhatsApp, which he noted dominates the Indian market for mobile internet communications.

“While I’m talking with my family or my friends, if they’re saying something to me and I have to search something, I’m not going to go to Google,” said Thiyagarajan, of Chennai, India. “I’m just going to put the phone in the speaker mode, and I’ll immediately search through Meta AI.”

However, Thiyagarajan said he only uses Meta AI when he’s on his phone. If he’s at his computer, OpenAI’s ChatGPT is his preferred AI chatbot.

Not everyone is a fan of Meta AI’s bundling into WhatsApp’s search functions. 

Jawhar Sircar, 72, a retired government official in Kolkata, India, called the Meta AI search feature in WhatsApp “quite a nuisance.” That’s because whenever a user pauses while typing out a name in the search-find box, the Meta AI technology quickly “picks up whatever has been typed” and generates what he describes as unnecessary search prompt suggestions.

As far as the popularity of Meta AI in India, Sircar said he thinks the feature is mostly used by companies, technologists and other professionals who “are getting hooked to AI” alongside the Indian government’s continued investment in regional computing infrastructure.

“Professionals and companies have started using AI, but the general user has no need, at least not on the Meta platforms,” Sircar wrote in an email.

Meta said in December that over 1 million advertisers had used the company’s GenAI tools to create more than 15 million ads in a single month.

“We estimate that businesses using image generation are seeing a +7% increase in conversions,” Meta said at the time, regarding its image generation features.

While people may associate generative AI with the visually striking and sometimes surreal imagery derived from popular services like Dall-E or Midjourney, it’s more likely that the average small business advertiser uses Meta’s GenAI tools for more subtle tasks, said Stacy Reed, an online advertising and Facebook ads consultant. 

That includes using AI to create multiple versions of an ad’s headline, auto-resizing the size of ads so they look appropriate within users’ Instagram and Facebook apps, and repositioning certain images within the ads so that the promotions perform better, Reed said.

Advertisers that already write strong, creative copy can ask Meta’s GenAI tools for “a little bit more” help, Reed said. 

“That’s where you win with their AI tools,” she said.

Reed said the many small advertisers she supports aren’t associating the new features with AI. They “think that Meta is just enhancing the way you build ads,” Reed said.

Celina Guerrero, an independent corporate sales and training consultant, said she uses Meta’s GenAI tools to help with writing headlines for her ads, but she said she finds Meta’s advertising interface to be confusing and constantly changing.

“It is visually overwhelming from a user experience,” Guerrero said.

Ahead of a Facebook ad campaign planned for January, Guerrero said she is debating how to use Meta’s GenAI tools for more in-depth tasks, like modifying her ad’s entire in-line copy. 

“I don’t want my copy to sound like ChatGPT,” Guerrero said, referring to the sterile, run-of-the-mill AI-generated text that’s proliferating the web. “I have two options: One, I don’t use the variations, or two, I spend an inordinate amount of hours editing it.”

Most big companies and advertising agencies are turning to more marketing-specific tools for their generative AI-based ad campaigns, said Jay Pattisall, principal analyst at Forrester. Those services are more robust than Meta’s built-in AI ad tools, he said.

Still, the mere introduction of simple GenAI tools is beneficial to Meta considering it dominates the digital ads market along with Google. Meta’s generative AI tools just have “to be good enough to squeeze out more investment” from advertisers, said Maurice Rahmey, CEO of performance marketing firm Disruptive Digital and a former Facebook customer manager.

“It’s better for their business, even if it’s just those small, incremental changes,” Rahmey said “It’s a business of scale.”

With Meta’s hiring of Shih from Salesforce in November, some analysts say Meta could make an enterprise technology push with its Llama family of open-source AI models. 

Llama’s advancements “represent a significant opportunity for businesses to drive more efficiencies and significantly improve the experiences they offer their customers,” Meta monetization head John Hegeman said in a statement.

Shih, who was one of CNBC’s 2024 changemakers, rejoined Salesforce in 2020 after previously working at the company from 2006 through 2009. As part of her most recent role at Salesforce, Shih helped oversee Einstein GPT for Service and Sales, a GenAI product intended for sales and customer support staff.

During her first stint at Salesforce, Shih created a business app that let users connect their Salesforce customer relationship software with their Facebook connections. In 2009, she wrote “The Facebook Era,” a book intended for professionals to better understand how to use social networks for business.

Multiple former Meta AI and product leaders told CNBC that Shih’s vast experience will be helpful considering the company has failed in previous attempts at building enterprise software. 

Meta announced in May that it plans to shut down Workplace, its business communications product, by 2026. And after buying enterprise startup Kustomer for about $1 billion in 2020, Meta spun it out in 2023 in a deal that was reportedly valued at $250 million.

The most logical step for Meta would be to create a larger business around WhatsApp, said Ralph Schackart, an internet equity analyst at investment bank William Blair. Specifically, WhatsApp could help businesses build customer-service chatbots using Meta’s GenAI, Schackart said. 

“Longer term, this is going to evolve into customized sales agents, which is a $3 trillion-plus industry,” Schackart said about Meta’s WhatsApp business AI chatbot opportunity.

WATCH: If the U.S. bans TikTok, China could retaliate and target Tesla and Apple.