AI Startup Valuations Are Tripling in Months as Mega-Rounds Flood Healthtech and Deep Tech

Meta description: In 2025, AI startup valuations are doubling and even tripling within months as 49 US AI companies secure $100M‑plus rounds and healthcare AI startups dominate new funding — a boom that is now raising serious bubble concerns.

A funding boom that looks like a new era — and a familiar warning

The final week of November 2025 is ending with another surge of venture capital into artificial intelligence and healthtech.

US startups raised more than $1.5 billion between November 24 and 30, led by a $600 million round for AI robotics software company Physical Intelligence, a $300 million round for prediction‑market platform Kalshi, and a $298 million Series B for longevity‑focused health startup Function Health. AI and healthtech account for the bulk of that capital. ^[1]

These latest deals land on top of a year in which:

AI startup valuations are doubling or tripling within months, as companies raise back‑to‑back mega‑rounds.
At least 49 US AI startups have raised $100 million or more in 2025, matching last year’s record — but with far more repeat mega‑rounds. ^[2]
Healthcare AI has become one of the hottest verticals, with early‑stage investors calling out 18 “most promising” health startups and data showing AI‑driven companies capturing the majority of digital health funding. ^[3]

At the same time, new analysis out this weekend suggests OpenAI’s ecosystem is now tied to nearly $100 billion of debt carried by its partners, while separate estimates from HSBC indicate the company may need to raise more than $200 billion by 2030 to fund its ambitions. ^[4]

Together, the latest reporting from Fortune (via syndication), TechCrunch, Business Insider and a cluster of fresh analyses paints a picture of an AI funding climate that is hotter — and riskier — than any previous tech cycle.

Valuations that double and triple in months

The core story of 2025’s AI boom is speed: valuations that jump by tens or even hundreds of billions of dollars in a matter of months.

Syndicated reporting based on Fortune’s analysis, summarized this week by IndexBox and others, highlights a series of extraordinary trajectories: ^[5]

Anthropic raised a $3.5 billion Series E in March at a valuation of about $61.5 billion, then a $13 billion Series F just six months later that lifted its valuation to roughly $183 billion — effectively tripling in half a year.
OpenAI began 2025 valued near $157 billion after an October 2024 funding round, climbed to around $300 billion in a March 2025 deal, and then hit a reported $500 billion valuation in a recent tender offer. On average, its private valuation increased by nearly $29 billion per month between October 2024 and October 2025.
Recruiting and “AI food chain” startup Mercor reportedly moved from a $2 billion valuation after a $100 million Series B in February to approximately $10 billion following a larger round in October.

A growing list of companies have raised two or more rounds in 2025 alone with sharply higher valuations each time, including Cursor (Anysphere), Reflection AI, OpenEvidence, Lila Sciences, Harmonic, Fal, Abridge and Doppel, while others such as Harvey and Databricks are reported to be on their third funding cycles of the year. ^[6]

These patterns are reminiscent of the 2021 zero‑interest‑rate era — but with far larger numbers and a far more mature AI market.

Cursor’s 5‑month leap: from $9.9B to $29.3B

Few examples capture the new math of AI valuations as vividly as Cursor, the AI coding assistant developed by Anysphere.

In June 2025, Anysphere raised $900 million at a valuation of about $9.9 billion, in a round that already tripled its previous valuation from January. ^[7]
Just five months later, in November, Cursor announced a $2.3 billion Series D led by Coatue and Accel, with Nvidia and Google joining the cap table. That round pushed its valuation to roughly $29.3 billion, nearly tripling again. ^[8]

Reuters reports that Cursor has crossed $1 billion in annualized revenue and grown sales‑led revenue 100‑fold since the start of the year. Global venture funding in Q3 2025 climbed 38% year‑over‑year to $97 billion, with about half of that capital going to AI companies. ^[9]

That context helps explain why investors are willing to price high‑growth AI platforms so aggressively — but it also underlines how dependent the broader funding environment has become on a relatively small cluster of AI winners.

49 US AI startups with $100M+ rounds — and many are coming back for more

TechCrunch’s deep dive into 2025 AI deals, updated and synthesized this week by The Tech Buzz, shows that 49 US AI startups have already raised rounds of $100 million or more this year, matching 2024’s record. ^[10]

The new twist is repeat mega‑rounds. More companies have raised multiple nine‑figure rounds in a single calendar year, a sign that investors are racing to lock in positions in what they perceive as category‑defining winners.

Highlights from that dataset include:

OpenAI topping the list with a $40 billion tender‑style transaction at a $300 billion valuation earlier this year, before the more recent secondary deals that reportedly pushed its valuation toward $500 billion. ^[11]
Anysphere/Cursor, with its $900 million June raise and $2.3 billion November follow‑on, leaping from roughly $10 billion to $29.3 billion in valuation. ^[12]
Healthcare AI standouts such as Hippocratic AI, which has now raised roughly $267 million across multiple rounds, and Abridge, which has secured about $550 million in 2025 funding and doubled its valuation to $5.3 billion. ^[13]
Large infrastructure bets, including Cerebras Systems raising roughly $1.1 billion for AI chips and Lambda securing around $480 million to expand cloud infrastructure for training and inference workloads. ^[14]
Legal and professional‑services AI, where Harvey reportedly closed two $300 million rounds this year, taking its valuation from around $3 billion to $5 billion in just four months. ^[15]

Across coding assistants, healthcare AI, legaltech and infrastructure, one pattern repeats: capital is concentrating in companies that can credibly claim to be building new “operating systems” for entire industries.

Healthcare AI emerges as a breakout winner

If 2023 and 2024 were about foundation models and horizontal platforms, 2025 looks like the year of applied healthcare AI.

A new Business Insider feature, published November 24 and based on input from investors at firms like Andreessen Horowitz, NEA, Kleiner Perkins and Oak HC/FT, identifies 18 “most promising” healthcare startups for 2025 — and nearly all of them are using AI at their core. ^[16]

Rock Health data cited in that piece shows why: AI‑driven healthcare startups attracted nearly $4 billion of the $6.4 billion in US digital health funding in the first half of 2025, meaning AI captured well over half of sector capital. ^[17]

The list spans several themes:

1. Revenue cycle, payments and data plumbing

Startups attacking the financial and administrative backbone of healthcare — often the most immediate ROI use cases for AI — feature heavily:

Akasa uses AI to automate hospital revenue‑cycle workflows and already counts major systems like Cleveland Clinic and Johns Hopkins as customers.
Inbox Health, RapidClaims, Stedi and Courier Health are modernizing patient billing, claims management, and pharma‑patient interactions with AI‑driven platforms.
PayZen applies AI underwriting to create interest‑free payment plans for patients, aiming to improve collections for providers while easing the affordability crisis.

These companies are attracting attention because they unlock cash, not just abstract efficiency — a powerful selling point in a margin‑constrained sector.

2. Workforce, hiring and training

A second cluster tackles healthcare’s acute labor shortage:

Carefam runs an AI‑powered hiring marketplace that matches nurses and clinicians to open roles while helping with resumes and interview prep.
Clasp links student loan repayment to future employment in health systems — a sort of ROTC model for medicine — helping hospitals secure talent earlier in the pipeline.
Stepful delivers short, online training programs for entry‑level clinical roles and uses AI tools to coach students and flag those at risk of dropping out.

With systems struggling to staff hospitals and clinics, investors see labor‑focused AI platforms as both socially important and commercially attractive.

3. New care models and diagnostics — from obesity to oncology

Several picks reflect a shift toward continuous, data‑rich care:

Cadence offers remote monitoring and telehealth for chronic conditions and already manages tens of thousands of patients across large health systems.
Knownwell runs clinics and virtual programs for obesity and metabolic disease, pairing primary care with weight‑management tools and recently raising new capital led by CVS Health Ventures.
Diana Health partners with hospitals to redesign women’s care — from maternity to menopause — using hybrid virtual/in‑person models.

Others are commercializing new diagnostic and intervention technologies:

Teal Health won FDA authorization this year for an at‑home cervical cancer screening wand, the first of its kind in the US.
Twin Health builds AI “digital twins” of patients based on wearable, lab and clinical data; a recent study published in NEJM Catalyst found that its programs helped some patients reduce A1C levels without GLP‑1 drugs. ^[18]
Nudge is developing ultrasound‑based “whole‑brain interfaces” for neurological conditions, combining imaging and stimulation in a non‑invasive system.

Finally, Wellsheet and Infinitus reflect a quieter but crucial layer: Wellsheet sits on top of electronic health records to surface the most relevant data for clinicians, while Infinitus deploys voice AI agents that handle high‑volume phone tasks like prior authorizations and benefits checks.

Taken together, these companies show why AI is no longer just an experiment in healthcare — it is embedded in workflows, reimbursement and clinical decision‑making.

This week’s numbers: AI and healthtech dominate the deal sheet

The funding data for November 24–30, 2025 underlines how central AI and healthtech have become to the US startup landscape.

According to a News Strike weekly wrap, US startups raised over $1.5 billion during the period, with the largest checks going to AI‑driven companies: ^[19]

Physical Intelligence (San Francisco) – $600 million Series B
AI software that brings large‑model intelligence to robots in logistics, manufacturing and autonomous systems; the week’s single largest deal.
Kalshi (San Francisco) – $300 million Series D
A regulated prediction‑market platform riding surging interest in event contracts; separate filings and reports indicate it has also closed a subsequent $1 billion round at an $11 billion valuation this month, more than doubling its valuation in a matter of weeks. ^[20]
Function Health (Austin) – $298 million Series B at a $2.5 billion valuation
A longevity‑focused diagnostics platform that offers regular lab testing and an AI‑driven “medical intelligence” layer on top of user data; total funding now stands around $350 million. ^[21]
Smaller but notable AI‑related deals, such as Voio (AI for clinical diagnostics) and Tidalwave (AI mortgage processing), also featured on the list.

In parallel, Hippocratic AI disclosed a $126 million Series C at a $3.5 billion valuation earlier this month, aimed at scaling “clinically safe” generative‑AI agents for healthcare systems. ^[22]

This is on top of the already‑announced 2025 mega‑rounds like Abridge’s $300 million Series E and various follow‑ons for infrastructure players like Cerebras and Lambda.

The message: even in a broader venture market that remains more cautious than in 2021, AI and healthtech are defying gravity.

The growing chorus of bubble warnings

With valuations climbing so quickly, concerns about an AI bubble are now front‑page news.

A new piece from LiveMint on November 30, drawing heavily on Fortune’s reporting, asks directly whether the back‑to‑back funding rounds at companies like OpenAI, Anthropic, Mercor and Cursor are becoming “a dangerous liability.” Andreessen Horowitz general partner Jennifer Li cautions that multiple rounds go wrong “when the focus shifts from building to fundraising before the foundation is set,” and argues they only work when capital directly fuels product‑market fit and execution. ^[23]

On the infrastructure side, analysis from the Financial Times and follow‑up coverage by The Times of India highlight a rapidly expanding debt pile linked to OpenAI’s ecosystem:

Data‑center and financing partners including Oracle, SoftBank, CoreWeave and others have amassed nearly $100 billion in loans and bonds tied to OpenAI‑related capacity and contracts. ^[24]
HSBC estimates OpenAI may need to raise at least $207 billion by 2030 to fund its commitments while remaining loss‑making — more than many sovereigns issue over the same period. ^[25]

That structure — where an AI startup with relatively modest revenue relies on partners to load up on debt to build out compute — is drawing comparisons to leveraged infrastructure booms in other sectors.

Investors are divided:

Some see this as a rational response to a once‑in‑a‑generation platform shift, in which whoever builds the best models and data centers first could enjoy years of super‑normal profits.
Others argue the market is already pricing in perfection — flawless execution, rapid monetization, and no major regulatory shocks — in a field where both technical and political risks remain high.

Either way, the cost of being wrong is increasing, not just for VCs and founders but for banks, pension funds and sovereign wealth funds financing the AI build‑out.

Why this is not just 2021 all over again

Despite the bubble chatter, many industry insiders stress key differences between today’s AI boom and the 2021 venture frenzy:

Real revenue (for some)
Companies such as Cursor and Abridge are reporting nine‑ and even ten‑figure annualized revenue and rapid growth, suggesting that at least a subset of AI startups are already commercially validated. ^[26]
Mass‑market adoption
Consumer and enterprise use of AI assistants, copilots and medical scribes has exploded since ChatGPT’s launch, creating sustained demand rather than purely speculative hype.
Enormous capital and compute requirements
Training frontier models now requires billions of dollars in GPUs, data centers and energy infrastructure. That makes the sector look less like 1999 dot‑coms and more like capital‑intensive heavy industry, with high barriers to entry but also high systemic risk if demand falls short.
Concentration of capital
Unlike prior bubbles where thousands of tiny companies raised modest sums, today’s AI boom channels huge checks into a relatively small number of winners, which may limit some of the froth but also amplifies the consequences if any one of those winners stumbles.

As Bison Ventures co‑founder Tom Biegala told Fortune, the 2021 market often saw startups raise simply because “companies would raise a round… not because they’ve made any sort of real progress.” Now, investors are at least attempting to tie rounds to technical and commercial milestones — though the bar remains subjective. ^[27]

What this means for founders, investors and regulators

For founders

Momentum cuts both ways. If you are fortunate enough to have multiple term sheets, more capital can help you capture market share — but it also raises expectations and compresses the time you have to turn burn into durable moats.
Use capital to deepen product‑market fit, not just headcount. As Jennifer Li notes, multi‑round fundraising only “goes right” when the money directly fuels better products and execution, not vanity metrics. ^[28]
Underwrite your own compute future. Locking in GPU and data‑center access is critical, but copying OpenAI’s capital structure is not — smaller AI companies may need to be more conservative to avoid being crushed by fixed costs.

For investors

Risk is shifting from equity to credit. As the OpenAI ecosystem shows, much of the risk now sits on the balance sheets of data‑center partners, private credit funds and banks. Equity investors need to understand that leverage stack when evaluating late‑stage deals. ^[29]
Healthcare AI is a durable theme, but crowded. The Business Insider list and recent mega‑rounds in companies like Abridge, Hippocratic AI and Function Health attest to the opportunity — but they also signal intensifying competition for similar customer budgets. ^[30]

For regulators and policymakers

Systemic risk is no longer hypothetical. When a single startup’s ecosystem involves $100 billion or more of associated debt and trillion‑dollar procurement commitments, questions about financial stability and competition policy move from think‑tank panels to the desks of central bankers and antitrust authorities. ^[31]
Healthcare AI needs robust guardrails. Tools that automate billing, staffing and even medical decision‑support can improve access and efficiency — but they also introduce new failure modes, from biased models to over‑reliance on black‑box systems.

The bottom line: breathtaking growth with very little margin for error

As of November 30, 2025, the data is clear:

AI startup valuations really are doubling and tripling within months.
The US now counts at least 49 AI companies with $100M‑plus rounds this year alone, and a growing subset have crossed the billion‑dollar‑round threshold. ^[32]
Healthcare AI and healthtech are central to this story, from the 18 early‑stage startups picked by leading investors to mega‑rounds for companies like Abridge, Hippocratic AI and Function Health. ^[33]

Whether history remembers 2025 as the foundation of a new AI‑powered economy or as the blow‑off top of an epic bubble will depend less on the size of the rounds and more on what comes next: can these companies convert record‑breaking funding into real productivity gains, better health outcomes and sustainable business models?

This AI Medical Assistant Just Hit a $5.3 Billion Valuation in 4 Months