Artificial intelligence now draws close to half of all venture capital in the world. Private AI companies raised about $226 billion in 2025, according to CB Insights.
In the first quarter of 2026, they raised another $226 billion. Funding then slowed to $153.1 billion in the second quarter and $103.3 billion in the third.
This article looks at the latest AI startup funding statistics. It covers company counts, mega-rounds, valuations, country totals and failure rates.
AI Startup Funding Statistics at a Glance
Here are the main figures from CB Insights, Crunchbase and Stanford. The totals differ because each group counts deals in its own way.
| Metric | Figure | Source |
|---|---|---|
| AI venture funding, 2025 | $226 billion | CB Insights |
| AI share of global venture funding, 2025 | 48% | CB Insights |
| Global private AI investment, 2025 | $344.7 billion | Stanford |
| Total corporate AI investment, 2025 | $581.7 billion | Stanford |
| AI funding, Q1 2026 | $226 billion | CB Insights |
| AI funding, Q3 2026 | $103.3 billion | CB Insights |
| Mega-round share of AI funding, 2025 | 79% | CB Insights |
| Mega-round share of AI funding, Q1 2026 | 94% | CB Insights |
| New AI unicorns, Q3 2026 | 42 | CB Insights |
| Newly funded US AI companies, 2025 | 1,953 | Stanford |
Crunchbase counts $211 billion in AI venture funding for 2025. That is up about 85% from $114 billion in 2024.
The trend matters more than the exact number. Every major tracker shows a record year.
How Many AI Startups Are There?
There is no official global register of AI companies. Databases sort firms in different ways, depending on whether AI is the main product or only one feature.
CB Insights tracks more than 40,000 private AI companies worldwide. That does not mean there are 40,000 new generative AI startups.
The group includes firms in foundation models, robotics, self-driving cars, healthcare, cybersecurity and semiconductors. It also covers data infrastructure, finance, defense, enterprise software, computer vision and voice AI.
Stanford’s 2026 AI Index gives a narrower count. The US had 1,953 newly funded AI companies in 2025, more than 10 times the number in the next-largest country.
The number of newly funded AI companies worldwide rose about 71% in 2025. Investors are still backing new entrants, even as the biggest rounds go to established leaders.
| Measure | Figure |
|---|---|
| Private AI companies tracked by CB Insights | More than 40,000 |
| Newly funded US AI companies, 2025 | 1,953 |
| US lead over the next-largest country | More than 10 times |
| Growth in newly funded AI companies worldwide, 2025 | About 71% |
How Much Funding Is Going Into AI?
The answer depends on the method. CB Insights recorded about $225.8 billion for private AI companies in 2025, nearly double the year before.
Crunchbase put the figure at $211 billion. Stanford uses a wider method and counts $344.66 billion in global private AI investment.
Stanford also tracks total corporate AI investment, which includes mergers and acquisitions. That figure was about $581.69 billion in 2025.
The gaps are not always errors. Databases treat minority stakes, growth capital, private equity, strategic deals, secondary sales, infrastructure and acquisitions in different ways.
By every major measure, 2025 was a record year for AI funding.
| Source | What it measures | 2025 figure |
|---|---|---|
| Crunchbase | AI-related venture funding | $211 billion |
| CB Insights | Funding to private AI companies | $225.8 billion |
| Stanford | Global private AI investment | $344.66 billion |
| Stanford | Total corporate AI investment, including M&A | $581.69 billion |
AI Took Nearly Half of Global Venture Capital
CB Insights estimates that venture funding across all industries reached about $469 billion in 2025. AI companies took about $226 billion of that, or roughly 48%.
Crunchbase reaches a similar view. It estimates that about half of global venture funding went to AI-related businesses.
AI is no longer one sector among many, like biotech, fintech and software as a service. It has become the main driver of venture funding.
| Category | 2025 funding | Share |
|---|---|---|
| AI companies | $226 billion | 48% |
| All other sectors (calculated) | $243 billion | 52% |
| Total global venture funding | $469 billion | 100% |
Mega-Rounds Now Dominate AI Financing
A mega-round is a private funding round of at least $100 million. In 2025, these deals made up about 79% of all AI funding.
The share grew in 2026. In the first quarter, 94% of AI funding came from rounds of $100 million or more.
The average deal size in Q1 2026 reached about $160 million. That compares with $38 million for the full year of 2025.
This does not mean a typical startup raised $160 million. A few huge deals pushed the average up.
In Q3 2026, mega-rounds still made up 85% of AI funding. They came from 154 deals.
| Period | Mega-round share of AI funding |
|---|---|
| Full year 2025 | 79% |
| Q1 2026 | 94% |
| Q3 2026 | 85% |
The Biggest AI Funding Rounds
CB Insights reports that three companies raised the largest rounds in Q1 2026. OpenAI raised $122 billion, Anthropic raised $30 billion and xAI raised $7.5 billion.
Together, those three rounds made up about 71% of all AI funding in the quarter.
The trend was clear in 2025 too. Crunchbase found that 15 companies raised single rounds of at least $2 billion, worth more than $100 billion in total.
OpenAI led that year with a $40 billion round led by SoftBank. CB Insights also lists large rounds for Anthropic, Scale AI, xAI, Databricks and Aligned Data Centers.
The six largest AI rounds of 2025 added up to about $111 billion. That was 49% of all AI funding that year.
| Company | Q1 2026 round | Reported valuation |
|---|---|---|
| OpenAI | $122 billion | About $840 billion |
| Anthropic | $30 billion | About $380 billion |
| xAI (Series E) | $7.5 billion | About $230 billion |
AI Funding by Quarter in 2026
AI funding reached $226 billion in Q1 2026, equal to the whole of 2025 on CB Insights’ count. The total was lifted by the rounds for OpenAI, Anthropic and xAI.
Funding fell to $153.1 billion in Q2. It then dropped to $103.3 billion in Q3, down 33% from Q2.
A quarter with $103 billion in funding is still large by past venture standards. The money did not disappear, but investors became more selective.
CB Insights also reported that deal counts fell even as total AI funding rose. Fewer companies are getting larger checks.
| Quarter | AI funding | Change from prior quarter |
|---|---|---|
| Q1 2026 | $226 billion | Not given |
| Q2 2026 | $153.1 billion | Down about 32% (calculated) |
| Q3 2026 | $103.3 billion | Down 33% |
Jumbo Series A Rounds and New Unicorns
Mega-rounds are no longer limited to mature companies. By September 2026, Crunchbase counted at least 114 Series A rounds worth $100 million or more.
Those rounds raised about $33 billion in total. More than 70% went to AI-focused startups, and at least 12 were above $500 million.
Series A rounds used to be measured in single-digit or low double-digit millions. In frontier AI, robotics, defense and chips, investors now put in hundreds of millions much earlier.
A unicorn is a startup valued at $1 billion. CB Insights says 42 AI companies became unicorns in Q3 2026, the most in about 18 months.
The trend reaches higher valuations too. By early October 2026, 42 private companies across the venture market had become decacorns, valued above $10 billion. That puts the year on course for a record.
| Measure | Figure |
|---|---|
| Series A rounds of $100 million or more (by Sept 2026) | At least 114 |
| Total raised in those rounds | About $33 billion |
| Share that went to AI-focused startups | More than 70% |
| Rounds above $500 million | At least 12 |
| New AI unicorns in Q3 2026 | 42 |
AI Startup Valuations
Large rounds have pushed private valuations to levels once seen only at big public companies. By 2026, several private AI firms were valued above $100 billion.
OpenAI’s round came with a valuation of about $840 billion. Anthropic was valued at about $380 billion and xAI at about $230 billion.
Young companies are also getting high prices. Reuters reported in October 2026 that chip startup Nuvacore was seeking funding at a valuation of about $2.5 billion.
Nuvacore is only six months old and has no commercial product yet. Teams with respected researchers or chip engineers can attract such valuations before they earn revenue.
High valuations reflect a belief that AI markets may favor a few winners. But a high valuation is not the same as profit.
| Company | Valuation | Note |
|---|---|---|
| OpenAI | About $840 billion | Linked to its Q1 2026 round |
| Anthropic | About $380 billion | Linked to its Q1 2026 round |
| xAI | About $230 billion | Implied by its Series E |
| Nuvacore | About $2.5 billion | Seeking funds, six months old |
Why AI Companies Need So Much Money
Older software startups could rent cloud computing as they grew. Frontier AI companies face much larger bills.
Their costs include chips, data centers, electricity, networking gear, model training, inference, data and research talent. Chip developers face heavy design costs as well.
Stanford estimates that OpenAI spent about $16.3 billion on compute in 2025. Anthropic spent about $6.8 billion.
OpenAI has reportedly projected hundreds of billions of dollars in future compute needs. That is why AI rounds now look like infrastructure financing.
Stanford’s data shows where private money went. The top category in 2025 was AI infrastructure, models, research and governance, at about $143.2 billion.
Other funded areas included data management, healthcare AI, cloud, cybersecurity, AI agents, robotics, self-driving cars, fintech, defense, semiconductors and legal technology.
| Item | 2025 figure |
|---|---|
| OpenAI compute spending | $16.3 billion |
| Anthropic compute spending | $6.8 billion |
| Private investment in AI infrastructure, models, research and governance | $143.2 billion |
The United States Leads AI Funding
No country is close to the US in private AI investment. Stanford estimates that US AI companies received $285.9 billion in 2025.
Europe received $20.9 billion and China received $12.4 billion. US private investment was more than 23 times China’s total.
The US also had 1,953 newly funded AI companies. Many of the largest model developers, chip firms, cloud providers and venture firms are based there.
Private figures probably understate China’s total. The Chinese government also runs large guidance funds.
Stanford estimates that government-linked vehicles in China put about $184 billion into AI firms between 2000 and 2023.
| Region | Private AI investment, 2025 |
|---|---|
| United States | $285.9 billion |
| Europe | $20.9 billion |
| China | $12.4 billion |
How Many AI Startups Fail?
This is where the numbers become less reliable. Claims that 90% of AI startups fail are repeated often online.
There is no authoritative global dataset showing that 90% of AI startups have failed. Many AI companies were formed between 2023 and 2026, so they have not existed long enough to measure a long-term failure rate.
General startup data gives a firmer base. CB Insights studied 431 venture-backed companies that shut down from 2023 onward. They were not all AI firms.
The shares overlap because many companies failed for more than one reason. Together, the 431 companies had raised $17.5 billion.
The median failed company had raised about $11 million. The average was about $48 million.
Large funding does not guarantee survival. Healthcare automation startup Olive raised close to $1 billion and reached a valuation of about $4 billion before it shut down.
| Reason for failure | Share of 431 failed startups |
|---|---|
| Ran out of capital | 70% |
| Poor product-market fit | 43% |
| Bad timing | 29% |
| Unsustainable unit economics | 19% |
Seed-Stage and Project Failure Rates
Seed-stage data offers another benchmark. PitchBook-related figures show that 29.5% of stronger consensus seed rounds eventually went out of business or bankrupt.
For other seed-funded startups, the figure was 39.4%. About 53% of the stronger group reached Series A, compared with about 28% of the others.
These rates may rise as newer groups age. It takes years to see how a cohort ends up.
There is stronger evidence for AI project failure inside companies. S&P Global found that the share of firms abandoning most of their AI projects before production rose from 17% to 42%.
Firms also said about 46% of AI projects were dropped between proof of concept and broad rollout. That measures projects, not startups.
| Measure | Figure |
|---|---|
| Stronger seed rounds that closed or went bankrupt | 29.5% |
| Other seed startups that closed or went bankrupt | 39.4% |
| Stronger seed rounds that reached Series A | About 53% |
| Other seed startups that reached Series A | About 28% |
| Firms abandoning most AI projects before production | 42% (up from 17%) |
| AI projects dropped between proof of concept and rollout | About 46% |
A Two-Tier Market With Real Risks
AI funding is splitting into two markets. At the top are firms that can raise billions, such as frontier model developers, chip startups, data center builders and defense AI companies.
Below them are thousands of smaller companies competing for ordinary seed and Series A money. A few huge deals can make the sector look flooded with cash while smaller firms struggle to raise their next round.
Several risks remain for AI startups:
- Weak product differentiation, because many firms use the same foundation models.
- High inference costs, which can grow as fast as revenue.
- Falling model prices, which squeeze simple products built on top of models.
- Customer concentration, where a few large clients supply most revenue.
- Fast technology change, which can turn a new feature into a standard one.
- Capital needs that reach billions before a company scales.
- Competition for top researchers and engineers.
Big rounds do not remove these risks. They can add to them if a company scales before its economics work.
The Bottom Line
AI is in one of the largest private funding booms in tech history. Private AI companies raised about $226 billion in 2025 by CB Insights’ count, and Stanford’s wider measure gives about $345 billion.
The boom grew in early 2026, with $226 billion raised in Q1 alone. Funding then eased to $103.3 billion in Q3, with mega-rounds still making up 85% of the total.
Valuations rose with funding. Dozens of new unicorns appeared, and OpenAI, Anthropic and xAI are valued in the hundreds of billions.
Failure data needs care. No credible dataset shows that 90% of AI startups have failed, and most are too young to judge.
Older startup data points to real risk. Roughly 30% to 40% of seed companies in measured groups eventually shut down, and running out of cash is the top cause.
The key question is no longer whether investors will fund AI. It is whether enough of these companies can turn large rounds into lasting products and profits.
Sources
- Stanford University – 2026 AI Index Report: Economy
- Stanford University – 2026 AI Index Report
- CB Insights – State of AI 2025
- CB Insights – State of AI Q1 2026
- CB Insights – State of AI Q3 2026
- CB Insights – State of Venture 2025
- Crunchbase – AI Funding in 2025
- Crunchbase – Largest Funding Rounds of 2025
- Crunchbase – Global Venture Funding 2025
- Crunchbase – Jumbo Series A Rounds in 2026
- CB Insights – Why Startups Fail
- S&P Global – AI Adoption and Project Failure Rates
- Reuters – Nuvacore Funding and Valuation
