Generative AI Funding: A Quarter-by-Quarter Analysis

All funding deals in the generative AI market over the last years

Below is the table listing all the deals. You can find our methodology at the end of this page.

If you want a deeper understanding of the market and its current dynamics, get our report covering the Generative Ai Market.

Is generative AI funding really getting stronger right now?

Yes, generative AI funding is stronger today than it was in mid-2025, although the huge headline totals make the improvement look much more dramatic than it really is.

Total funding jumps from $3.82B in Q2 2025 to $30.57B in Q3, falls to $8.88B in Q4, surges to about $161.18B in Q1 2026, then drops to an audited $73.67B in Q2. Looking at that chart alone, you would think funding conditions were changing completely every three months.

The rest of the market is much steadier. Capital from rounds of $50M or less stays between $1.13B and $1.50B in every quarter we studied. At the same time, the median round rises from $10M in Q2 and Q3 2025 to $14M in Q4 and roughly $20M in both 2026 quarters.

There is real improvement underneath the giant rounds. Startups are generally raising more money, while OpenAI, Anthropic and a handful of other companies are making the total funding figures swing by tens of billions of dollars.

Quarter	Deals	Total capital	Capital excluding >$50M	Median round
Q2 2025	121	$3.82B	$1.38B	$10.0M
Q3 2025	137	$30.57B	$1.34B	$10.0M
Q4 2025	97	$8.88B	$1.13B	$14.0M
Q1 2026	142	≈$161.18B	$1.44B	$20.0M
Q2 2026	115	$73.67B	$1.50B	$19.3M

Was Q3 2025 really a generative AI funding boom?

Q3 2025 looked like a huge generative AI boom in dollars, but most startups barely felt an improvement in funding conditions.

Capital went from $3.82B to $30.57B in one quarter, an eightfold increase. Deal count moved much less, from 121 to 137, while the median round stayed exactly at $10M. Funding from deals below the $50M mark even slipped slightly, from $1.38B to $1.34B.

A few very large rounds explain most of the gap. Anthropic raised $13B, OpenAI raised an $8.3B tranche, Mistral AI announced a €1.7B Series C worth roughly $2B at the time, and Thinking Machines Lab raised another $2B.

Anthropic alone gives a good sense of what was happening. When the company announced its $13B Series F, it said run-rate revenue had grown from roughly $1B at the start of 2025 to more than $5B by August, while the number of customers spending more than $100,000 a year had increased nearly sevenfold.

That kind of commercial growth can justify enormous amounts of capital at the frontier. It says far less about what a normal application startup could raise during the same quarter.

Q3 was the point where frontier-model funding really started pulling away from the rest of the generative AI market.

Why did generative AI funding cool in Q4 2025 while round sizes got bigger?

Generative AI funding became much narrower in Q4 2025, yet investors were still willing to write bigger checks to the companies they backed.

Deal count fell from 137 to 97, the sharpest quarterly contraction in our period. Total funding also dropped from $30.57B to $8.88B as the giant Anthropic and OpenAI rounds from Q3 stopped inflating the comparison.

The median round moved the other way, climbing from $10M to $14M. Rounds below $5M also fell from 30.7% of transactions to 21.6%.

Several large financings kept the upper end of the market busy. Cursor raised $2.3B after reaching more than $1B in annualized revenue. Reflection AI announced a $2B round around its plan to build frontier open models, while Luma AI raised $900M and paired the financing with plans for a 2-gigawatt AI compute cluster.

There were fewer funded companies, but companies with strong usage, revenue or serious infrastructure ambitions could still raise very large amounts.

What really changed in generative AI funding in Q1 2026?

Q1 2026 was the first quarter in our period where the generative AI market improved broadly at the same time as frontier funding went off the charts.

Deals rose from 97 to 142, a 46% increase. The median financing went from $14M to $20M, and funding from rounds of $50M or less climbed 28% to $1.44B. Those changes affected far more companies than one or two giant deals.

Then OpenAI made the quarter extraordinary.

OpenAI initially announced $110B in new investment from Amazon, NVIDIA and SoftBank while the financing was still open. On March 31, the company said the round had closed with $122B in committed capital. We use that later company disclosure, which takes our Q1 funding estimate to roughly $161.18B.

OpenAI also gave a clear reason for needing so much capital. The company described compute as a strategic advantage and said its infrastructure now spans cloud providers including Microsoft, Oracle, AWS, CoreWeave and Google Cloud, alongside NVIDIA, AMD, AWS Trainium, Cerebras and its own planned chip.

The quarter combined two developments that had previously been easier to separate: a healthier funding market across normal-sized rounds and an unprecedented infrastructure buildout at the very top.

Did generative AI funding collapse in Q2 2026?

No. Generative AI funding cooled in Q2 2026, but the underlying market held up far better than the 54% fall in total dollars suggests.

Deal count did weaken from 142 to 115, so Q2 was clearly less active than Q1. The median round stayed close to its recent high at $19.3M, though, and early-stage deals recovered from 66.9% to 73.9% of transactions.

The smaller end of the funding market was also remarkably stable. Capital excluding deals above $50M actually reached $1.50B, its highest level across the five quarters.

Most of the apparent crash comes from comparing two extraordinary frontier rounds. Q1 now includes OpenAI's $122B close. Q2 contains Anthropic's $65B Series H.

Anthropic requires one extra adjustment. The company said the $65B Series H included $15B of previously committed investments from Amazon and Google. Counting those commitments separately as well would record the same capital twice. After removing that overlap, Q2 comes to $73.67B rather than $88.67B.

The quarter was softer by number of deals, while financing conditions for the broader market remained fairly solid.

How much are OpenAI and Anthropic distorting generative AI funding now?

OpenAI and Anthropic currently distort generative AI funding totals so much that quarterly capital can be a poor measure of what is happening to the average startup.

In Q2 2025, the largest deal represented 23.6% of all funding. OpenAI's updated $122B financing alone accounts for roughly 76% of Q1 2026 capital. Anthropic's $65B round represents about 88% of audited Q2 funding.

This concentration goes beyond the single largest transaction. The top three deals account for roughly 95% of Q1 capital and 91% in Q2. Deals above $50M absorb about 99% and 98% of the money in those two quarters.

Still, the upper end is getting broader. We counted six $100M-plus rounds in Q2 2025, 14 in Q3, 12 in Q4, 22 in Q1 2026 and 17 in Q2.

OpenAI and Anthropic have an almost absurd impact on the dollar totals, while the number of other companies capable of raising nine-figure rounds is also much higher than it was a year earlier.

Are generative AI investors writing bigger checks across the whole market?

Yes, investors are writing bigger generative AI checks across several funding bands, rather than only pouring more money into the two biggest labs.

The clearest change is at the bottom. Deals below $5M represented 30.6% of financings in Q2 2025 and 30.7% in Q3. By Q2 2026, they were down to just 10.4%.

The middle remains very active. In the latest quarter, 40% of deals were between $5M and $20M, while another 20.9% were between $20M and $50M. Roughly six out of ten financings therefore landed between $5M and $50M.

Large rounds have become much more common as well. The $100M-plus bucket went from 5.8% of deals in Q2 2025 to 16.5% in Q2 2026.

Combined with the move in the median from $10M to around $20M, the picture is pretty clear: the typical funded generative AI company is raising more money these days.

Deal-size bucket	Q2 2025	Q3 2025	Q4 2025	Q1 2026	Q2 2026
Under $5M	30.6%	30.7%	21.6%	18.3%	10.4%
$5M–<$20M	38.0%	32.1%	36.1%	30.3%	40.0%
$20M–<$50M	19.8%	16.8%	17.5%	20.4%	20.9%
$50M–<$100M	5.8%	6.6%	10.3%	12.0%	12.2%
$100M+	5.8%	13.9%	14.4%	18.3%	16.5%
Median round	$10.0M	$10.0M	$14.0M	$20.0M	$19.3M

Are new generative AI startups still getting funded, or is all the money going to repeat winners?

New generative AI startups are still getting funded at a healthy rate, although most of the actual dollars now go to companies that have already raised before.

First financings account for 43.8% of deals in Q2 2025, 47.4% in Q3, 40.2% in Q4, 39.4% in Q1 2026 and 42.6% in Q2. That range is surprisingly stable given everything else happening in the market.

The newer companies are also starting with more money. Median first financing rises from $4M in Q2 and Q3 2025 to about $10M by Q2 2026.

Follow-on companies dominate capital because the gap in round size is enormous. In Q1 2026, after incorporating OpenAI's final close, roughly $160B of the quarter's capital went to companies that had already raised before.

So the pipeline of new startups is alive. The financial split appears later: once a company proves it can scale, the amount of capital available can jump by one or two orders of magnitude.

Are generative AI investors moving toward later-stage and older companies?

Generative AI investors briefly moved much harder toward later-stage companies in Q1 2026, but early-stage funding remains the core of the market by deal count.

Seed and Series A together represented around 77% to 79% of deals throughout the three 2025 quarters we studied. That share dropped to 66.9% in Q1 2026, while later-stage deals jumped to 28.2%.

The shift did not continue in Q2. Early-stage deals recovered to 73.9%, while later-stage transactions fell back to 19.1%.

The checks at Seed and Series A are getting bigger at the same time. The Seed median rises from roughly $4.5M in Q2 2025 to $7M in Q2 2026, while the Series A median moves from $17.3M to $27M.

Company-age data point in a similar direction, although we trust that evidence less. Among companies where founding dates are available, the median age moves from two years in Q2 2025 to 2.5 years in Q2 2026, and follow-on companies appear to be drifting closer to three years old.

Coverage of founding dates falls below 50% in the latest quarter, so we would not make a big claim from age alone. The stage data are much cleaner: investors still fund plenty of young generative AI companies, and they are paying more to do it.

Which generative AI categories are actually gaining funding momentum?

AI governance has the clearest momentum by number of funded companies, foundation models dominate the money, and developer platforms are moving toward fewer but much larger rounds.

Enterprise applications still make up about half of generative AI deals today, so they remain the biggest category by breadth. Their median financing has also risen from $10M in Q2 2025 to $16M in Q2 2026.

Governance is growing much faster from a smaller base. Its deal share moves from 5.0% to 13.0% over the period, while quarterly financings increase from six to 15.

Foundation Model APIs sit at the other extreme. They represent 8.7% of audited Q2 2026 deals and 91.9% of the quarter's capital.

Developer platforms have become especially interesting. Their share of deals falls sharply in 2026, while their median round reaches $41M.

Creative GenAI is harder to read. Deal share moves up and down across the five quarters, and its median financing spikes in Q1 before dropping again in Q2. There is not enough consistency there to call it a durable funding shift.

Category	Q2 2025 deal share	Q2 2026 deal share	What is happening now
Enterprise GenAI Applications	62.0%	50.4%	Still the broadest category
GenAI Developer Platforms	19.8%	15.7%	Fewer deals, much larger rounds
Foundation Model APIs	2.5%	8.7%	Huge capital concentration
GenAI Governance Tools	5.0%	13.0%	Strongest consistent breadth growth
Creative GenAI Tools	10.7%	6.1%	Volatile
GenAI Services	Minimal	6.1%	Growing from a very small base

Are generative AI developer platforms actually losing momentum?

Generative AI developer platforms are getting more selective, with fewer companies funded and far more money going into the ones that make it through.

Developer platforms reached around 30% of deals in Q3 and Q4 2025. Their share then slipped to 24.6% in Q1 2026 and 15.7% in Q2.

Round sizes went almost straight up: the median progresses from $6.5M to $7.1M, $11M, $15.8M and eventually $41M.

Cursor shows what this can look like at the top end. When the company announced its $2.3B Series D, it said annualized revenue had crossed $1B, up from more than $100M when it announced its Series B earlier in 2025.

A category can produce fewer financings while its strongest companies become much more valuable and capital hungry. For developer platforms, that explanation fits the data better than a simple loss of investor interest.

Why is AI governance funding picking up now?

AI governance funding is picking up because companies now have real compliance and deployment problems to solve, and regulation has made some of those problems much more urgent.

The funding pattern is unusually consistent. Governance grows from 5.0% of deals in Q2 2025 to 6.6%, 7.2%, 7.7% and finally 13.0% in Q2 2026. Its median round also rises from $4.45M to $12M.

Europe gives us a useful timing clue. General-purpose AI obligations under the EU AI Act started applying in August 2025. Providers have requirements around technical documentation, copyright policy and training-content transparency, while the most advanced models face extra risk, incident-reporting and cybersecurity obligations.

The regulatory pressure is more concrete today. The European Commission's enforcement powers for general-purpose AI rules started applying in August 2026, alongside additional AI transparency requirements.

That sequence overlaps closely with the steady rise we see in governance deals. Regulation does not explain every financing, since companies also need model evaluation, observability, access controls and internal AI policies for ordinary commercial reasons. But it helps explain why governance has moved from a future concern to something enterprises are buying now.

Have foundation-model companies become their own capital market?

Foundation-model companies now operate on a completely different capital scale from most of generative AI.

Foundation Model APIs went from 0.9% of capital in Q2 2025 to 85.6% in Q3, 31.6% in Q4 and roughly 95% in Q1 2026 after updating OpenAI's financing. The category then takes 91.9% of audited Q2 capital.

Those numbers become easier to understand when we look at what the companies are actually buying.

Anthropic said its $65B Series H would help expand compute capacity as Claude demand grew, and reported more than $47B in run-rate revenue by May. Amazon separately said Anthropic had committed to spend more than $100B on AWS technologies over ten years and secure up to 5 gigawatts of Trainium capacity.

OpenAI describes the same basic constraint. Its latest financing is tied to a broad infrastructure strategy across several cloud and chip suppliers, while the company says durable access to compute is becoming one of its main strategic advantages.

Luma's $900M Series C came with plans for a 2-gigawatt compute cluster. Thinking Machines Lab has since announced a separate NVIDIA partnership targeting at least one gigawatt of Vera Rubin systems.

Once AI labs start discussing infrastructure in gigawatts, normal venture-capital comparisons stop being very useful. These companies increasingly resemble capital-intensive infrastructure businesses layered on top of software and research labs.

Is generative AI funding becoming even more concentrated in North America?

Generative AI deal activity remains international, but the largest checks are increasingly concentrated in North America.

North America accounts for roughly 64% to 72% of transactions across the period. Its share by number of deals therefore changes much less than its share of dollars.

Capital tells a more extreme story. North America received 84.7% of funding in Q2 2025, 90.9% in Q3, 86.1% in Q4 and an even larger share in 2026 as OpenAI and Anthropic dominate the totals.

The typical North American financing has also moved up. Its median rises from $12M in Q2 2025 to $25M in both 2026 quarters.

Europe keeps roughly 18% to 22% of deals throughout the period and reaches a $13.4M median round in Q2 2026. Asia-Pacific remains around 8% of transactions in most quarters, with a $14M median in the latest one.

Europe is therefore still producing plenty of funded generative AI companies. Mistral's €1.7B Series C, led by ASML, also shows that Europe can produce very large strategic financings. The biggest difference today is that North America contains several companies with capital needs measured in tens of billions of dollars.

Who keeps funding generative AI: the same VCs or more strategic investors?

The familiar venture firms still show up constantly in generative AI, while the faster change is the rise of strategic investors that can provide chips, cloud capacity and distribution alongside capital.

Andreessen Horowitz appears as a lead or co-lead 24 times across the period we studied. Accel appears 17 times, Sequoia and Insight Partners 15 times each, followed closely by General Catalyst, Lightspeed, Kleiner Perkins, Index Ventures and Khosla Ventures.

The broader investor pool remains healthy. We count roughly 86 to 124 unique disclosed lead investors per quarter, depending on deal volume. There is a recurring inner circle, but plenty of other firms still lead rounds.

Corporate involvement has moved much faster. Strategic-backed deals increase from 14.9% of transactions in Q2 2025 to 27.8% in Q2 2026. Median strategic-backed rounds rise from $13.5M initially to roughly $40M–$55M in every later quarter.

The names explain why. ASML led Mistral AI's €1.7B Series C with a €1.3B investment and a commercial partnership. Amazon's Anthropic relationship combines equity investment with an AWS commitment worth more than $100B over ten years. OpenAI's major investors include Amazon and NVIDIA, both of which also sit directly inside its compute strategy.

These days, the biggest strategic investors can be shareholders, suppliers and customers at the same time. That is one of the clearest changes in how frontier generative AI gets financed.

Investor	Lead/co-lead appearances in our dataset	Broader role seen in the market
Andreessen Horowitz	24	Recurring financial investor
Accel	17	Recurring financial investor
Sequoia Capital	15	Recurring financial investor
Insight Partners	15	Recurring financial investor
General Catalyst	14	Recurring financial investor
Lightspeed Venture Partners	14	Recurring financial investor
Kleiner Perkins	13	Recurring financial investor
Index Ventures	13	Recurring financial investor
Khosla Ventures	12	Recurring financial investor
Amazon / AWS	Strategic	Capital + cloud + compute
NVIDIA	Strategic	Capital + GPU infrastructure
ASML	Strategic	Capital + industrial partnership

What is the generative AI funding trend actually telling us?

Generative AI funding is genuinely stronger than it was at the start of our period, but record fundraising totals are a terrible way to measure that improvement. The cleaner evidence comes from larger median rounds, the sharp decline in sub-$5M deals and the continued strength of the $5M–$50M middle.

Funding has also become much more concentrated at the top. One OpenAI or Anthropic transaction can now change the apparent direction of the entire market. Anyone reading quarterly capital figures without separating those rounds is going to get a distorted picture.

The startup pipeline remains active. Roughly four in ten deals are still first financings, and Seed plus Series A continue to represent most transactions. Investors are clearly willing to fund new companies; they simply reserve vastly larger pools of capital for companies that have already proved something.

Category momentum is splitting in interesting ways. AI governance has the clearest growth in deal breadth, developer platforms are becoming more selective and capital intensive, and foundation-model developers increasingly operate on their own financial scale.

Strategic capital is also becoming more important. The combination of investment, chips, cloud capacity and commercial partnerships makes companies such as Amazon, NVIDIA and ASML much more central to funding today than a simple investor leaderboard would suggest.

Across these five quarters, the ordinary generative AI funding market improved gradually. The frontier layer did something much more extreme: it turned into a capital-intensive infrastructure race capable of making every quarterly headline look far more dramatic than the rest of the market really is.