Toolradar Research
Two AI Labs Took 75% of 2026's Software Funding
Software companies raised $272B across 449 rounds in 2026, but OpenAI ($110B) and Anthropic ($95B) took 75% of it. The other 414 companies split a quarter.
Key findings
What the data shows.
- 01
Two AI labs took 75% of the money. OpenAI raised $110 billion and Anthropic raised $95 billion across two rounds, a combined $205 billion. That is 75% of the roughly $272 billion raised by all 416 software companies we tracked in 2026.
- 02
Third place is a rounding error. After the top two, the largest round of the year is Decagon at $4.5 billion, followed by Synthesia at $4 billion and CoreWeave at $3.5 billion. The gap between second and third is larger than the entire rest of the market combined.
- 03
AI dominates by count, not just by dollars. The categories with the most funded companies are AI and Automation (98), AI Agents (71), and AI Assistants (51). The money and the deal flow both point the same direction.
- 04
For everyone outside AI, funding looks normal. Strip the two mega-labs and the rest is ordinary startup machinery: seed and Series A rounds in the millions, not billions. The bubble, if it is one, is concentrated in a handful of frontier-model companies, not spread across software.
About the research
How we built this report.
Toolradar tool database. Editorial review with weekly pricing verification.
2026. Snapshot taken August 4, 2026. Refresh due Nov 4, 2026.
Public scoring rubric. See how we rate for the full criteria.
Creative Commons BY 4.0. Quote, link, and reuse with attribution.
Venture funding is supposed to be a spread bet: many companies, many rounds, a portfolio. In software in 2026, it stopped being one. We tracked 449 funding rounds raised by 416 software companies this year, worth roughly $272 billion in total. Two companies took three-quarters of it. This is not a funding market with a few big winners. It is a funding market with two winners and everyone else.
The dataset
This report covers 449 distinct funding rounds announced in 2026 by 416 software companies in the Toolradar catalog, deduplicated so a single company round is counted once even when the company ships many products. Dollar totals are the sum of disclosed round amounts. We report reliable aggregates: total raised, round counts, top rounds, and the number of funded companies per category. We deliberately do not break dollars down by category, because companies span multiple categories and the sums would double-count. This is first-party structured funding data, released under CC BY 4.0.
Two companies, three-quarters of the money
| Company | 2026 raised | Share of total |
|---|---|---|
| OpenAI | $110B | 40% |
| Anthropic | $95B | 35% |
| Everyone else (414 companies) | $67B | 25% |
Read that bottom row twice. Four hundred and fourteen software companies, every startup and scale-up that raised money in 2026 other than the two frontier labs, split a quarter of the total between them. The concentration is not close to anything a normal venture market produces. It is the financial signature of a platform shift, where the market decides the winners are the model providers and pours capital into them at a scale that dwarfs everything downstream.
The top rounds
| Company | Round | Amount |
|---|---|---|
| OpenAI | Series F | $110B |
| Anthropic | Series H | $65B |
| Anthropic | Series G | $30B |
| Decagon | Series D | $4.5B |
| Synthesia | Series F | $4B |
| CoreWeave | Growth | $3.5B |
| Parloa | Series D | $3B |
| Filevine | Late-stage | $3B |
Look at what the biggest non-mega rounds are: Decagon (AI customer service agents), Synthesia (AI video), CoreWeave (AI infrastructure), Parloa (AI contact centers). Even outside the top two, the largest checks are going to AI applications and the picks-and-shovels underneath them. The one non-AI name near the top, Filevine, a legal case-management platform, stands out precisely because it is the exception.
AI dominates the deal flow too
Dollars are one story, deal count is another, and they agree. Ranking categories by how many companies raised money in 2026, AI leads on both.
| Category | Funded companies (2026) |
|---|---|
| AI and Automation | 98 |
| Analytics | 83 |
| Finance | 72 |
| AI Agents | 71 |
| Workflow Automation | 64 |
| AI Assistants | 51 |
| Security | 42 |
| Cloud and Infrastructure | 33 |
AI and Automation, AI Agents, and AI Assistants are three of the top six categories by funded-company count. Analytics, Finance, and Security still raise steadily, the ordinary business of software financing has not stopped. But the center of gravity, in both dollars and deals, has moved decisively to AI.
What this means
- For founders outside AI: the headline numbers are not your market. The $272 billion is a mirage created by two rounds. The realistic funding environment for a normal software company is seed and Series A checks in the millions, competitive but not transformed. Do not benchmark your raise against OpenAI.
- For buyers evaluating AI vendors: enormous funding is not durability, it is runway and expectation. A tool sitting on billions can afford to underprice and outspend, which is great until the terms change. Weight the product and the review corpus, not the size of the last round.
- For anyone reading the AI bubble debate: the concentration is the argument. When two companies absorb 75% of an entire sector's capital, the risk is not spread across software, it is stacked on a very small number of bets. Whether that is visionary or fragile depends entirely on those two companies delivering.
Methodology
Source: 449 funding rounds announced in 2026 by 416 software companies in the Toolradar catalog, deduplicated to one round per company even when a company ships multiple products. Totals sum disclosed round amounts; company-level shares are exact. We report totals, round counts, top rounds, and funded-company counts per category, and deliberately omit dollar-per-category breakdowns because multi-category companies would double-count. Round amounts are as disclosed and collected from public sources. Released under CC BY 4.0. Reuse it with a link back to this page.
Frequently asked questions
How much did software companies raise in 2026?
Roughly $272 billion across 449 rounds and 416 companies in this dataset. But that number is dominated by two rounds: OpenAI's $110 billion and Anthropic's $95 billion together make up 75% of the total, so the figure describes two companies far more than it describes the software market.
Which companies raised the most in 2026?
OpenAI ($110B) and Anthropic ($95B across two rounds) are far ahead of everyone. The largest round after them is Decagon at $4.5 billion, then Synthesia at $4 billion and CoreWeave at $3.5 billion. The gap between the top two and the rest is enormous.
Is AI taking all the venture funding?
In this data, largely yes at the top and by deal count. Two AI labs took three-quarters of the dollars, and AI categories lead by number of funded companies. Non-AI software still raises money, but the concentration of capital in frontier-model companies is unlike anything in the rest of the market.
Should I trust a tool because it raised a lot of money?
No. Large funding buys runway and time, not reliability or fit. A heavily funded tool can underprice to grab share and later change terms once it needs margin. Evaluate the product, the pricing, and the real review corpus rather than the size of the last round.
Cite this report
Use the data, credit the source.
Released under Creative Commons BY 4.0. You may quote, link, and reuse the data with attribution.
More research
How Software Is Priced in 2026: The Median Tool Costs $18, the Average Costs $178
We read the pricing pages of 5,194 tools. The median starts at $18/mo but the mean is $178, 51% offer a free tier, and three tiers is the industry standard. Software pricing is bimodal.
AI for Marketing 2026: AI Writes, It Does Not Yet Send
First-party data on where AI concentrates in the marketing stack. AI is deepest in generative tasks (copywriting 91%, content marketing 63%) and thinnest in operational ones (email marketing 32%): in marketing, AI writes, it does not yet send.
AI for Legal 2026: Deep AI, Almost No Free Tier
First-party data on AI in the legal stack. AI is deepest in document work (80% of document automation, 100% of AI-for-legal tools), yet legal is the least-free corner of software: legal practice management offers a free tier in just 3% of tools.
