SVB just released their latest 2026 State of the Markets report, and it’s a great data-rich snapshots of the venture and startup ecosystem I’ve seen in a while. We dug through all 39 pages so you don’t have to.
Screenshot
Here are the 10 biggest takeaways — and what they actually mean if you’re building or funding a SaaS company today.
#1: Venture Capital Has Split Into Two Completely Different Industries
This is the single most important structural shift in the report, and it affects everything downstream.
In 2025, nearly $340B flowed into US VC-backed companies — the second-highest year ever. Sounds great, right? Except the top 1% of companies by valuation captured a full third of all that capital. Meanwhile, the bottom 50% of companies got just 7%.
And the number of deals actually fell 15% year-over-year even as dollars invested jumped 53%.
What does this mean practically? There are now two completely separate venture industries operating under the same label. At the top, you have what is effectively late-stage private asset management — billion-dollar checks going into Series H and Series J rounds for companies like OpenAI and Databricks. These funds generate returns primarily through management fees, not carried interest. At the other end, you have traditional early-stage venture, where founders are still scrapping for $5M Series A rounds and the fundraising environment remains genuinely difficult.
If you’re a founder outside the AI mega-round club, recognize that the headline numbers about “record VC investment” have almost nothing to do with your fundraising reality. Plan accordingly.
#2: The Revenue Bar to Raise Your Next Round Is Higher Than Ever — And Growth Is Slower
This is the compounding challenge that’s quietly crushing a lot of startups right now.
SVB’s proprietary data shows that median revenue at time of raise has increased substantially across every series compared to 2021. A median Series A company today is raising at around $2.5M in trailing revenue. A median Series B is at $6M. Series C is at roughly $15M.
That alone would be manageable. The problem is that top-quartile revenue growth rates have been cut roughly in half since 2021 at every stage. Seed-stage top-quartile growth dropped from around 960% to 320%. Series A dropped from roughly 480% to 165%. Series B from about 240% to 125%.
So you need more revenue to raise your next round, but you’re growing slower getting there. This is why graduation rates have plummeted — only about 3% of seed companies are making it to Series A within 12 months today. The median seed company needs to grow revenue roughly 11x to reach median Series A benchmarks. That’s a massive leap when growth rates are compressing.
The practical takeaway: build your financial plan around 24-30 months of runway minimum, not 18. And be honest with yourself about whether you’re tracking to the revenue benchmarks for your next round, because the goalpost has moved.
#3: Extension Rounds Have Become the New Normal — And That’s Actually OK
Here’s something that would have been stigmatized three years ago but is now completely mainstream: nearly 18% of all Series A deals completed in 2025 were raised by companies that had previously done a seed extension round.
The extension round has become the bridge between the old fundraising cadence and the new reality of higher benchmarks and slower growth. And the data suggests it’s not as punitive as founders fear. SVB found that the median return per dollar invested from an extension round to the next series is only about $0.61-$0.67 less than investing in the original round.
For founders, the lesson is clear: don’t treat an extension round as a failure. Treat it as a tool. If you’re 60-70% of the way to your next-round benchmarks and you have strong investor relationships, an extension can be the smart play versus trying to raise a full round at a discount.
#4: AI Valuations Are Real, But the Bubble Signals Are Unmistakable
SVB makes a striking comparison: the top five AI unicorns (OpenAI, xAI, Anthropic, Databricks, and Scale) are collectively worth more than $1.2 trillion — exceeding the total market value of every single IPO during the entire dot-com era, adjusted for inflation.
Total AI venture investment since 2022 has hit $560B, roughly matching the entire dot-com era’s VC deployment.
The parallels are instructive but not perfectly symmetrical. SVB notes clear signs of overexuberance in AI: elevated burn multiples, low revenue per employee, and valuation premiums for companies generating less revenue than their non-AI peers. AI companies command valuation premiums of anywhere from 10% at seed to 222% at Series D compared to non-AI peers.
But here’s what matters for SaaS founders specifically: companies that have integrated AI into their products are getting a valuation premium regardless of sector. This isn’t just about pure-play AI companies. Enterprise, fintech, consumer — across the board, AI integration translates into higher valuations.
The actionable insight: you don’t need to be building an LLM to benefit from the AI wave. But you do need to have a credible AI story woven into your product. As David Frankel of Founder Collective put it in the report, calling yourself an AI company today is like calling yourself an internet company in 2004 — it should just be table stakes.
#5: The IPO Window Is Open, But Public Markets Are Punishing
2025 was the best year for VC-backed tech IPOs since 2021. That’s the good news. The bad news is that the bar is extraordinarily high and the market’s reception has been lukewarm at best.
The median revenue at IPO in 2025 was $246M — the highest of any cohort in the modern venture era. Median operating margin at IPO was -16%. Median revenue growth was 35%. Companies were an average of 12.5 years old and had raised a median of $400M in equity before going public.
And after all that? Only about half of the 2025 IPO cohort was trading above their last private valuation at the time of the report. Less than a third were above their initial public market value.
What this means for late-stage SaaS companies: the IPO is no longer the triumphant liquidity event it once was. It’s increasingly a necessary step for companies that have grown too large for private markets to support, not a celebration of arrival. If you’re running a company approaching IPO scale, the secondary market can buy you time — but as Eric Thomassian from Forge noted in the report, secondaries are a complement to an IPO, not a replacement for one.
#6: VC Fundraising Has Fallen to a 7-Year Low — But That Creates Opportunity
Here’s a stat that doesn’t get enough attention: US VC fundraising dollars fell almost 20% year-over-year to their lowest level since 2019. GPs are spending nearly two years in market to close a fund — double what it took in 2021. And when they finally close, they’re typically right around their initial target, not blowing past it like in the boom years.
But there’s a silver lining embedded in this data. About one-fifth of all US VC fundraising in 2025 came from first-time fund managers. And 61% of those new firms were founded by at least one investor spinning out of a firm with over $10B in AUM.
The LP community is shifting capital toward early-stage managers and away from large platform funds. This is creating a new generation of investors who are building their reputations right now and will be hungry for deals. For founders, these emerging managers can be excellent partners — they’re motivated, they have deep networks from their prior firms, and they need portfolio companies to succeed to establish their track record.
#7: Defense Tech Has Gone From Taboo to $21B — But Buyer Beware
US VC investment in defense tech nearly doubled its prior high-water mark in 2025, hitting over $21B. The number of companies funded in the space reached 306. Defense spending was 3.4% of US GDP in 2025 and could reach 4-5% by 2027 under the proposed $1.5T defense budget.
But here’s where it gets complicated: recent defense tech IPO performance has been poor. SVB’s data shows that 2025’s three defense tech IPOs returned less than 150% annualized relative to the S&P 500. Many of the de-SPAC defense companies from 2021-2022 have performed terribly, with multiple companies showing triple-digit negative annualized returns.
The challenge is that defense companies face fundamentally different dynamics than typical SaaS businesses. Government procurement cycles are long and complex. Revenue can be lumpy and hard to predict. And profitability often lags even when revenue growth is strong.
For SaaS founders looking at government and defense as a market: the opportunity is real, but go in with eyes open about the sales cycle and the path to profitability. The investors flooding into this space are used to simple SaaS contracts, and government is a very different animal.
#8: Cash Runway Has Improved, But Don’t Get Comfortable
After years of deterioration, cash runway conditions finally improved across all sectors in 2025. The median VC-backed tech company now has about 14 months of runway, up from 13 the year prior. The percentage of companies with less than 12 months of cash has stabilized.
But the improvement is modest, and there’s significant variance by sector. Frontier tech companies have the shortest median runway at around 10-12 months. Fintech companies have the longest at around 14-16 months. Enterprise software sits in between.
The more interesting finding is what’s happening post-raise. SVB’s data shows that in 2025, companies began accelerating both burn and revenue growth after raising. On average, burn increased 50% and revenue growth increased 75% in the year following a raise. This is a meaningful shift from the 2022-2024 period when companies were focused almost exclusively on extending runway and improving efficiency.
The signal here: we may be entering a new phase where the best companies start leaning back into growth. If capital markets remain supportive, expect burn multiples to tick up in 2026 as companies invest in capturing the AI opportunity.
#9: San Francisco Is the Only Tech Hub Where Company Formation Is Actually Growing
US company formation rates are at their lowest point in six years. Exclude AI companies and they’re at the lowest in over a decade.
But San Francisco is the exception. It’s the only major tech hub to see growth in company formation since 2022, driven almost entirely by AI. Meanwhile, cities like Atlanta, which was an emerging tech hub, have seen VC deals, company formation, and investment slow considerably — down 47% since 2022.
What’s interesting is that while company formation is concentrated in SF, office space demand is broadening nationally. WeWork data shows tour activity up across the board, with Boston (+103%), Miami (+66%), and Chicago (+58%) seeing the biggest increases. The return-to-office trend is settled, and companies are building flexible hybrid footprints across multiple cities.
For founders: if you’re building an AI company, the gravity of San Francisco’s talent pool and investor network is stronger than ever. For everyone else, the geographic arbitrage play still works — you can build a great company in many cities, but the fundraising relationship still tends to flow through California-based investors, who participate in 30% of all US deals.
#10: The Dot-Com Parallel Is Real — But So Is the Opportunity
SVB’s most provocative comparison runs throughout the report: AI investment today mirrors the dot-com era in both scale and speed. Total AI VC deployment since 2022 roughly matches the entire dot-com era. Domain registrations on .ai are following the same trajectory as .com domains in the late 1990s. And adoption curves suggest we’re at roughly 1998 levels of internet adoption, with about a third of US adults using an AI chatbot.
But here’s the crucial difference from the dot-com era: unlike internet adoption, which was opt-in, AI is being embedded into services people already use. And unlike the dot-com bust, where most of the value was concentrated in newly public companies, the AI value is concentrated in private companies that don’t need to go public.
History suggests that the bubble will deflate at some point. SVB is quite direct about this. But history also shows that every major platform shift — internet, mobile, cloud, SaaS — created lasting businesses that defined the next decade of technology. The question isn’t whether AI will create massive long-term value. It’s whether you can catch the wave without getting caught in the inevitable correction.
The bottom line from this report is that we’re in one of the most bifurcated markets in venture history. If you’re at the top of the power law — AI-native, hyper-growth, repeat founder — capital has literally never been more abundant. If you’re everyone else, the environment is harder than the headline numbers suggest, but it’s improving. Deal activity is starting to tick up. Graduation rates are stabilizing. Exits are recovering. And a new generation of fund managers is hungry for deals.
Build for 24+ months of runway. Know your revenue benchmarks cold. Have a credible AI story. And remember that the best companies in history were built during periods that felt exactly like this one.
