One word to describe Q2 2026: Concentrated.

Venture capital had its biggest half-year ever, and almost nobody felt it. US VC deployed $412.7 billion in the first half of 2026, already nearly 30% more than all of 2025, but 87.5% of those dollars went into rounds of $100 million or more. Records at the very top, contraction almost everywhere beneath it. Here is what we are seeing in the market:

1. AI is still eating the venture market. 86.0% of all venture dollars in H1 went to AI companies. Five of Q2's seven billion-dollar-plus rounds were AI, led by Anthropic's $65 billion round at a $965 billion post-money valuation. Valuations have now pushed past their 2021 highs at every single series, and this time it is not cheap money driving it, it is AI expectations. Last quarter I said we want to see a greater delta emerge between the winners and losers in enterprise AI before we lean back in. That delta is starting to show up in revenue, but it has not shown up in price yet, which means entry discipline matters more than ever. When valuations assume flawless execution, we underwrite for the execution, not the hype.

2. Florida is having a record year. Tampa Bay is on track for its best funding year in more than a decade. Companies in the Tampa-St. Pete metro raised $209.8 million across 25 deals in Q2, up from $129 million in Q1. When you include the Sarasota-Bradenton metro, the region pulled in more than $650 million in the first half of this year. A big driver of that number is TENEX.AI, the Sarasota-based cybersecurity company that raised a $250 million Series B led by Crosspoint Capital at a valuation north of $1 billion in Q1. TENEX.AI is an AI-native managed detection and response company, was seeded by a16z, set up its world headquarters in Sarasota, and was just named the fastest-growing cybersecurity company in the country. A Florida-built AI cybersecurity unicorn is not a fluke, it is exactly what our thesis says should be happening here. Statewide, Florida companies raised roughly $3.1 billion in the first half, per PitchBook, Miami pulled in $832 million in Q2 and is tied with Austin as the number five metro in the country by deal count, and one more stat I will shamelessly include: PitchBook noted the largest fund raised outside South Florida was our own $60 million FLF Fund 3 (total commitments).

3. We are building bridges at pre-seed. This quarter we went live with two new programs that push us even earlier in the funnel:

• The FLF Accelerate Fund. A $5 million hard-cap vehicle that invests in companies coming out of a16z's speedrun accelerator, the most competitive accelerator in the world right now with an acceptance rate under 1.0%. a16z puts up to $1 million into each company and wraps them in a 600-person operator platform. We are not approaching this cold: we have attended two cohorts, have a standing relationship with the speedrun team, and have direct access to the General Partner who sponsors the program. These companies move fast, and getting on their cap tables at the accelerator stage is a wedge we could not buy any other way.
• Florida First Check. In partnership with The Florida Institute, we launched an evergreen pre-seed strategy, managed by Florida Funders, that writes $75,000 to $100,000 first checks to technology founders across the state, working alongside Florida's accelerators, universities, and incubators. Florida has never had a shortage of brilliant founders; what has been missing is a reliable institutional first check that keeps them building here instead of relocating for capital. Our first check went to Stello AI, a Tampa-based AI HR platform.

Both programs will continue to compound our proprietary deal flow, and that matters because of what is happening at the earliest stage of the market right now.

4. More companies are starting. Fewer are breaking through. The US is on pace for more than 10,000 companies raising their first round of venture capital this year, an all-time record, while seed round sizes have stayed flat and the tourists have left: corporate VC participation is at a decade low, crossover fund deal counts are the lowest since 2019, and first-time fund formation is the lowest since 2016. That is a favorable buyer's environment at our entry point — pre-seed and seed — even as later-stage prices balloon. But here is the catch: It has never been easier to start an AI company, so more entrepreneurs are flooding in, and with them comes more competition for the same revenue. We are seeing a lot of companies get to $1 million in ARR very quickly, and some even sprint to $5 million, but fewer companies are breaking through that $5 million mark. Because we invest early, we need to see the evidence at the time of investment that a company can punch through that wall: real pipeline traction and larger contract sizes, usually deals above $100,000 in ACV. Small logos prove somebody will try your product. Six-figure contracts prove an enterprise will run its business on it.

5. Liquidity is thawing from the top down. SpaceX went public in Q2 at a $1.7 trillion valuation, raising $75 billion in the largest IPO ever completed and making Q2 the biggest exit quarter on record. OpenAI and Anthropic have both confidentially filed, and M&A value is at a decade high of $375.4 billion year to date. We are always looking to prioritize liquidity. With our most recent exit of RepScrubs in Fund 1, that fund now sits above the 75th percentile on DPI, measured on paid-in capital, against Carta benchmarks for peer funds of the same vintage, and we are seeing real strength from secondary buyers and private equity acquirers for our companies. On SpaceX: as I said last quarter, this is a massive event for our industry. But let me be straight with you about how it has traded. Insiders and private-market investors have not yet been able to sell their positions, and after a record day one the stock has since declined. That is not a great sign for the IPOs to come, including Anthropic and OpenAI, but there is no need to panic, as it is typical for the IPO market. Of the ten largest US tech IPOs, only three traded positive in their first year as public companies. The flywheel still turns: lockups expire, distributions flow, and LPs recycle capital. We believe we will have several more exits this year.

When looking at the FLF portfolio, Q2 2026 delivered two of the biggest results in our firm's history:

1. Flex, a FLF Fund 2 portfolio company, raised a $70 million Series B1 led by Halo Fund at a $1.2 billion post-money valuation, making it our second unicorn in the Fund 2 portfolio. We are now carrying the position at approximately 8x gross MOIC on capital invested across all FLF vehicles in the company, and approximately 16x on our original entry. Both are unrealized marks based on the new round's price.
2. RepScrubs, a FLF Fund 1 portfolio company, was acquired, delivering a realized 11.84x gross MOIC on capital invested and returning approximately 30% of Fund 1's paid-in capital in a single exit. RepScrubs is our strategy working exactly as designed: a Florida-based company, a low entry valuation, and a company that never raised additional capital, so we were never diluted. It was not a unicorn, and it did not need to be. Disciplined entry plus zero dilution turned a good outcome into a fund-defining one.

Overall, Q2 was a milestone quarter for FLF: record activity in our home state, two new engines for proprietary deal flow, a new unicorn, and an exit that returned roughly 30% of Fund 1's paid-in capital on its own. We will keep prioritizing liquidity, especially in Fund 1 as we approach the end of that fund's lifecycle, and we will keep being disciplined buyers in a market that is anything but. As always, thank you to our founders and investors, because without you, we would not be able to continue on our mission to invest in the best founders in the country.

Thank you,

Saxon Baum