Series B Funding Diversifies as Round Sizes Hit Record Highs

Series B funding broke expectations in the past six months. Everyone assumed AI would dominate the pipeline. Reality proved different.

The rounds spread across biotech, robotics, security, and more. Sure, AI leads the pack. But the Nvidia-backed companies getting series b funding don’t follow a single playbook. Diversity returned to growth-stage investing after years of concentration risk.

The numbers tell the story. Annual Series B capital climbed steadily after hitting bottom in 2023. This year started strong. Round counts held steady too—no dramatic thinning of the funding pipeline that plagued late-stage deals.

Biggest checks of the past six months.

Reflection AI grabbed the largest round. The foundation model developer raised $2 billion in a Nvidia-led financing. Founded in 2024 by former Google DeepMind researchers, the company went from launch to mega-round in months.

Kailera Therapeutics raised $600 million in October. The obesity treatment developer was just one year old. Oral medications for weight loss attracted massive capital as the sector exploded beyond injectables.

Physical Intelligence pulled $600 million in November. The AI robotics startup landed Google‘s CapitalG as lead investor. Robotics plus AI equals checks VCs can’t resist writing.

Those three deals alone totaled $3.2 billion. But here’s the thing: They span foundation models, biotech, and robotics. Not cookie-cutter AI software plays.

Round sizes keep growing.

Average series b funding hit $68 million so far in 2025. That appears to be the highest on record. The trend started years ago—average round sizes inched higher since 2020. Not as dramatic as late-stage record-setting, but the trajectory points up.

Meanwhile, small Series B rounds vanished. From 2020 through 2023, about 150 rounds between $1 million and $10 million closed annually. Last year? Only 44 such rounds. VCs consolidated capital into fewer, larger bets.

When I was at Greycroft, we saw this pattern emerging. Funds raised bigger vehicles, wrote bigger checks, took bigger ownership stakes. The math only works if you can deploy $50 million-plus into winners. Writing $5 million checks doesn’t move the needle for a $500 million fund.

Follow the money. Fund economics drive everything.

Where the capital went.

Healthcare and biotech captured over 25% of series b funding over the past six months. Obesity treatments, rare disease therapies, diagnostic platforms—all attracted growth-stage capital. The sector delivers binary outcomes VCs understand: FDA approval or bust.

Robotics and hardware grabbed roughly 15%. Physical world applications came back into fashion after a decade of pure software dominance. Investors remembered that atoms plus bits can generate defensible moats.

AI-related categories took about half of Series B investment across multiple industries. That’s the headline number everyone expected. But the categories span vertical AI, foundation models, robotics, security, and enterprise tools. Not monolithic.

Software-focused companies captured the majority of capital. No surprise there. Software margins and scalability remain unmatched for VC return math.

What this actually means.

Series B investors consolidated their bets somewhat. Fewer small rounds, bigger average checks. But they spread capital across sectors and technologies. No single vertical dominated completely.

Question is whether diversification signals confidence or confusion. Confident investors back multiple themes because opportunity spans sectors. Confused investors spray capital around hoping something works.

I’d argue this looks like confidence. To reach Series B requires technological edge, early traction, or both. These aren’t risky seed bets on unproven ideas. Companies proved something by the time they raise $50 million-plus in growth capital.

VCs won’t tell you this, but Series B is where pattern recognition matters most. Too early and you’re guessing on founders. Too late and valuation compresses returns. Series B hits the sweet spot: proven enough to reduce risk, early enough to generate 10x-plus if things work.

Here’s what the term sheet doesn’t say at this stage. Liquidation preferences start stacking. Participating preferred shows up. Ratchets get negotiated. VCs taking $68 million average positions demand downside protection. Founders celebrate the raise, but dilution and terms matter more than headline valuation.

Valuation is vanity. Terms are sanity.

The Boston VC ecosystem knows this cold. Firms here wrote the playbook on growth-stage investing. Battery, General Catalyst, Spark—they built franchises on Series B expertise. Pick right at this stage and returns compound beautifully. Pick wrong and capital disappears into zombie companies that can’t exit.

What comes next.

Round sizes will keep growing. Fund sizes keep growing, so check sizes follow. The $68 million average probably hits $75 million-plus by year-end. Late-stage rounds already set fresh records quarterly.

Small Series B rounds won’t come back soon. The fund economics don’t support them. Emerging managers might write $10 million checks, but established funds focus capital on bigger swings.

Sector diversification probably persists. AI hype will fade eventually—always does. When it does, capital rotates to healthcare, climate, infrastructure, whatever narrative captures imagination next. For now, at least Series B investors aren’t putting all capital into one basket.

The real test comes in 3-5 years. These companies need exits. IPO markets need to reopen. Acquisitions need to return. Without liquidity, even great Series B picks generate zero cash for LPs.

I’ve seen this movie before. Strong Series B vintages in 2012-2014 generated exceptional returns by 2018-2020 because exit markets opened. Strong vintages in 2020-2021 still wait for exits because public markets slammed shut.

Timing matters as much as picking.

For now, Series B looks healthy. Diverse sectors. Growing round sizes. Steady deal counts. Whether diversity translates to returns depends on what happens between now and exit. That timeline stretches 5-7 years minimum.

Patience required. Capital deployed today won’t return until 2030-plus.