State of Australian Startup Funding 2025 × NuScore: What the Data Says About What VCs Actually Fund

The 2025 State of Australian Startup Funding report (Cut Through Venture / Folklore Ventures) paints a picture of recovery: $5.1 billion across 390 deals, up 24% from 2024. But the headline numbers hide structural shifts that change how founders should approach their raise.

We cross-referenced the report's findings with 197 NuScore-evaluated pitch decks to understand what is actually driving investment decisions in the Australian market.

What Are the Key Numbers for 2025?

What Do NuScore Evaluations Reveal About These Trends?

1. AI positioning matters more than "tech"

61% of all capital flowed to startups with an AI offering. In our NuScore data, AI/ML-specific decks score 5.91 on average — vs 5.08 for generic "technology" decks. That is a 0.83-point gap on a 10-point scale.

Implication: If you are building with AI, say so explicitly. Generic tech positioning leaves points on the table.

2. Diligence is up 40% — and product is what they are diligencing

The report notes "diligence requests are up 40%" with "execution risk as the new emotional trigger." Our NuScore data confirms: product execution is the #1 predictor of evaluation outcomes (r=0.869). VCs are spending more time on product diligence, not just market narratives.

3. Victoria overtook NSW — different cultures, different evaluation

VIC: $1.9B across 134 deals. NSW: $1.7B across 160 deals. First time Victoria has led on capital.

Our thesis analysis reveals why: Melbourne VCs are the most team-focused globally (24.1% mention founders), writing fewer but larger conviction-driven cheques. Sydney VCs balance team and market equally (15.6% / 16.9%), spreading capital across more deals.

4. Founders are raising younger — traction data is sparser

Pre-seed median company age dropped from 1.2 years (2024) to 0.7 years (2025). Founders are going to investors earlier, which means less traction data in the deck.

NuScore handles this through stage-aware weight redistribution: pre-seed decks are not penalised for missing financials or thin traction. The engine shifts weight to team and market — the dimensions that matter at 0.7 years old.

5. Female founder capital doubled — but deal participation declined

Capital share: 15% (2024) → 24% (2025). Deal share: 28% → 24%. More capital per deal for female founders, especially in Queensland (34% female founder participation). This represents meaningful progress on capital access but warrants continued monitoring for systematic evaluation bias.

What Are the Current AU Raise Benchmarks?

If you are raising in Australia in 2026, these are the median round sizes from 2025 (source: State of Australian Startup Funding 2025, Cut Through Venture / Folklore Ventures):

Stage	2025 Median (AUD)	2-Year Change
Pre-Seed (Angel)	$1.0M	+67%
Seed	$2.5M	+25%
Series A	$11.0M	+38%
Series B+	$30.0M	+50%

NuScore now calibrates against these 2025 benchmarks. When you upload your deck, your raise target is compared against AU-specific medians — not Silicon Valley numbers. A $2.5M AUD seed raise is at the median for Australia but would be below median in the US ($3M+ USD).

See how our Say-Do Gap research explains what Australian VCs actually evaluate, and why Melbourne investors are uniquely team-focused compared to other global hubs.

See how your deck scores against AU benchmarks → nuvc.ai

Data Sources and Methodology

Market data: State of Australian Startup Funding 2023, 2024, 2025 — annual reports by Cut Through Venture and Folklore Ventures, drawing on 1,000+ ecosystem contributors. Data covers VC, angel, and venture debt investment in Australian-founded startups.

Scoring data: 197 pitch decks evaluated by VCGrade 2.0 across 5 dimensions (team, market, product, traction, financials). Hybrid LLM+rules engine with 4-tier cross-check. Sector and industry classification from deck metadata.

Investor data: 7,800+ investor records in the NUVC database. Thesis analysis based on investors with written investment theses (>20 words). Geographic classification from self-reported location.