Investing in women is investing in the future of AI

AI will determine who owns the next generation of economic infrastructure. We're in the early innings of an exponential take-off where AI will soon be utilised by nearly every person, at every company, in every country, every day. But there's a problem hiding in plain sight. Some of the people most likely to catch a model's failure modes, whether that be in hiring algorithms, credit scoring and medical diagnostics, are the people least likely to be in the room when those models are being built, and it has a real cost. 

The AI industry is attracting unprecedented capital, but 98% of venture capital still bypasses women. In a world where the founding team's perspective is literally baked into the model's training objectives, evaluation criteria, and deployment context, the gap is a meaningful design constraint. 

We have already seen this play out at the nation-state level, where governments are actively intervening to ensure their cultural norms are embedded in AI training objectives. Yet, almost no one is focused on the fact that roughly 50% of the population (women) remain structurally excluded from where those objectives are set. Investing in women is not a diversity initiative, but a strategic decision for the future of AI. 

What a fully supported ecosystem could look like 

The AI investment narrative of the last three years has been dominated by foundation model funding; the race to scale parameters, training compute, and benchmark performance. That race has produced extraordinary science and extraordinary valuations. It has also produced a crowded, capital-intensive layer where differentiation is increasingly difficult to sustain.

Women hold senior leadership and research roles in some of the most prominent AI companies, including Anthropic. But the next wave of durable value will be built at the application layer; verticalised AI that works within regulated industries, earns trust with skeptical enterprise buyers, and navigates compliance frameworks and messy real-world data, and real-world consequences.  

At this layer, the question is not how to optimise the model but whether it will actually work in a hospital, a courthouse, or a bank; where the data is messy, the stakes are real, and getting it wrong has significant consequences. These tools are already shaping hiring decisions, healthcare outcomes, and who gets access to credit. Having women's influence and impact on the broader ecosystem would be far greater if capital flowed more equitably. 

The network problem compounds at the worst possible moment 

Venture capital is a relationship business, and AI is compressing timelines in ways that make relationships even more consequential, not less. Pre-seed and seed rounds are closing faster. The window between first check and meaningful commercial traction is shortening. The founders who move through that window quickest are those who already have warm introductions to the right investors, design partners, and engineering and research talent.

Female founders are structurally underrepresented in these networks; not because they are less connected, but because the networks themselves were built around a founder archetype that has looked historically different. In the world of AI, early access to compute partnerships, research collaborations, and enterprise pilot programs can determine whether a company exists in two years. 

The market failure here is well-documented, but what's less discussed is what the ecosystem loses; founders who bring adversarial thinking about failure modes, genuine domain expertise from adjacent careers in healthcare and financial services, and a commercial orientation toward the unsexy, sticky, high-retention problems that enterprise buyers actually pay to solve. The challenge is not a lack of capital flowing into AI, it's how that capital is distributed. 

Advancing capital for women founders  

Addressing this imbalance requires deliberate, structured intervention, particularly at the earliest stages of company formation. 

The female founder's gap is not a talent problem, it is a capital and network access problem, and both are solvable with deliberate design. Investment models like the Anthemis Female Innovators Lab Fund (AFIL) demonstrate how strategic capital can  provide distribution leverage, hands-on operational support, and bridge the post-investment gap where many early-stage company's stall. 

And, as companies build, scale, and exit, their founders are entering the broader ecosystem as angel investors, board members, and mentors, with their own capital, networks, and hard-won pattern recognition to deploy. Each successful company creates the conditions for the next cohort to move faster and raise more efficiently. That is how you build a self-reinforcing pipeline; not through mandates, but through momentum.

The allocation decision in front of us

AI will define productivity, labour markets and global competitiveness for decades. The decisions being made today, about who gets funded, whose problems get resourced, and whose networks shapes what gets built, will be embedded in economic infrastructure for a generation. These are not abstract choices, but allocation decisions with long tails and compounding consequences.  

Women are more than contributors to AI, they are central to building resilient, robust, and ethical growth of the sector. As AI becomes core infrastructure, diversity in leadership and ownership becomes a matter of long-term economic strength.

Investing in women in AI is not about shifting opportunity at the edges. It is about backing the full range of builders working on the hardest problems. Investors who recognise this early will not be doing diversity a favour, and they will be doing themselves one.