From Text to Action: Innovation Learns to Move Through Space and Time

Innovation has always had its emblematic moments – the garage in Palo Alto, the dorm room in Cambridge – but 2026’s defining image may be something far less romantic and far more revealing: billions of hours of video game footage quietly powering the next frontier of artificial intelligence.

New York–based General Intuition has just raised $320 million in a Series A round at a $2.3 billion valuation, one of the largest AGI-oriented financings of the year.[2][4] The round is led by Khosla Ventures, with participation from Jeff Bezos, Eric Schmidt, General Catalyst, and researchers connected to MIT and Google DeepMind – a roster that reads like a roll call of technology’s most aggressive risk-takers.[2][3][9]

What they are backing is not another text-trained chatbot. General Intuition is training AI on billions of gameplay clips from the Medal platform, using the embedded action labels – which button was pressed, at what moment, and with what consequence – to teach models spatial-temporal reasoning, the ability to perceive and act in space and time.[2][4] In other words, innovation is shifting from systems that describe the world to systems that can move through it.

This is more than a clever data hack. It signals a deeper turn in innovation culture: towards embodied intelligence, towards AI that can operate in games, simulations, robotics, industrial automation, and autonomous systems.[4][5] The bet is that video games encode the physics, constraints, and decision sequences that text alone cannot capture – and that those patterns are the missing bridge from clever language models to general-purpose agents.

The Venture Capital Machine Learns to Experiment Faster

A second, quieter revolution is unfolding not in labs but in term sheets. At the NBER Spring 2026 “Productivity, Innovation, and Entrepreneurship” conference, researchers report that venture capital has structurally shifted toward more frequent, smaller investments, a pattern that is boosting new firm formation.[10] According to the conference summary, generative AI is lowering the cost of entry, enabling investors to place more bets on more ideas, more often.[10]

This matters because capital structure is innovation infrastructure. When financing moves from a handful of large, high-stakes rounds to a portfolio of smaller, rapid experiments, it changes who gets to try, what they can try, and how fast they can iterate. Gen AI tools make it possible to prototype products, automate back-office work, and test business models with dramatically less capital – and VC is adjusting in real time.

The numbers behind that adjustment are stark. Ecosystem data for H1 2025 – now framing 2026 analyses – show $205 billion raised, up 32% from H1 2024, marking the strongest half-year for VC since 2021.[1] AI startups captured 65% of total deal value, and more than half of new unicorns are AI companies.[1] Investors are calling 2026 “the year of the consumer,” but the underlying story is the same: AI-first software and automation are becoming the default lens through which new consumer and enterprise products are conceived.[1][10]

Innovation, in this view, is no longer about heroic, singular breakthroughs. It is about systematized experimentation, powered by generative tools and funded by capital that has finally learned to match the tempo of software.

Founders Everywhere: Entrepreneurship Becomes a Default Identity

If capital is innovating, so are careers. At the World Economic Forum Annual Meeting 2026, LinkedIn data presented in the session “New Dawn for Entrepreneurship?” showed a 60% year-over-year rise in people globally adding “founder” to their profiles.[7] That 60% growth is three times the rate seen since 2021, suggesting a structural tilt towards entrepreneurship over traditional employment tracks.[7]

Panelists linked this surge directly to AI tools that cut startup time and costs, making it feasible for solo operators and small teams to do what once required large organizations: build products, reach global audiences, automate operations, and access sophisticated analytics.[7] In today’s ecosystem commentary, this data point is being cited as proof of a broad-based global entrepreneurship boom in 2026.[1][7]

The cultural implications are as significant as the economic ones. For a generation that came of age in an era of layoffs, automation anxiety, and institutional distrust, building something – anything – is increasingly seen as a rational default, not a risky exception. Innovation is ceasing to be a specialized activity; it is becoming a mass occupation.

Frontier Tech Grows Up: Climate and Fusion Move Toward IPO Reality

At the other end of the maturity curve, 2026 is quietly marking a coming-of-age moment for technologies that once sat firmly in the “too early” bucket. Recent analyses of climate tech VC trends highlight aggressive investment into carbon removal, next-generation energy systems, and AI-powered climate verification platforms.[1] The headline examples read like a checklist of previously speculative fields entering late-stage territory.

Fervo Energy has raised $462 million in a Series E as it approaches potential IPO milestones.[1] Commonwealth Fusion Systems has amassed about $2.9 billion in total funding and is now described as nearing IPO-stage readiness.[1] These are not science experiments any more; they are capital-intensive companies marching along familiar pre-IPO trajectories.

Innovation here is not about novelty; it is about scale and industrialization. Carbon removal and fusion were once the province of labs and government grants. In 2026, they sit inside cap tables and roadshow decks. The shift tells us that the frontier of “deep tech” is being pulled into the same financing, accountability, and execution structures that built cloud computing and e-commerce.

The New Shape of Innovation: Systems, Not Slogans

Taken together, these developments outline a new center of gravity for innovation.

- Massive capital commitments to AGI and deep tech – from General Intuition’s $320 million Series A at a $2.3 billion valuation to multi-billion-dollar fusion plays – show that investors are prepared to underwrite long-duration, high-complexity bets.[2][4][1] - Structural shifts in venture capital, documented by NBER research, reveal an ecosystem that is adapting its investment mechanics to the speed and modularity of Gen AI.[10] - A 60% global jump in founder identities indicates that entrepreneurship is no longer confined to Silicon Valley; it is a mainstream career choice, amplified by accessible AI tools.[7] - Climate and fusion startups nearing IPO milestones demonstrate that yesterday’s frontier technologies are becoming tomorrow’s regulated industries.[1]

Innovation, in this moment, is less about the lone genius and more about the systems that allow many people, in many places, to act on their ideas: data-rich platforms like Medal, capital structures that reward rapid iteration, AI tools that compress the distance from concept to company, and deep-tech ventures that translate research into infrastructure.

If there is a single thread running through 2026’s startup landscape, it is this: the future is no longer waiting to be discovered. It is being trained, funded, and shipped – frame by frame in game engines, line by line in generative models, and megawatt by megawatt in geothermal wells and fusion testbeds. Innovation has become a continuous process, not an isolated event, and the world is reorganizing itself around that fact.