.svg){kind=icon}
ENTERPRISE 5.0
The BGV Beat: Q1
KEY TRENDS
Architecting the Next Intelligence Stack
Powering the Stack: The Energy Infrastructure Supercycle
Behind every AI copilot and agent lies a transformation most users never see. The U.S. is entering a multi-trillion-dollar buildout of data centers, power grids, cooling systems, and semiconductor supply chains. Inference, not training, is becoming the dominant driver of power consumption, and the mix between large and small language models will shape demand for decades. Even in the low case, energy needs rise substantially. The next phase of AI will be governed less by silicon and more by steel, copper, concrete, and regulation.
Emergent Capabilities: World Models & Nested Learning
For five years, large language models absorbed most of AI’s capital, talent, and attention. A quieter shift is now unmistakable. World models learn physics, causality, and 3D environments, opening markets that LLMs cannot reach, from robotics to drug discovery. Google’s Nested Learning challenges a core assumption: that learning ends once deployment begins. Together, they move AI from pattern matching in language to grounded intelligence that keeps learning as it runs. The last decade of AI learned from the internet. The next will learn from the world itself.
DEEPER INSIGHTS
Enterprise 5.0
Enterprise 5.0 Expanded: Systems of Action
The first wave of enterprise AI was built on copilots that helped humans work faster. The next shift is already underway: agent-first systems of action, where software becomes the default executor of work. Systems of record stored information and systems of engagement improved interaction. Systems of action own execution across high-friction workflows. At BGV, we focus on industries where labor is constrained, workflows are dense, and data is fragmented, such as retail, healthcare, and energy. The companies that own this layer will shape the future of enterprise software.
Agentic AI & the Future of Work
Opening the HAI Co-Innovation Summit in Paris, Xavier Vasquez of IBM explained why most enterprises remain stuck between pilots and scaled AI: real returns come from redesigning the operating model, not adding technology on top. IBM practiced what it preaches, embedding AI in 70% of its workflows and generating $3.5 billion in productivity gains. He also looked ahead to what comes next: small domain-specific models, energy-efficient chips, and quantum computing moving from research to real-world use cases.
Events
Financing Events
XCaliber Health raises $6.5M to build an agentic operating system for healthcare. The startup replaces fragmented point solutions with AI agents that run workflows end to end, from prescription refills to referral coordination and lab notifications, while keeping humans in control of critical decisions.
Webinars
Perspectives
Webinars
What’s Next in AI Innovation — Investor Panel JPMorgan, BGV, Forgepoint Capital, Eurazeo
Investors from JPMorgan, BGV, Forgepoint Capital, and Eurazeo discussed where AI innovation is heading as the platform shift accelerates. The panel explored the rise of physical AI, the unbundling of traditional software and the move toward outcome-based pricing, and why discipline on unit economics matters more than ever. Panelists agreed foundation models are overfunded, while systems of action, robotics, and security for AI remain under-looked opportunities.
Watch here
Webinars
What We Need from AI Startups — Corporate Panel
Novo Nordisk, Santander, LexisNexis Risk Solutions, White & Case
Novo Nordisk, Santander, LexisNexis Risk Solutions, White & Case
Moderated by Novo Nordisk, leaders from Santander, LexisNexis Risk Solutions, and White & Case shared what it really takes for AI startups to win in the enterprise. The panel’s message was clear: fall in love with the customer’s problem, treat governance and compliance as part of the product, define success before a pilot starts, and co-create with enterprises. As one panelist put it, a great demo is not the same as being enterprise ready.
Watch here
Webinars
The Next Wave of Enterprise AI — Startup Showcase
Multiverse, Fiddler, Dataiku, Neo4j
Multiverse, Fiddler, Dataiku, Neo4j
Multiverse , Fiddler, Dataiku, and Neo4j showcased how they help enterprises move AI from pilots into production. Multiverse compresses models to run AI faster, cheaper, and on sovereign infrastructure. Fiddler presented its control plane to monitor, secure, and govern AI agents. Dataiku made the case for a reasoning layer to re-engineer complex business decisions, while Neo4j showed how a knowledge graph gives agents the context and memory they need.
Watch here
Webinars
The Next Wave of Enterprise AI — Startup Showcase
Maisa, Greenly, Qevlar, Asenion
Maisa, Greenly, Qevlar, Asenion
Maisa, Greenly, Qevlar, and Asenion showcased how AI is moving into some of the most demanding enterprise environments. Maisa deploys auditable digital workers that turn complex processes into verifiable code. Greenly uses AI to scale carbon accounting across entire supply chains. Qevlar’s agentic SOC platform investigates every security alert in minutes, and Asenion helps organizations govern their AI and become audit ready for the EU AI Act.
Watch here
Contact
Let's Talk
Whether you are a founder building the future of enterprise AI, an investor seeking long-horizon alpha, or a corporate leader looking for the frontier, we want to hear from you.