Call for Founders

The first in a bi-annual series outlining where Forum is most actively looking to meet founders, whether you are exploring an idea, validating it, or already building.

At Forum Ventures, we closely watch major technology and economic shifts, and reflect on their second and third order effects. As these downstream impacts compound, they often expose gaps that existing tools, business models, and institutions weren’t built to address. 

The themes below highlight where we see those gaps emerging today and where we’re excited to explore new ideas alongside founders and builders, whether that’s through our AI Venture Studio, accelerator, or pre-seed fund. 

Compute Starts Acting Like a Commodity 

As demand for inference grows, compute is beginning to take the characteristics of global commodities markets. For most enterprises, compute will be a massive capital expenditure.  Much the same as airlines, manufacturers, logistics and agriculture companies utilize financial markets to minimize financial risks of core input costs to their businesses, new products will emerge that allow the same degree of risk mitigation for heavy compute businesses. 

We’d love to meet with founders who are thinking about: 

• Compute- Energy Arbitrage Networks: platforms where compute is priced and routed by time, location, energy availability, etc. - optimizing for cost/ predictability not just speed.
• Forwards and Hedging Market: products that let companies lock in future GPU or inference capacity at known prices, manage volatility/ supply risk, and create liquidity around compute and power commitments (including secondary markets for PPAs).
• Bundled Compute and Power Contracts Integrated procurement of compute capacity with power guarantees, hedging both simultaneously. 
• Compute Insurance and Risk Rating:  tools that price, underwrite, and guarantee compute availability and performance, with financial penalties for violations.
• AI Infrastructure Tokenization: mechanisms that tokenize compute capacity so spiky, perishable workloads can be pooled, traded, and resold, creating liquidity beyond bilateral contracts.
• Compute Usage Analytics and Optimization FinOps tools specifically for AI compute. Real-time cost monitoring, usage attribution, waste identification, and automated workload optimization across training and inference workloads.

Acute Care Moves Into the Home

In December, the House passed a bill to extend hospital at home services for another 5 years. This will open up more opportunities to deliver acute and recovery care in the home via telehealth, reducing costs and increasing patient comfort/convenience and experience. There are broad platforms partnering with Health Systems to extend their care to at-home services. One well funded company in the space shut down among uncertainty around the bill but then it ultimately got extended for another 5 years. 

We’d love to meet with founders who are thinking about: 

• Logistics for labs, medications, medical equipment needed at the home, etc. 
• Billing/RCM and quality reporting for home based care.
• Care coordination and labor capacity planning. 
• Virtual nursing infrastructure.

Medical Malpractice Litigation and Insurance Changes in the AI Era

The Medical Malpractice insurance industry in the US is approximately $8B per year and growing; and the medical malpractice litigation industry in the US is another $3B+. As AI permeates through our healthcare system in the form of AI scribes, AI diagnostic support tools, increased data collection from patients and eventually AI diagnostics; this will change how we underwrite malpractice insurance and how these cases are litigated. 

We’d love to meet founders who are thinking about:

• Unique Insurance Models: In this new reality are there different ways to underwrite and price this risk.
• Claims Predictions: Can you leverage AI to predict and prevent claims before they even happen. 
• Malpractice Prevention Layer: QA for AI scribes to mitigate liability and adhere to compliance. 
• AI System for Law Firms: AI system to ingest case information and identify real opportunities that today are going under litigated.
• AI Native Malpractice Defense Law Firm: AI generated strategies, motions and settlement negotiations.

Price Transparency Regulations Is Creating New Opportunities

There has been a push over the last several years to bring more price transparency to the healthcare industry and this current administration is pushing on it even more with stricter enforcement and larger penalties for not adhering to it. This is creating opportunities to help patients and self insured employers navigate to better, less expensive options to help drive costs down. We already made an investment in this space with Outfox Health and are excited about the long term potential there. 

We’d love to meet founders who are thinking about: 

• Real Time Compliance Monitoring for Health Systems: Especially for small to mid-size health systems that don’t have Epic’s compliance module and don’t want to pay for consultants. 
• Payer Contract Platform: Health systems work with dozens of payers and have to translate those payer contracts into clear rates to meet pricing transparency requirements
• Claims Co-Pilots & Automation: Eligibility checks, responsibility estimates, post-claim auditing, repricing, denial triage, and automated appeals.
• FHIR Native Prior Authorization Platforms: Infrastructure built for the CMS Prior Authorization and Interoperability Rule, enabling electronic prior auth and real-time data exchange across payers and providers.

The Winners of the Infrastructure Supercycle Will Build for Efficiency 

The first phase of the infrastructure supercycle has been a land grab: securing compute at any cost amid tight constraints across chips, data centers, networking, and energy. Hyperscalers and NeoClouds are deploying capital aggressively, even as unit economics at scale remain uncertain. As the cycle matures, power, networking, regulation, labor, and capital efficiency become limiting factors. The next generation of infrastructure will be built around operating within these limits. 

We’d love to meet with founders who are thinking about: 

• Compute Architecture Geography: Where to build becomes as important as what to build. Proximity to power, fiber, water, favorable regulations.
• Waste Heat Monetization: Data centers generate massive heat. Capturing and selling it for heating, industrial processes, and other uses. Turning a cost center into revenue.
• Construction and Deployment Speed: Modular data centers, pre-fabricated components, construction automation. Compressing 24-month builds to 6-12 months.
• Water Infrastructure: Water sourcing, treatment, cooling tower optimization, and water reuse systems for data centers.
• Next-Gen Networking & Interconnect: Photonic and optical transitions, integrated light-source chips, and network observability tools.

Critical Materials As a Constraint to Industrial Power

The United States cannot achieve strategic independence without reliable access to critical materials. We expect to see material investment from both the public and private sectors in technology to modernize our mining capabilities. The mine of the future will leverage ai across the entire mining lifecycle. The big opportunity here is vertically integrated and getting to a venture scale outcome will require owning several parts of the value chain. 

We’d love to meet with founders who are thinking about: 

• Exploration and Resource Assessment: Agentic geological discovery, dynamic real time orebody models. 
• Drilling and Sampling: Autonomous rovers leveraging multimodal agents for mapping and sample collection, AI Vision ore sorting.
• Mine Planning, Permitting & Development: Compressing feasibility studies, environmental review, permitting, and mine planning to reduce timelines in project development.
• Recycling & Alternative Materials: Urban and industrial mining, advanced recycling, novel materials that substitute for scarce inputs.
• Cleaner & More Efficient Processing Technologies: Enable domestic refining of rare earths and critical minerals at scale.

The Legacy Factory Modernization Challenge

Rebuilding US industrial resilience will require a dramatic expansion of domestic manufacturing capacity. Most domestic manufacturers operate with minimal digital infrastructure and legacy equipment. They were never built to operate in today's manufacturing environment and lack the skilled labor,  capital or institutional knowledge to bring their manufacturing processes into the modern age. The idea that we can simply distribute an autonomous factory stack into this market is unrealistic. Therefore, we will either need to build ground up AI native advanced manufacturing, or leverage innovative strategies to retrofit the legacy manufacturing base.

We’d love to meet with founders who are thinking about: 

• Manufacturing Roll-Ups: Acquiring and integrating fragmented machining, fabrication, and forging shops into digitally native manufacturing platforms.
• PPA like Modernization Credits: Private credit and other investors fund the buildout, factories pay on output or uptime, credits get repackaged in asset based securities (ABS) and other asset classes. 
• Automation-as-a-Service & the Evolution of the Integrator: Evaluates capacity, proposing modernization, implements solutions, manages financing and repayment/ rev shares and monitors compliance/ traceability.
• AI Native Manufacturing Facilities: New factory models built ground-up around automation, vertical integration, and advanced processes, including cryogenic manufacturing for next generation materials, semiconductors, and quantum components.
• Financing and Insurance for Modernization Risk: Products that underwrite the risk of factory modernization projects, insure against production disruptions during upgrades, or provide performance guarantees for automation investments.
• Skilled Labor Force Multipliers: Tools that amplify the output of limited skilled machinists and technicians, capture their institutional knowledge through observation and demonstration, and use that data to train robotic systems and autonomous manufacturing processes.
• Permitting & Compliance Software for Industrial Buildout: Compress site selection, permitting, inspection, and regulatory reporting.

Full-Stack AI Disrupts Professional Services

In a world where software shifts from tools to owned workflows, and success is measured in outcomes, service providers are disrupted by full-stack, ai native service firms. Sectors that are best positioned for disruption are ones where:  (1) outcomes are well defined, (2) the work is predominantly a product of consulting or pure services (3) software has historically failed due to workflows fragmentation and reliance on unstructured data from multiple sources/ stakeholders. 

We’d love to meet with founders who are thinking about: 

• Outcome First Service Firms: AI-native companies in categories like claims and recovery, revenue cycle management, warranty administration, title insurance, construction claims, and transaction advisory.
• Outcome Based Pricing & Billing Infrastructure: Systems that support pricing, invoicing, and revenue recognition when customers pay for results.
• Insurance & Risk Infrastructure for AI Native Services: Insurance products and underwriting models that cover outcome risk when AI driven services fail (e.g. “AI lawyers” or “AI nurses”).
• Quality Assurance and Audit Trails: Systems that create verifiable records of AI decision-making for regulated services. Who made what decision, what data was used. Critical for legal/ medical/ financial services where you need to defend outcomes in court or to regulators.
• Credentialing & Compliance for Semi Autonomous Systems: Certifying models, workflows, and human supervisors to operate in regulated environments.

There’s a Need to Manage Mixed Labor Systems

AI is beginning to reorganize labor markets, firm structures, and macroeconomic systems. White collar roles are absorbing the earliest impact as AI substitutes for routine knowledge work, while the physical world economy faces severe skilled labor shortages. As AI changes the relative value of skills and organizational design, the core mechanics of employment, from education through compensation to taxation, will evolve. We are entering a period where human labor, AI labor, and hybrid labor increasingly coexist (and at times, compete), which requires new forms of governance.

We’d love to meet with founders who are thinking about:

• Human + AI Work Orchestration: Routing tasks between humans, AI agents, and hybrid workflows, markets that match skills (human or AI based) to work in real time.
• “Balance-Sheet Capitalism”: Financial platforms emerge to manage corporate capital as corporations turn into asset managers with more idle cash. They lend to one another without banks and leverage autonomous treasury solutions.

• Performance Management: Measuring output in AI heavy environments, addressing increased shadow automation, payroll alignment with ai/ human contributions.
• Shift Toward Project Based Employment: Greater reliance on contract, fractional, and hybrid human/ai labor.Analytics that separate human vs. AI contribution to outcomes. Who gets credit when a human-AI team ships something?
• The Formation of Prominent Employee Archetypes: (1) AI-supervisors (prompting, auditing, improving models) (2) human specialists (trusted for judgment, relationships, or taste).
• Regulatory and Tax Implications Software to navigate emerging regulations around AI displacement and tax treatment of AI labor costs vs. human wages. Compliance infrastructure as governments figure out how to tax AI productivity.

Energy Continues to be a Bottleneck for AI

ClimateTech is in vogue again, catalyzed by AI demand. The first part of the stack is discussed in earlier themes through evolutions of hardware/ chips as well as datacenters, however capital will also coalesce around energy generation and transmission, as the bottlenecks at the bottom of the stack cannot be mitigated via chip and data center innovation alone. 

We’d love to meet with founders who are thinking about:

• Novel Power Architectures and Alternative Energy Sources: Thermal recycling systems, micro-modular compute powered by excess flare gas at oil and gas sites, greater incorporation of wind, solar, tidal, and hybrid renewable systems.
• AI Optimized Participation in Grid Markets: Dynamically shifting energy loads based on price and supply.
• The “Vertically Integrated Developer”: Data center development “wars” require end to end developer ownership: power + land + permits + building.
• Energy & Grid Security Infrastructure: Systems that address the growing cyber and operational vulnerabilities of energy networks.
• Energy Storage as a Product: Platforms and technologies that make storage modular, financeable, and bundleable.

The Edge Stack Diverges from Cloud

The growth of edge AI is driving demand for memory that delivers high bandwidth without draining power. As AI inference moves into autonomous vehicles, robotics, industrial sensors, and AR/VR devices, memory becomes a critical bottleneck. Data center memory solutions can't meet the strict power, thermal, and size constraints these edge devices require. While SK Hynix and Micron are developing next-generation DRAM for edge applications, different use cases require different tradeoffs between speed, latency, power consumption, and cost. 

Beyond memory, interoperability across hardware platforms and robust developer tooling remain critical gaps in the edge infrastructure stack.  This fragmentation creates openings for startups to build targeted solutions, innovative architectures, and software tools for specific edge AI workloads.

We'd love to meet with founders who are thinking about:

• Memory Architectures and Processing-in-Memory: Application-specific memory systems optimized for edge workloads like vision processing, sensor fusion, or real-time control. Systems that perform computation directly within or adjacent to memory to eliminate data movement, reducing power consumption and latency.
• Memory Management and Testing for Edge Deployments: Platforms that dynamically allocate and optimize memory usage across edge device fleets. Tools and equipment that verify memory performance, reliability, and power efficiency under real-world conditions like temperature extremes, vibration, and extended duty cycles.
• Hardware-Agnostic Frameworks: SDKs, abstraction layers, and universal standards that enable model portability across different edge platforms without vendor lock-in.
• Edge Observability and Developer Tools: Real-time monitoring for model performance, hardware utilization, and thermal management. Safety and audit infrastructure for regulated or safety-critical edge deployments.

AI Infrastructure Shifts from Training to Orchestration

The investable opportunity in AI is shifting from who has the best model to who controls the software layer between models and real work. Progress in AI increasingly comes from post-training and inference system design. Companies are competing on how they route requests, verify outputs, orchestrate tools, and enforce governance rather than raw model scale. As top models converge in capability, differentiation now happens in production based on how reliably they work, what they cost to run, and how safely they can be governed.

The next generation of AI products will be agentic systems that combine models with tool execution, information retrieval, persistent memory, and human oversight. Startups win when they are provider-neutral, become a system of record for agent behavior, plug into existing enterprise controls, and deliver measurable outcomes like lower cost, higher reliability, or faster time to production.

We'd love to meet with founders who are thinking about:

• Evaluation and Audit Infrastructure for Agents: Trace what agents actually do, test workflows before deployment, score risk by business impact, manage production failures with rollbacks and emergency stops.
• Data Access Control and Governance: Granular permissions with delegated authorization, audit logs mapped to compliance requirements, automated data redaction, human approval workflows for sensitive operations.
• Cost and Latency Optimization: Route requests across models based on quality requirements, compress context and cache intermediate outputs, enforce spending limits and latency targets per workflow.
• Agent Operations Platforms: Single source of truth for what agents do in production, how they fail, and whether they meet audit and compliance standards across different model providers.

Many of the most important companies of the next decade will come from founders who understand not just the technology shift, but the downstream systems it forces to change. 

We’re investing in 80+ companies this year through our AI venture studio, accelerator, and pre-seed fund. If you are spending time in these spaces, we’d be excited to meet and explore whether there is an opportunity to work together.

‍

Forum Ventures® is the leading early-stage fund, accelerator and venture studio for B2B SaaS startups. Founded in 2014 and based in New York, San Francisco, and Toronto, we’re on a mission to make the B2B SaaS journey easier, more accessible, and successful for early-stage founders. We invest in founders at the earliest stages and work together to launch, build and scale their businesses. To date, we have made 550+ pre-seed and seed investments globally across industries like fintech, healthcare, applied AI, AI infrastructure, vertical AI, supply chain, and more.