A permanent capital vehicle that combines venture building, AI infrastructure, and compute partnerships to create and scale companies at unprecedented speed and efficiency.
We are entering an era where systematic discovery and innovation will be powered by compute at scale. The companies that harness this capability will win. Those that do not will be disrupted.
The history of value creation follows a clear pattern: when capital and compute converge on a domain, the economics of that domain change permanently. In pharmaceuticals, computational drug discovery reduced the average cost of bringing a new molecule to Phase I from $2.6B over 12 years to under $400M in under 4 years. In materials science, companies like Citrine Informatics use machine learning to compress decades of materials development into months. In financial markets, quantitative strategies now account for over 35% of total equity trading volume.
The same transformation is now arriving in entrepreneurship itself. The tasks required to build a company, including market research, competitive analysis, product design, brand development, legal formation, financial modeling, customer acquisition, and operational management, are increasingly executable by AI agents. This is not a marginal improvement. It is a structural shift in the economics of company creation.
The thesis: What happened in drug discovery and materials science when compute and capital combined to unlock value at scale is now happening in entrepreneurial innovation. Share Holdings is the permanent capital infrastructure to capture this transformation.
Traditional venture capital deploys $500K to $5M per company, funds 20 to 40 companies per portfolio, and expects 1 to 3 winners to return the fund. The autonomous creation model inverts these economics entirely. By collapsing the cost and time of testing a venture hypothesis toward the marginal cost of compute, it becomes possible to run orders of magnitude more experiments per year, dramatically increasing the surface area for discovering breakout outcomes while maintaining concentrated ownership in the ones that work.
Share Ventures operates through two complementary vehicles, backed by a permanent capital holding company. This structure enables both traditional venture returns and long-term compounding through owned companies.
The dual vehicle approach is inspired by the most successful capital allocators in history. Berkshire Hathaway combines operating businesses with a public equity portfolio. Pioneering permanent capital vehicles hold high-performing companies indefinitely rather than forcing exits on a traditional fund timeline. Share Holdings applies this same permanent capital philosophy to autonomously created companies, with the added advantage that the creation cost per company is orders of magnitude lower.
Venture lab and fund unlocking human potential
Venture capital fund investing in external companies across human performance verticals
Venture lab that builds companies from scratch using ShareOS and compute partnerships
Operating subsidiary that runs the innovation engine, builds and launches new ventures
Permanent capital vehicle for long-term ownership of foundry-born companies. No forced exits. Compounding value indefinitely.
The Fund captures returns from external investments with traditional VC economics. The Foundry creates companies from scratch at dramatically lower cost, with the Holdings vehicle enabling permanent ownership and long-term compounding. Fund investments also generate strategic insights that inform Foundry creation decisions.
Investors benefit from two distinct return profiles: venture-style exits from the Fund and long-term cash-flowing assets from Holdings. Both are powered by the same platform, team, and compute partnerships, creating natural synergy between the vehicles.
Traditional VC funds operate on 10-year timelines with pressure to exit positions for LP distributions. This forces premature exits and misaligns incentives. Share Holdings, as a permanent capital vehicle, can hold winning companies indefinitely, allowing compounding to work over decades rather than being constrained by fund lifecycle. When a foundry-born company reaches sustainable cash flow, it transfers to Holdings where it can compound value without exit pressure.
The advantage: Companies built in the Foundry benefit from shared infrastructure, agent orchestration, and operational playbooks from day one. This dramatically reduces the cost and time to build a company, improving returns for all partners. Winners compound inside Holdings indefinitely.
Large companies have distribution, capital, and customers but struggle to discover new value. Their moats are eroding. They need external innovation engines. This is true across every industry.
McKinsey's 2025 State of Innovation report found that 84% of Fortune 500 executives consider innovation "critical to growth," yet only 6% are satisfied with their innovation performance. The reason is structural: large organizations optimize for execution, not discovery. Their incentive systems, governance structures, and risk tolerance all work against the kind of rapid experimentation needed to find new value.
Meanwhile, the cost of building a new company has dropped by over 90% in the last decade. Cloud infrastructure replaced server rooms. Open-source software replaced licensed toolkits. And now, AI agents are replacing the teams of specialists required for market research, product development, brand creation, and go-to-market execution. The result is a new category of opportunity: systematically building companies using compute rather than headcount.
Have distribution but cannot discover. Traditional moats are eroding. Need external innovation capability to remain competitive.
Can be created systematically when discovery is powered by compute at scale and codified operational playbooks.
Share Holdings serves both: It is the discovery and innovation engine for companies that need to remain competitive, and the infrastructure to create entirely new companies from scratch. The same platform, the same agents, the same playbooks, applied in two modes.
The percentage of company-building tasks that can be automated by AI has followed an exponential curve since 2020. We are approaching the inflection point where autonomous company creation becomes not just possible, but economically superior.
In 2020, AI could handle roughly 5% of company-building tasks: basic text generation, simple data analysis, and template-based outputs. By 2023, with the arrival of GPT-4 and Claude, that figure jumped to approximately 25%, covering market research, competitive analysis, basic product design, financial modeling, and content creation. Today in 2026, with agentic AI systems orchestrating multi-step workflows, we estimate 45-50% of company-building tasks can be fully automated, with another 20% significantly augmented.
The trajectory is clear. By 2028, we project 65-70% automation. By 2030, over 80%. The companies that build infrastructure to capitalize on this trend now will own the most valuable capability of the next decade.
The inflection point is now. Between 2025 and 2027, AI transitions from augmenting human workers to autonomously executing complete workflows. The organizations that build infrastructure to harness this shift will define the next era of company creation. Those that wait will face a structural disadvantage that only widens with time.
ShareOS is the autonomous company creation platform that powers everything. It orchestrates 130+ AI agents, manages company lifecycles, and turns compute into real businesses.
Think of ShareOS as the operating system for building and running companies. Just as iOS provides the infrastructure layer that makes it possible for millions of apps to exist, ShareOS provides the infrastructure that makes it possible to create, launch, and operate companies at scale. Every company built on ShareOS inherits authentication, payment processing, database management, deployment pipelines, monitoring, and agent orchestration from day one.
Every company, whether created de novo or onboarded as an existing business, is managed through seven execution workstreams. Each workstream has dedicated agent groups, KPI tracking, and decision gates that ensure systematic progress.
36 agent groups across 7 workstreams coordinate autonomously. Agents discover opportunities, validate hypotheses, build products, and run operations around the clock. Each agent group has a parent agent managing sub-agents with specific KPI responsibilities.
De Novo: Create companies from scratch using compute-powered discovery. Existing Company: Onboard any company, detect gaps across workstreams, and fill them dynamically with agents.
Pre-built architecture for rapid company setup: authentication, payments, databases, APIs, deployment pipelines, and monitoring. A new company goes from concept to live product in days, not months.
Every dollar of value is tracked from company level through workstreams, goals, milestones, and tasks. Performance and execution scores derive from real dollar valuations, not arbitrary point systems.
The vehicle is built on strategic partnerships with the world's leading providers of AI compute and capability. Each partner contributes unique infrastructure and expertise that no single provider could offer alone.
These are not vendor relationships. They are strategic partnerships where both sides contribute meaningful value. Compute partners gain a high-volume deployment channel that showcases their technology in real production environments. We gain infrastructure, engineering talent, early access, and distribution that would cost tens of millions to build independently.
TPUs, Gemini, Cloud
GPUs, CUDA, DGX
Azure, Copilot
GPT, Reasoning
Claude, Safety
TPU/GPU clusters for training and inference, dedicated capacity for high-throughput experimentation across all portfolio companies.
ML engineers and researchers embedded in our workflows, architecture reviews, and production optimization.
Preview access to unreleased models, fine-tuning capabilities, and new features before general availability.
Co-sell motions, marketplace listings, and integration partnerships for go-to-market acceleration.
Model-agnostic infrastructure with automatic failover, vector storage, orchestration, and real-time agent coordination at production scale.
Pre-built architecture for rapid company setup: auth, payments, databases, APIs, and deployment pipelines.
Codified workflows across 7 execution workstreams with decision gates and instrumentation.
Deep connectivity across human performance domains: longevity, cognitive enhancement, organizational health, wellness, and connected care.
The vehicle systematically discovers white space, identifies problems worth solving, rapidly builds solutions, and leverages partner distribution to scale. Every step is instrumented, measured, and optimized.
Every company moves through a defined lifecycle with clear gates and metrics at each stage. This is not a loose framework. It is a disciplined, repeatable process with quantitative thresholds for advancement.
AI agents scan seven performance verticals for unmet needs, generate venture concepts, validate through landing pages and smoke tests, build MVPs, and launch. The entire process from discovery to live product can happen in days. Human judgment enters at key decision gates: concept approval, pilot commitment, and scale decisions.
Onboard any company onto ShareOS. Agents diagnose gaps across all seven workstreams, deploy targeted solutions, and continuously optimize operations. This mode serves both portfolio companies and external partners who need innovation capability without building it internally.
When the cost and time of testing a venture hypothesis collapse toward the marginal cost of compute, the math of autonomous company creation becomes transformative compared to the $500K+ cost of launching a traditional company. Volume, speed, and signal quality replace capital intensity and guesswork.
Traditional venture capital is a high-cost, low-volume game. A typical fund invests $2M to $5M into each of 25 to 40 companies, hoping that 2 to 3 become breakout successes. The failure rate is 60-70%, and the capital at risk per attempt is enormous. The autonomous creation model inverts this entirely.
By running experiments at compute cost rather than headcount cost, we can test 1,000 venture hypotheses per year. Each experiment includes market validation, competitive analysis, landing page creation, smoke testing, and initial signal measurement. The ones that show traction advance. The ones that don't are discarded at minimal cost.
$2M - $5M
For a full year of experiments yielding a handful of exit-ready companies. Compare this to a traditional fund deploying $50M to $100M for 25 to 40 portfolio companies.
The funnel is self-correcting. Capital only flows to experiments that demonstrate traction. 95% of spend is concentrated in the top 50 performers. Failed experiments cost nearly nothing. This is venture capital with the risk profile of systematic R&D.
The math is simple: a traditional fund spends tens of millions to test 25 to 40 ideas. The autonomous model tests orders of magnitude more hypotheses for the same capital, because the cost of a failed experiment collapses toward the cost of the compute that ran it. The expected value per dollar deployed is structurally superior.
Autonomous company creation fundamentally changes the math of portfolio construction. Higher ownership, lower cost per company, and permanent capital create a return profile that traditional venture cannot match.
When you build a company from scratch, you own it entirely at inception. Even after bringing in leadership, advisors, and early employees, the holding entity retains 30-60% ownership. Compare this to a traditional VC fund that acquires 15-20% ownership by investing $2-5M in a Series A round. The economics are fundamentally different.
| Metric | Traditional VC Fund | Autonomous Creation |
|---|---|---|
| Fund Size | $100M | $200M |
| Experiments / Year | 8-12 investments | 1,000+ experiments |
| Cost Per Attempt | $2M - $5M | Marginal compute cost |
| Ownership at Entry | 15-20% | 30-60% |
| Time to First Signal | 12-18 months | Days to weeks |
| Portfolio Concentration | 25-40 companies | 50 cash-flowing + 15 IP-qualifying |
| Decision Data per Company | Qualitative (pitch, references) | Quantitative (508 KPIs per company) |
| Exit Pressure | 10-year fund lifecycle | Permanent capital, no forced exits |
Based on conservative assumptions about funnel conversion rates and average company valuations at each stage, the model generates compelling portfolio-level outcomes.
$150M
Portfolio value across 50+ cash-flowing companies and IP assets
$500M
Compounding returns from permanent holdings plus new vintage creation
$2B+
Multi-generational portfolio with compounding cash flows and strategic exits
Conservative assumptions: These projections assume a 0.5% experiment-to-scale conversion rate, an average enterprise value of roughly $24M at scale, and no outlier outcomes. On a $200M vehicle, the base case models an 8.5x gross MOIC over ten years, with the upside case reaching 19.8x. A single breakout company (comparable to early-stage investments in companies like Sensate or QOVES) would significantly exceed these targets.
Three structural advantages combine to create a return profile that traditional venture capital cannot replicate: lower cost per experiment, higher ownership at entry, and faster iteration cycles.
10,000x
Radically lower initial cost per experiment, at compute cost rather than the $500K+ of a traditional launch. Capital is preserved for companies that demonstrate traction, not consumed by early-stage uncertainty.
2-4x
Higher ownership per company (30-60% vs. 15-20%). Every dollar of enterprise value created flows to the holding entity at 2-4x the rate of a traditional VC investment.
50x
Faster iteration from hypothesis to signal (days vs. 12-18 months). Failed experiments are identified and discarded in hours, not years.
| Scenario | Traditional VC Fund | Autonomous Creation |
|---|---|---|
| Capital Deployed | $100M | $100M |
| Companies Funded / Created | 30 | 25,000+ experiments, 125 scaled |
| Avg. Ownership | 18% | 45% |
| Winners (assume 5%) | 1-2 companies | 5-6 companies per vintage year |
| Avg. Winner Valuation | $200M | $50M (lower bar needed) |
| Value of Winners (ownership-adj.) | $36-72M | $112-135M |
| Portfolio Cash Flow (annual, yr 5) | Minimal (growth stage) | $15-30M from 50+ companies |
| Gross MOIC (10-year) | 2.5-3.5x (top quartile) | 8.5x base case, up to 19.8x (modeled) |
The compounding effect: Unlike a traditional fund that deploys capital once and waits for exits, the autonomous model generates cash flow from early-stage companies that funds the next generation of experiments. By year 3, the system becomes partially self-funding. By year 5, portfolio cash flows can exceed new capital deployment, creating a compounding engine that accelerates over time.
This model is built on a structural inflection point: the cost of creating a company has collapsed 10,000x in the last decade. What once required $500K to $5M in capital, time, and human labor to test a single venture hypothesis now requires little more than the compute to run it. That is not an incremental improvement. It changes the entire equation of venture creation. When failure is nearly free, you can run orders of magnitude more experiments. When ownership is established at creation rather than acquired at premium, you retain structurally higher stakes. And when a permanent capital vehicle holds the outcomes indefinitely, returns compound without the drag of forced exits. Share Holdings is the entity built to exploit this inflection point at scale.
Every deployment generates data that improves the system. Research outputs, experimental data, workflow traces, and commercialization learnings compound over time. This is not a theoretical moat. It is a measurable, quantifiable advantage that widens with every experiment.
After 1,000 experiments, the system has learned which markets respond to which messages, which product architectures scale, which pricing models convert, and which operational playbooks work for which company types. After 10,000 experiments, this dataset becomes the most valuable asset in the portfolio, more valuable than any individual company, because it makes every future company more likely to succeed.
Data moat: Unlike traditional venture, the vehicle generates proprietary data that improves the system. More deployments produce better models, which increase throughput, which create more deployments. The advantage compounds exponentially, and it cannot be replicated without running the experiments.
Share Ventures focuses on the seven domains that define human performance. Every company we build or invest in advances one or more of these verticals. This focus creates deep domain expertise that compounds across the portfolio.
Strength, endurance, recovery
Attention, learning, memory
Regulation, resilience, stability
Trust, culture, cohesion
Metabolic health, immune, sleep
Execution, operating models
Cash flow, allocation, wealth
Cross-vertical synergy: Companies in the portfolio share data, distribution, and infrastructure. A biological health company feeds insights to cognitive and emotional wellness companies. An organizational intelligence platform informs financial and social tools. The whole becomes greater than the sum of its parts.
Every company Share Holdings creates or acquires produces value on two distinct tracks. We hold them separate, we never collapse them into one number, and we never use a bare "impact" label for money. Two tracks, always labeled.
The dollar contribution to enterprise value, what an investor would pay for the company: revenue, margin, moat, and strategic de-risking. This is the return the vehicle underwrites.
The dollar contribution to societal and stakeholder value: health gained, harm avoided, and productivity and access created for people outside the firm. Valued with the same discipline as enterprise value, never folded into it.
Social value gets the same rigor we apply to enterprise valuation: an explicit framework, a defined population, a per-unit benefit tied to a cited benchmark, and a stated sensitivity range. No vibes. The three frameworks we use most:
QALYs gained × $50K to $150K per QALY, plus disease burden reduced through avoided healthcare cost.
Hours saved for people outside the company × the value of their time.
Populations reached × the value of gaining access to a service they previously lacked.
Social Valuation Impact = population affected × per-person benefit × timeframe. Worked example, Feno: at 100,000 active users, roughly 15% with diabetes or pre-diabetes, at $2,500 average avoided cost, the modeled annual Social Valuation Impact is about $37M, against a target enterprise Valuation Impact of $82.1M. That figure is deliberately conservative: it counts only the diabetic-cohort medical offset and excludes cardiovascular, pregnancy, and productivity gains. Thirty percent fewer users takes it to $26M; thirty percent more to $48M.
Because our seven domains are chosen where the innovation gap and the human stakes are both highest, the two tracks tend to move together. We go after the largest markets we believe can also be the most consequential, and we report both kinds of value on their own terms.
Four converging factors make this the right moment to deploy capital into autonomous company creation:
Foundation models can now handle complex discovery, validation, and execution tasks. Agentic AI makes autonomous multi-step workflows possible at production scale. This was not true even 18 months ago.
Large companies across every industry need external innovation engines. Traditional defensibilities, including brand, distribution, and regulatory capture, are less durable in an AI-native world. The market for innovation-as-a-service is emerging.
Cloud compute, open-source models, and deployment pipelines have reached the point where building a company is 10x cheaper and faster than five years ago. The infrastructure to run thousands of experiments per year exists today.
LPs are increasingly drawn to permanent capital structures that avoid forced exits. The success of leading permanent capital funds and similar vehicles has validated the model. Share Holdings applies it to a new asset class.
The parallel: Drug discovery and materials science were transformed when compute and capital combined to run experiments at scale. The same transformation is now possible for entrepreneurial innovation. The infrastructure exists. The partnerships are forming. The first movers who build this capability now will own the most defensible position in venture for the next decade.
The vehicle is designed for institutional partners who bring capital, distribution, and domain expertise. Multiple paths for strategic alignment exist across every industry vertical.
Exposure to a new asset class: systematic, compute-powered company creation. Diversified across hundreds of experiments per year with structural ownership advantages (30-60% entry ownership vs. 15-20% in traditional VC).
An external innovation engine that discovers new products, validates markets, and builds companies in domains where their existing businesses need to evolve. The engine runs 24/7 without organizational politics or legacy constraints.
Strategic AI capability paired with an ownership-aligned permanent capital vehicle to transform real businesses through applied deployment.
The Vista Equity Partners model is also instructive. Vista demonstrated that applying consistent, technology-driven operational playbooks to software companies could produce top-decile returns at scale. Their Perennial fund pioneered permanent capital in the PE context. Share Holdings combines this insight with the cost structure of autonomous creation: the playbooks are not applied after acquisition at $500M+ enterprise value, but embedded from inception at near-zero marginal cost.
Share Ventures has an active portfolio spanning multiple human performance verticals. Each company benefits from shared infrastructure, compute partnerships, and the ShareOS platform.
Connected oral health platform. Smart toothbrush with AI-powered health insights. $1M+ monthly revenue, 30% MoM growth, selling on Amazon and direct.
Organizational intelligence platform. AI system of record for culture and team dynamics. New CEO (ex-Patagonia CPO). Pursuing $30M revenue acquisition target.
AI facial analytics. $1.2M monthly revenue, profitable, 40M monthly uniques. 65% male users. Growing virally on social platforms.
AI performance coaching via text, voice, and calendar integration. Multi-modal coaching powered by behavioral science and chronobiology.
Financial performance. Autonomous financial intelligence and wealth creation tools. Launch stage.
Social performance. Social trust and community cohesion platform. Launch stage.
Biological performance. Connected care and health data intelligence. Early stage.
External fund investments include companies like Sensate (vagus nerve stimulation), QOVES (facial aesthetics AI), Jocasta Neuroscience, and others across the human performance landscape. These investments generate returns and create strategic synergies with Foundry-built companies, validating market hypotheses and providing distribution channels.
We have built the foundation. The platform is live. The agents are running. The portfolio is generating revenue. Now we are assembling the partnerships to execute at scale.
We welcome the opportunity to share more detail, walk through the platform live, and discuss how we can build together.