Generative AI, the creator economy, and autonomous driving commercialization emerge as the core themes of Goldman Sachs' private company software and internet conference

Overview

This report summarizes Goldman Sachs' fifth Private Company Software and Internet Conference held in San Francisco on April 8-9. The report covers generative AI, the creator economy, autonomous driving, food delivery, and health and wellness, drawing on discussions among private companies, venture capitalists, and other industry participants. The overall tone is constructive: AI adoption is advancing quickly, and the industry is moving from an infrastructure capacity constraint phase into one focused on efficiency gains, cost leverage, and vertical use cases.

Core views

The core views are: first, adoption of generative AI applications is accelerating, and the industry is shifting from simply pursuing compute capacity toward improving efficiency and lowering costs; second, agentic flows may reshape enterprise software, causing traditional systems to play more of a data warehouse role while AI agents directly orchestrate business workflows; third, AI is also seen as an important tool for projecting global influence in the future, while energy and chip infrastructure remain constrained and create investment opportunities; fourth, the creator economy still depends on genuine human connection, but AI live streaming, social commerce, and fan-community tools are improving monetization efficiency; fifth, the focus across autonomous driving, food delivery, and health tech is moving from technical validation to scaled operations, asset utilization, subscription retention, and broader TAM expansion.

Analysis framework

The report uses a conference-minutes style analysis method, distilling remarks from different companies and industry participants into thematic trends and validating technology adoption, business model evolution, operating efficiency gains, and market expansion pathways through company case studies. The emphasis is not on financial modeling or valuation derivation, but on industry trends, commercialization stages, operating metrics, and future growth drivers.

Methodology notes

• conference_takeaways主题归纳法

  Theme synthesis method

  The report distills discussions from private companies, venture investors, and industry participants into five themes: generative AI, the creator economy, autonomous driving, food delivery, and health tech, in order to identify common growth drivers and constraints.

• equity_research_disclosureGS Factor Profile

  Goldman Sachs Factor Profile

  The disclosure section explains that GS Factor Profile compares stocks with the market and industry peers through growth, financial returns, valuation multiples, and a composite factor, but this conference note did not assign factor scores to any specific company.

• equity_research_disclosureM&A Rank

  Merger and acquisition probability score

  The disclosure section explains that Goldman Sachs uses an M&A rank from 1 to 3 to assess acquisition probability, where 1 indicates high probability, 2 indicates medium probability, and 3 indicates low probability; this report did not assign a specific M&A rank to any participating company.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• AI infrastructure and the chip/energy supply chain

  Benefits from generative AI adoption and expectations for large-scale infrastructure spending

  Strengths

  High demand visibility, with conference participants expecting related infrastructure spending of about $7.5 trillion by 2030.

  Weaknesses

  Energy and chip infrastructure remain constrained, which may limit near-term expansion.

  Comparison

  Compared with application-layer companies, infrastructure is closer to the underlying constraints and investment bottlenecks of AI expansion.

  Risks

  Capex cycles, the speed at which supply bottlenecks ease, and lower unit demand from efficiency improvements.

• Eleven Labs

  A representative private company in voice AI

  Strengths

  About $400 million in ARR and expansion from text-to-speech into emotion, dialect, and multimodal voice.

  Weaknesses

  Enterprise communication use cases still need continued validation for scale and differentiation.

  Comparison

  Compared with traditional voice tools, the emphasis is shifting from synthesized voice to context understanding and immersive communication.

  Risks

  Intensifying competition, model costs, copyright issues, and compliance concerns.

• RunwayML

  A representative company in AI video generation and world models

  Strengths

  Can generate frames in under 100 milliseconds and supports real-time generation adjustments.

  Weaknesses

  Expanding from media and entertainment into personal and team use cases still needs validation of paid conversion.

  Comparison

  Compared with traditional rendering workflows, RunwayML emphasizes real-time, personalized, and cross-hardware content generation.

  Risks

  Compute costs, content copyright, model quality, and the pace of enterprise customer adoption.

• Patreon

  Creator economy platform

  Strengths

  Strengthens creator-fan connection through free memberships, fan communities, and in-app media hosting.

  Weaknesses

  Historically, user growth has depended heavily on creators' own promotion, and direct-to-fan marketing by the platform is still being expanded.

  Comparison

  Compared with AI-generated content platforms, Patreon emphasizes the enduring value of human connection and paid super-fan relationships.

  Risks

  Creator supply, fan willingness to pay, platform competition, and content discovery efficiency.

• SuperOrdinary

  Social commerce and brand infrastructure platform

  Strengths

  Treats creators as scalable marketing infrastructure and uses AI live streaming to maintain 24/7 engagement.

  Weaknesses

  Dependence on the TikTok ecosystem means platform rule changes could affect growth.

  Comparison

  Shifts from traditional pay-to-play marketing to a cost structure driven by sales commissions.

  Risks

  Platform dependence, social commerce regulation, and M&A execution risk.

• May Mobility

  Representative company in autonomous driving commercialization

  Strengths

  Uses a Driver-as-a-Service model and partners with platforms such as UBER and LYFT to solve demand aggregation and fleet management.

  Weaknesses

  The industry still needs to balance expansion into new cities with high asset utilization.

  Comparison

  Compared with Waymo's modular system and Tesla's end-to-end route, May Mobility emphasizes adapting more easily to new scenarios.

  Risks

  Regulation, financing, fleet management, OEM partnerships, and unit economics.

• CookUnity

  Food delivery and chef infrastructure platform

  Strengths

  Provides chefs with supply-chain, packaging, and logistics infrastructure and is expanding into B2B enterprise customers.

  Weaknesses

  Requires continued investment in automation and a centralized kitchen network.

  Comparison

  Compared with traditional food-delivery platforms, CookUnity is more like an infrastructure layer for scaling chef operations.

  Risks

  Food quality control, logistics costs, customer retention, and returns on automation investment.

• Eight Sleep

  Sleep health hardware and wearable-plus platform

  Strengths

  Collects biometrics through Pod hardware and intervenes through temperature and height adjustments; it plans to integrate wearable devices such as Whoop and Oura.

  Weaknesses

  Expanding from hardware into predictive health analytics and clinical use cases requires stronger evidence and a clearer regulatory path.

  Comparison

  Compared with ordinary wearables, Eight Sleep emphasizes moving from monitoring to active intervention.

  Risks

  FDA approval, expansion into medical use cases, hardware costs, consumer affordability, and data privacy.

Key data

• Conference dates and locationApril 8-9, 2026, San FranciscoThe fifth Goldman Sachs Private Company Software & Internet conference.
• AI infrastructure spending expectationApproximately $7.5 trillion by 2030Meeting participants viewed AI infrastructure investment as enormous, with energy and chip infrastructure still constrained.
• Eleven Labs ARRAbout $400 million in ARRThe company is expanding from text-to-speech into emotion, dialect, and multimodal voice applications.
• SuperOrdinary operating streamlining exampleThe number of people needed to run a beauty brand has fallen from about 25 to roughly 3-5This reflects the potential impact of AI on brand operating efficiency and organizational structure.
• RunwayML content generation viewMore than 50% of the content users see over the next five years may be AI-generatedThe company believes video generation and personalized content will expand from media and entertainment into many hardware interfaces.
• RunwayML generation speedThe world model can generate frames in under 100 millisecondsThis supports real-time adjustment during generation rather than pre-rendering all frames.
• CookUnity Unity Pass penetrationAbout 15% of customers have joined the programThe subscription plan uses health and personalization to increase customer frequency and AOV.
• Goldman Sachs global coverageAs of April 1, 2026, Goldman Sachs has investment ratings on 3,074 stocksFrom the disclosure section, this is background information on Goldman Sachs' global equity research coverage.

Impact & implications

The report suggests that software and internet investment themes are shifting from point solutions toward comprehensive competition around infrastructure, automation, and operating leverage. AI infrastructure, voice and video generation, social commerce tools, autonomous driving operating systems, restaurant supply-chain infrastructure, and wearable health intervention devices could all benefit. At the same time, monetization capability, asset utilization, energy and chip supply, regulatory approval, and user retention will become the key variables determining whether long-term value is ultimately realized.

Risks

• Constraints in energy and chip infrastructure may limit the expansion of AI applications and model capability.
• Competition in generative AI applications is intense, and model costs, copyright, compliance, and differentiation capability may affect profit quality.
• Agentic flows reshaping enterprise software are still early, and customer adoption, data governance, and workflow reliability remain unproven.
• Autonomous driving commercialization is constrained by regulation, financing, OEM partnerships, fleet management, and asset utilization.
• The creator economy and social commerce depend on platform traffic, user willingness to pay, and the extent to which AI content substitutes for real human connection.
• Expansion of health tech into clinical use cases may face FDA approval, data privacy, and medical liability risks.

What to watch

• The pace at which AI infrastructure investment is realized, and whether constraints in energy, chips, and compute supply ease.
• Whether companies move from pilot AI applications to large-scale, top-down deployment.
• The extent to which agentic workflows replace traditional enterprise software, dashboards, and record systems.
• Paid conversion for AI voice, AI video, and AI live streaming in enterprise communications, media production, and social commerce.
• Whether autonomous driving companies can maintain high asset utilization and healthy unit economics as they expand into new cities.
• CookUnity's Unity Pass penetration, customer frequency, AOV, and B2B expansion performance.
• Eight Sleep's FDA approval progress, expansion of marketing to women and non-athlete consumers, and deployment in hospital and elder-care settings.