Helion Venture Partners

TechCrunch Disrupt 2026 Adds Six Themed Stages

TechCrunch announced the agenda structure for TechCrunch Disrupt 2026, scheduled for October 13–15 at Moscone West in San Francisco. The conference will gather 10,000+ founders, investors and operators for 200–250+ sessions organized across six themed stages: Disrupt Stage, Builders Stage, Smart Money Stage, Smart Systems Stage, AI in the Real World Stage, and the AI Stage (presented by Google Cloud). Programming spans topics from fundraising and go‑to‑market tactics to fintech, physical infrastructure for AI, and production-grade AI deployments. Notable speakers and panelists listed include partners and CEOs from Index Ventures, Sapphire Ventures, Gusto, Gamma, Google, CapitalG, Airwallex, Emergence Capital, Inertia and Helion. Early bird tickets are available with savings up to $410 and 50% off a second ticket.

Race to Power the Grid by 2035 Widens

Tech companies' growing AI compute needs are driving investment and commercial deals across fusion, small modular fission, natural gas, and long-duration battery storage as contenders to supply reliable 24/7 power by 2035. Natural gas remains the current baseload choice but faces supply-chain vulnerabilities and multi-year turbine backlogs. Several small modular reactor (SMR) firms and fusion startups aim to connect commercial plants to grids in the early 2030s; notable projects include Kairos Power's Hermes 2 demo, Oklo's 2028 target, Commonwealth Fusion Systems' demo and 400 MW Arc, and Helion's aggressive plan to deliver electricity to Microsoft by 2028 and to supply OpenAI at multi-gigawatt scale by 2030–2035. Cost and scale remain key hurdles: current levelized costs for new nuclear and fusion are high versus gas and renewables paired with batteries, while new long-duration storage (iron‑air, organic fluid designs) could undercut alternatives.

Meta Moves Toward Tokenized, Agent‑Driven Work

The newsletter reports that Meta is building a personal AI “CEO agent” led by Mark Zuckerberg that can pull operational data across the company and that AI usage is being tied to employee evaluations. The piece argues that agent-based measurement could make white‑collar output traceable and priced in tokens, exposing roles that are hard to quantify today. The issue also summarizes related AI signals: a USC paper finding that “expert” persona prompting can reduce factual accuracy on benchmarks, OpenAI’s request to UK regulators to include ChatGPT on Android/Chrome choice screens, and LinkedIn’s inconsistent enforcement toward visibly AI‑run accounts. Additional briefs note Anthropic’s Claude gaining desktop‑control features, Jensen Huang’s AGI comments, funding and M&A activity (Littlebird, Lovable), and enterprise/security moves by Cisco and Google Cloud.