Disrupt Stage Announced: Where Tech Breaks First

The Disrupt Stage is designed as a stress test on what comes next, a live reading of the diagnostic where the future of tech breaks first and who’s doing what to push hardest on which fault lines. The lineup for this year’s class ranges from moonshots to model scaling, autonomy to on-chain computing, cloud economics to consumer AI — a cross section that reflects the way investment, research and regulation are all converging.

A lineup made to push limits across AI, cloud, and compute

Alphabet’s moonshot chief Astro Teller joins Microsoft CTO Kevin Scott and Netflix CTO Elizabeth Stone for an unprecedented, high-res look at how frontier AI is moving out of the lab to ingest all kinds of useful data for your product. Vinod Khosla, with his signature candor on hard-tech bets, and Sequoia’s Roelof Botha dissecting the next venture cycle. They’re joined by Solana co-founder Anatoly Yakovenko on decentralized compute, Box CEO Aaron Levie on cloud-first reinvention, and operator-investor Elad Gil on the AI-capital feedback loop.

Look for fresh debuts and real-world demonstrations. Slate Auto is unveiling a completely customizable electric truck — a sign that the modular hardware space, a software standard for so long, has come to EVs. On the consumer side, Phia’s co-founders Phoebe Gates and Sophia Kianni break down what Gen Z commerce looks like when AI curates, converses and converts in one stream. Culture creation also strikes a chord with Cluely’s Roy Lee, and then NBA champion turned co-founder Tristan Thompson on how TracyAI combines performance analytics with fan engagement.

AI Leaders Distinguish Themselves on the Accelerator Circuit

Scott’s perspective ranges from developer tools to datacenter design. GitHub has stated: “AI pair-programming could represent 30–50% of code suggestions already, and the next horizon is organizational scale — policy, provenance, and cost control for model use at thousands or tens of thousands of teams.” Look for a frank discussion of inference economics and the move from custom copilots to platform-wide AI infrastructure.

Stone, who oversees one of the largest personalization engines in the world, brings hard numbers to the hype. With hundreds of millions of users and billions of daily engagements, streaming provides a perfect crucible for reinforcement learning, program discovery, and synthetic media safety. Look for reports of experiments at scale — everything from A/B frameworks to guardrails that help recommendation systems stay aligned with the trust of the audience.

And the moonshotian perch Teller is partial to is spotlighting those uncomfortable truths: around compute scarcity, around energy limitations, and about this long tail of safety work.

The Stanford AI Index has tracked the steep increase in training cost and power use for state-of-the-art models; this conversation will focus on how to bend that curve via new architectures, dedicated hardware, robotics integration, and responsible deployment.

Autonomy and Hardware Graduate From Pilot to Proof

Waymo co-CEO Tekedra Mawakana is about to clear the air on driverless readiness. Independent examinations from safety organizations like NHTSA and the Insurance Institute for Highway Safety have underlined the importance of exposure and edge cases rather than marketing milestones. Just look for a conversation on measurable safety deltas vs. human baselines, fleet telemetry, and the regulatory paths that will either speed or slow the scale-up process.

On trucks, the timing is tactical. The International Energy Agency expects EVs to jump toward a fifth of global sales, but commercial and pickup segments remain behind on total cost of ownership and charging infrastructure. A customizable electric truck sounds like a gamble on component-level innovation: battery-swappable platforms; software-defined torque profiles; and configurable beds that can be ordered up for utility work, delivery, or recreation without having to redesign the whole blasted vehicle.

Levie’s vantage point in the cloud world adds to that push. Synergy Research has been tracking cloud infrastructure spending exceeding a quarterly trillion dollars since 2018; and AI workloads are revolutionizing the bill. The tough question for any CIO — and every startup — is how to balance unit economics once inference becomes a meaningful line item. Watch for down-to-earth advice on data pipelines, retrieval augmentation and the evolving practice of AI cost governance.

Decentralized Compute and Rails V2 for Finance

Yakovenko will push past token prices to throughput, fees, and, crucially, developer experience. With advancements like local fee markets and dense on-chain data, the question is moving from speculation to scale: can decentralized systems credibly support real-time payments, games, and social apps without compromising security?

Botha and Gil will map the terrain of the capital. Venture deal volume cooled from its pandemic ebb while AI mega-rounds saw capital concentrated into a few model and infrastructure plays, according to PitchBook and NVCA analysis. The Disrupt Stage will reveal how the barbell market impacts both the formation of seed, second-time founder plays and the new “full-stack AI” thesis that goes from chips all the way through to application UX.

Founders rewriting consumer playbooks for AI-first commerce

The premillennial analog might be something like Phia’s blueprint — part of the larger shift toward social discovery, chat-native shopping and creator-fueled trust. Accenture has estimated that social commerce could exceed a trillion dollars by the middle of the decade, and the leaders are constructing lightweight funnels in which content and checkout fall into each other. Look ahead to hard lessons on attribution, authenticity and keeping hallucinations out of the cart.

Cluely’s Roy Lee provides us with the creative industry’s version of this: as generative tools commoditize production, taste, curation and rights management become defensible moats. A similar playbook is being used in sports. With league-wide tracking systems and nascent data rights, TracyAI reads the market as one where predictive insights and fan personalization are as important as box scores.

Why place still matters for innovation and local ecosystems

San Francisco’s mayor, Daniel Lurie, is likely to urge that density is destiny. The Bay Area consistently commands a significant slice of U.S. venture dollars, according to NVCA, and access to research labs, cloud providers and policymakers accelerates iteration. The takeaway: innovation compasses are local, even in a remote-first world.

What to watch onstage: metrics, safety, and real outcomes

Seek evidence, not poetic proclamations:

• Operating metrics for AI rollouts
• Safety standards for autonomy
• Concrete paths to economically viable cloud-era software
• Credible returns on pseudo-decentralized systems
• Consumer AI that sells without spamming

“Disrupt has become a bellwether exactly in the sense that it rewards specifics — and because the people on that stage are building what the rest of the industry will be reacting to next.”