Guest: Vipul Ved Prakash, CEO & Co-founder, Together AI — together.ai
Host: Sumant Mandal, Founder & Managing Partner, March Capital

Vipul Ved Prakash has spent two decades building systems that run at internet scale. One was a spam filter that still screens about a fifth of the world's email. Another was a real-time social search engine that Apple bought in 2014. When he co-founded Together AI in 2022, just before ChatGPT launched, he bet that AI would need an open ecosystem as much as it needed frontier labs. That bet now looks like the consensus view.

In this episode, March Capital's Sumant Mandal sits down with Vipul to trace his path from Cloudmark and Topsy to Apple's Siri group, and then to Together. Together is a platform that makes open-source models easy, cheap, and safe to run in production.

Tune in as Vipul and Sumant get into what he calls this year's "quiet revolution" in open models, why infrastructure must keep pace as tokens become more valuable, and what he expects the future to look like as open models match frontier intelligence.

***

Together AI is the AI Native Cloud for inference, compute and model shaping. AI natives and enterprise use Together AI's full stack AI platform for every step in their AI development journey.

The Montgomery Summit returns to Santa Monica on March 9-10, 2027.

***

00:00 Introduction to Together.ai and Vipul Ved Prakash
02:58 The Journey of AI: From Spam Filters to Foundational Models
05:52 Building an Open Ecosystem for AI
08:48 The Economics of Open Source vs. Closed Models
12:11 Regulatory Challenges and Industry Pushback
14:58 The Future of LLMs and AI Models
18:01 Infrastructure Challenges in AI Development
20:56 Global Perspectives on AI Infrastructure
23:53 Governance and Regulation in AI
27:03 The Vision for Together.ai
30:12 Final Thoughts and Future Directions

Evan Smith (Co-Founder & CEO, Altana) in conversation with Robert Silvers (Partner & Co-Chair, National Security Practice, Ropes & Gray)

Smith and Silvers — former DHS Under Secretary and co-author of the Uyghur Forced Labor Prevention Act — discuss how geopolitical fragmentation, from US-China trade restrictions to EU carbon border adjustments, is forcing companies to map multi-tier supply chain networks for the first time, since CFOs can't forecast tariff impacts without knowing their full value chain.

The most striking disclosure: a Chinese logistics middleware called "Loginc," embedded in global port and logistics systems, reportedly gives Beijing visibility into roughly 90% of global trade — part of a broader strategy that includes giving away free border management software to nations under the Belt and Road Initiative. Altana's federated AI architecture, already used by U.S. Customs and Border Protection, offers an alternative: mapping global trade connections without centralizing sensitive data, with EU customs authorities reportedly calling it "light years ahead" of alternatives.

Umesh Sachdev (Co-Founder & CEO, Uniphore) in conversation with Gajen Kandiah (CEO, Rackspace) and Sumant Mandal (Co-Founder & Managing Partner, March Capital)

Kandiah — six months into his tenure as Rackspace CEO after serving as Chief AI Officer at Hitachi — and Sachdev discuss how geopolitical risk has turned AI infrastructure sovereignty from a data-control preference into a business continuity imperative, as non-US enterprises ask how they'd keep operating if trade policy cut off access to Silicon Valley AI platforms.

Uniphore serves 2,500 enterprise customers, including 200+ Fortune 500 companies, with a "sovereign AI stack" that runs across any cloud, on-prem environment, LLM, or GPU type. The core thesis: proprietary data and fine-tuned models — not the underlying LLMs, which every enterprise can access — are the real source of differentiation, and owning the full stack is what makes that differentiation defensible.

Grant Lee (Co-Founder & CEO, Gamma) in conversation with Jamie Montgomery (Co-Founder & Managing Partner, March Capital)

Lee traces Gamma's founding story — started in mid-2020 in London — through three product phases, culminating in a March 2023 relaunch that fixed the product's first-30-seconds experience and triggered explosive, entirely organic growth driven by creators sharing the tool with their audiences.

The relaunch took Gamma from months of effort to reach 60,000 users to tens of thousands of new signups a day within a week. The company has kept its original 12 employees since launch, operates with fewer than 70 people total under a "player-coach" model with no traditional management layer, and Lee credits the growth to personally onboarding creators one by one rather than paid acquisition.

Ram Palaniappan (Founder & CEO, EarnIn) in conversation with Madison McIlwain and Claudia Laurie (Hosts, The Room Podcast)

Palaniappan argues payroll is the last major financial system still running on a multi-day batch settlement lag, and traces EarnIn's origin from a manual practice he ran at RushCard into a consumer fintech platform that now serves roughly 1% of all U.S. paycheck recipients, with a roadmap toward continuous, streaming pay.

A University of Oregon study of 1.5 million EarnIn users found incomes rose 11.5% (about $335/month) with on-demand pay access — gains at twice the rate of inflation — while EarnIn's LivePay product has already surpassed one million transactions since its late-2025 launch. Palaniappan also describes a deliberately asymmetric capital strategy: buying hospital medical debt in bulk at a discount and forgiving it, writing off nearly $10 million on behalf of customers.

Ketan Karkhanis (CEO, ThoughtSpot) in conversation with Greg Emerson (Managing Director & Senior Partner, Boston Consulting Group)

Karkhanis describes the strategic pivot he made within 17 months of becoming CEO — shutting down business units that no longer fit, rebuilding engineering culture around AI-first principles, and reorienting sales and go-to-market entirely — framed against BCG's estimate of a $3 trillion agentic AI opportunity, roughly double today's B2B software market.

ThoughtSpot has crossed $140 million ARR with 33 customers above $1 million ARR and is projecting $180 million ARR at 48% year-over-year growth by fiscal year-end, with 60–70% growth targeted the following year. Karkhanis is candid that the company hasn't fully solved the shift away from seat-based pricing but is betting on unlimited-usage contracts as the model that captures the most value.

Carina Hong (CEO, Axiom Math) in conversation with Matt McIlwain (Managing Director, Madrona Venture Group)

Hong describes Axiom's mission to build the world's first AI mathematician — a system that doesn't just solve advanced math problems but formally proves mathematical conjectures and verifies logical correctness without requiring human expertise to audit the result.

The headline proof point: Axiom Prover scored a perfect 120/120 on the Putnam Exam (only 5 humans have done that in 98 years, and the best LLM scored 103/120), then went on to resolve four open mathematical conjectures that no human mathematician had solved. Hong connects the breakthrough to a bigger commercial opportunity in chip verification, code verification, and building trust into agentic AI as it deploys at scale.

Wardah Inam (CEO, Overjet) in conversation with Quentin Hardy (Principal, LGTM Communication)

Inam explains how Overjet applies AI across the full dental treatment workflow — diagnosis, patient communication, insurance claims — for a half-trillion-dollar global market where treatment acceptance rates run as low as 50%, founded after she noticed wide diagnostic variation between dentists.

The operational case is stark: dentists spend 5+ hours a week documenting patient communications, hygienists spend 15+ minutes per patient on periodontal charting, and front-desk staff spend 20 hours a week verifying insurance — all of which Overjet automates or compresses, including a 3x cut in charting time. The company now serves thousands of practices and 60,000+ users, processes claims for 115 million+ insured members, and is eyeing expansion into adjacent outpatient verticals like veterinary and dermatology.

Pankaj Patel (Co-Founder & CEO, Nile) in conversation with Quentin Hardy (Principal, LGTM Communication Services)

Patel, a 22-year Cisco veteran, lays out Nile's cloud-native, zero-trust network-as-a-service platform designed to replace legacy enterprise LAN infrastructure with a fully autonomous, security-first architecture — citing a 2025 Google study tying 50% of zero-day vulnerabilities to legacy vendors, specifically Cisco and Fortinet.

Proof points include standing up an entire network for 200,000 daily visitors at a 2-million-square-foot Saudi Arabia conference with just 4 people in 3 days, a 4.9/5.0 Gartner Voice of Customer score, and land-and-expand accounts growing from $50K to $1.4M+ ACV. Nile has raised $300 million to date and is targeting 300–500% growth in ARR and TCV in 2025.

Vipul Ved Prakash (Co-Founder & CEO, Together.ai) in conversation with Sumant Mandal (Co-Founder & Managing Partner, March Capital)

Prakash, a former Apple engineer who worked on Face ID and Siri, argues that generative AI represents a new computational substrate, not just an advancement in language processing. He positions Together.ai as the cloud layer for the emerging AI workload era, running the top open-source models across modalities on Grace Blackwell rack-scale systems.

The conversation frames inference efficiency at scale as the defining competitive moat of the AI era, comparable to what AWS did for general-purpose cloud compute, with Together.ai serving both frontier labs and enterprises adapting models to proprietary data.

Ryan Meadows (Chief Revenue Officer, Lovable) in conversation with Seth Siegler (Chief Innovation Officer, eXp Realty)

Siegler explains how eXp Realty, a publicly traded brokerage with 83,000 agents across 29 countries and $4 billion in revenue, deployed Lovable's vibe-coding platform across its non-technical workforce to replace commercial SaaS products and ship custom internal software, governed by a small team of internally promoted "vibe coders."

The results include international operations NPS jumping from the 20s into the 80s, millions saved in SaaS spending, and internal tools built to replace products including Meta Workplace and Slack. CEO Glenn Sanford personally built 70% of the company's internal social network himself. The session closes with a larger idea: the next step is not incremental software replacement, but full organizational reinvention around fewer departments and fewer handoffs.

Lochmiller traces Crusoe's evolution from flare-gas-powered Bitcoin mining to a vertically integrated AI infrastructure provider, covering land and power development, data center construction, GPU cluster deployment, and managed inference. The company is building gigawatt-scale campuses for OpenAI and Oracle under the Stargate initiative.

The centerpiece is Crusoe's Abilene, Texas campus: 1.2 gigawatts of power capacity, equivalent to the city of Denver, financed as a $15 billion project with JP Morgan and Blue Owl. Crusoe won the original OpenAI contract against 32 competitors by delivering a 100-megawatt facility in 11 months versus a competing 2.5-year bid. Lochmiller argues that AI data centers sized for peak demand will eventually sell excess power back to the grid, even as U.S. data center electricity use is projected to triple by the end of the decade.

Zac Bookman (Co-Founder & CEO, OpenGov) and Garrett Langley (Founder & CEO, Flock Safety) in conversation with Wes Nichols (Partner, March Capital)

Bookman and Langley describe how AI is transforming public safety and government operations. Flock Safety's camera network and LLM-powered 911 analysis helped apprehend roughly a million criminals last year, about 20% of all U.S. crimes solved, while OpenGov, competing with Tyler Technologies, is overhauling San Francisco's permitting system and helping Los Angeles prepare for a carless 2028 Olympics.

Both CEOs frame acute government staffing shortages as the real driver of rapid AI adoption. On privacy, Langley argues AI should augment rather than replace human judgment, with humans initiating and concluding decisions while AI runs the analytical middle.

Steve Westly (Founder & Managing Partner, The Westly Group) in conversation with Daniel Weiss (Co-Founder & Managing Partner, Angeleno Group).

Westly and Weiss, both longtime clean energy investors, argue that the AI and data center buildout represents the largest energy investment opportunity in history. They note that Anthropic and OpenAI, two of the fastest-growing companies ever by user adoption, already consume more energy than any manufacturing firm.

Weiss's key comparison is that hyperscalers will spend more on CapEx in 2026 alone than all U.S. investor-owned utilities are projected to spend over the next five years combined. Both frame U.S. AI leadership as the strategic counterweight to China's dominance in solar, batteries, and EVs.

DJ Sampath (SVP, AI Software & Platform, Cisco), Jon Jones (Chief Revenue Officer, CoreWeave), and George Mathew (Managing Director, Insight Partners) discuss where durable AI value resides with moderator Rohit Choudhary (Founder & CEO, Acceldata)

Jones describes CoreWeave's trajectory from GPU rental to near-Fortune-100 scale in under five years. Sampath outlines a "trifecta" of enterprise AI challenges: infrastructure constraints, a trust deficit around agents, and a data gap from insufficient machine-data training.

Mathew argues that compound AI systems built around domain-specific data, not generic models, will capture most enterprise value.

Charles Zedlewski (Chief Product Officer, Together.ai), Robert Brennan (Co-Founder & CEO, OpenHands), and Nader Khalil (Director, Developer Technology, Nvidia) discuss open-source AI with moderator Yvonne Wassenaar (Senior Advisor & Board Director).

The panel argues that open AI systems, including accessible weights, open training methods, and community-driven inference stacks, are reaching performance parity with proprietary frontier models for most enterprise use cases at a fraction of the cost. Khalil frames Nvidia's role as enabling the full open-source stack.

Brennan positions OpenHands as proof that production-grade agentic AI can be built entirely on open foundations, while Zedlewski notes that nearly every major coding agent already runs some meaningful share of inference on open-weight models.

Rory O'Driscoll (Partner, Scale Venture Partners), Charly Kevers (CFO, Carta), Kris O'Toole (Principal, Semler Brossy), and Sean Feller (Partner, Gibson Dunn) discuss equity compensation strategy with moderator Deanna Kempinski (Senior Manager, HR Consulting, Baker Tilly)

The panel traces how extended time-to-liquidity, now averaging 13--14 years versus 4--5 historically, has reshaped the employee value proposition for equity. O'Driscoll notes that 75--80% of secondary transactions concentrate in a handful of marquee names, leaving most startup employees without meaningful liquidity for nearly a decade.

The group highlights a widening divide: SaaS companies are now heavily dilution-constrained while high-growth AI companies routinely offer engineers $10 million or more in equity for the same role

Micheline Nijmeh (CMO, ThoughtSpot), Ozge Tuncel Ozcan (CRO, Forter), and Meghan Gendelman (CMO, B2B, Canva) discuss AI-driven go-to-market strategy with moderator Ishan Mukherjee (Co-Founder & CEO, Rox)

Nijmeh describes ThoughtSpot's CEO mandate that every employee learn agentic AI, with no new headcount or tools approved until teams have first tried building their own agents. Ozcan says Forter's board no longer gives credit for adopting AI because it is now a baseline expectation, with revenue per employee as the primary metric boards track.

Gendelman offers a cautionary tale from a previous company where a mass AI content rewrite tanked SEO traffic because it lacked a point of view, underscoring that AI output still needs human editorial judgment.

Micheline Nijmeh (CMO, ThoughtSpot), Ozge Tuncel Ozcan (CRO, Forter), and Meghan Gendelman (CMO, B2B, Canva) discuss AI-driven go-to-market strategy with moderator Ishan Mukherjee (Co-Founder & CEO, Rox)

Nijmeh describes ThoughtSpot's CEO mandate that every employee learn agentic AI, with no new headcount or tools approved until teams have first tried building their own agents. Ozcan says Forter's board no longer gives credit for adopting AI because it is now a baseline expectation, with revenue per employee as the primary metric boards track.

Gendelman offers a cautionary tale from a previous company where a mass AI content rewrite tanked SEO traffic because it lacked a point of view, underscoring that AI output still needs human editorial judgment.

Part I features Ali Riley (4-Time Olympian, New Zealand Soccer), Ezra Frech (2-Time Paralympian, Team USA), and Ryan Murphy (9-Time Olympic Medalist; VP, Norwest Venture Partners) with moderator Johnny Hooper (2-Time Olympian, Water Polo).

Part II features David Michael (Chief Information Officer, LA28), Gene Sykes (Board of Directors, United States Olympic & Paralympic Committee, LA28), and Michelle McGuire (Partner & Chief Data and AI Officer, Deloitte) with moderator Jamie Montgomery (Co-Founder & Managing Partner, March Capital).

Athletes share first-person accounts of competing at the highest level and preparing for a home Olympics. Frech makes the case that the Paralympics, representing 15% of the global population, is a once-in-a-generation opportunity to shift how America views disability and elite performance. LA28 leadership then details the logistics of hosting: 351 medal events across more than 50 sports using 75% existing venues, a $2.5 billion sponsorship target already exceeded, and a ticket lottery offering 72 million tickets.

Michael Milken (Chairman, Milken Institute) in conversation with Jamie Montgomery (Co-Founder & Managing Partner, March Capital)

Milken revisits a formula he developed at Berkeley in 1965: human capital is the world's largest asset, and connecting capital to people with new ideas is the highest-multiplier act a society can perform. He argues that U.S. financial assets of roughly $250 trillion are dwarfed by an estimated $2,000--3,000 trillion in social and human capital.

He illustrates the stakes with striking mobility data and proposes a bold residency pathway for high-net-worth immigrants as a more powerful fiscal lever than conventional tax policy.

Ryan Daniels (Co-Founder & CEO, Crosby), Ori Eldarov (Co-Founder & CEO, OffDeal), and Nico Laqua (Founder & CEO, Corgi) discuss AI-native professional services with moderator Marc Bhargava (Managing Director, General Catalyst).

The founders, who are building a law firm, an investment bank, and an insurance company from scratch, share a thesis that the most durable AI opportunity lies in fully integrated companies, not software sold to incumbents. Daniels frames legal work as second only to code generation in its transformability by AI, and has eliminated the billable hour entirely at Crosby.

Eldarov built OffDeal around the $3.7 trillion in businesses needing to change hands over the next decade. Laqua describes Corgi becoming a licensed insurance carrier rather than building software on top of existing players.

Aklil Ibssa (Coinbase), Farooq Malik (Rain), Yoshi Yokokawa (Alpaca), and Bentzi Rabi (Utila) discuss stablecoin infrastructure with moderator Hans Morris (Managing Partner, NYCA; former President, Visa).

Morris frames the market opportunity: total stablecoin volume sits around $60 trillion, but over 90% remains in crypto settlement rather than real-economy use. Remittances and payroll, a combined $200 billion market, remain under 1% stablecoin-denominated.

The panel identifies five criteria for institutional-scale adoption: data standardization, interoperability, money quality, settlement infrastructure, and governance. They note meaningful progress since the Genius Act's passage and the emergence of more than 50 national regulatory frameworks.

Jason Lee (Head of Digital Natives, OpenAI), Michael Reitblat (CEO, Forter), and Kai Wu (Chief Revenue Officer, Airwallex) discuss autonomous transactions with moderator Kasey Lobaugh (Chief Futurist, Consumer Industry, Deloitte).

Reitblat's data shows fully agentic transactions remain under 10 basis points of all commerce, but are growing roughly 20% month-over-month. More than 12% of agentic transactions on Forter's network were fraudulent that morning, versus 1--3% for conventional e-commerce, as fraudsters move faster than legitimate adoption.

Market-size estimates range from Morgan Stanley's $200 billion by 2030 to McKinsey's $5 trillion, and the panel agrees that trust, not technology, is the real bottleneck to adoption.

Kahlil Lalji (Co-Founder & CEO, Natural), Paul Klein (CEO, Browserbase), and Sudheesh Nair (Co-Founder & CEO, TinyFish) discuss agentic infrastructure with moderator Matt Vail (Head of VC Partnerships, OpenAI).

The panel debates the shift from human-to-web to agent-to-web interaction, with Klein arguing that AI can operate on the internet as it exists today, with no rebuild required. Lalji predicts that agents will execute a majority of global payment volume within a decade.

Nair offers the most sweeping claims: that 2024 was the first year humans became a minority of web traffic, that 90% of B2B transactions worth roughly $15 trillion will be disintermediated by agents, and that brand equity built for human attention may become functionally obsolete.

Stelios Diamantidis (Chief Product Officer, Cognichip), Ofer Shacham (Co-Founder & CEO, Majestic Labs), and Joseph Costello (CEO, InPho) discuss AI hardware bottlenecks with moderator Shahin Farshchi (General Partner, Lux Capital).

Framed around Nvidia's $20 billion acquisition of Groq, the panel examines why chip development still runs on multi-year timelines despite rapid software innovation. Shacham argues that even a hypothetical zero-power, infinite-performance processor would yield only marginal gains because memory subsystems cannot feed it fast enough.

Diamantidis describes sophomore-level students designing entire custom CPUs in five-day hackathons using AI-assisted tools, pointing to a collapse in chip-design barriers. The panel's consensus is that system-level architectural rethinking, not incremental silicon improvements, will drive the next 100x efficiency gain in AI compute.

Youngjun Choi (Global Robotics & Physical AI Leader, EY) moderates a panel with Amit Goel (Head of Edge Computing Product & Robotics Ecosystem, Nvidia), Jagdeep Singh (Founder & CEO, Rhoda AI), Junghee Ryu (Founder & CEO, RLWRLD), and Ali Agha (CEO, Field AI).

The panel converges on manufacturing, logistics, and construction as the most immediate opportunities for physical AI, driven by acute labor shortages. Panelists debate general-purpose versus purpose-built robot form factors, converging on a near-term hybrid of wheeled mobility with a dexterous humanoid upper body.

Goel compares humanoid robots to the iPhone platform shift, while the group agrees that the biggest misconception in the field is overconfidence in teleoperation data as the path to scale. They argue instead for simulation and architectural approaches that embed physics into the model.

Izhar Medalsy (Co-Founder & CEO, Quantum Elements), Carmen Palacios-Berraquero (Founder & CEO, Nu Quantum), Pete Shadbolt (Co-Founder & Chief Science Officer, PsiQuantum), and Katie Pizzolato (VP, Quantum Platform & Global Community, IBM) discuss fault-tolerant quantum computing with moderator Prineha Narang (Professor, UCLA; Operating Partner, DCVC).

The panel traces the engineering path from today's noisy qubits to million-qubit error-corrected machines. Shadbolt describes PsiQuantum's 90,000-square-foot, 22-megawatt quantum facility under construction in Chicago, while Pizzolato outlines IBM's 2029 roadmap for a first fault-tolerant system.

All panelists converge on computational chemistry and molecular simulation as the industry quantum will transform first, with Shadbolt adding a provocative prediction that meaningful progress on code-breaking could be feasible within five years.

Raj Subramaniam (CEO & President, FedEx) in conversation with John Chambers (Founder & CEO, JC2 Ventures)

Subramaniam describes FedEx's three-layer AI strategy: operational efficiency, differentiated services, and a new external data unit called FedEx DataWorks. The strategy is built on a network that moves $2 trillion in global commerce and 2 petabytes of data daily across 17 million packages. The company rolled out AI training to all 530,000 employees in December.

Chambers argues that AI is moving five times faster than the internet transition. Both stress that cultural readiness and CEO-led change management, not technology access, determine whether large incumbents actually transform.

Atif Zaim (Deputy Chair & Managing Principal, KPMG U.S.) and Umesh Sachdev (Co-Founder & CEO, Uniphore) discuss enterprise AI deployment with moderator John Chambers (Founder & CEO, JC2 Ventures).

Zaim describes KPMG's shift from AI pilots to enterprise-wide industrialization across 35,000 U.S. professionals, arguing that the biggest challenge is human, not technical. Cultural change cannot be delegated to IT. Sachdev, drawing on Uniphore's 200+ C-suite conversations over the past year, says enterprise buyers have stopped asking what AI can do and now demand dollar-value outcomes within 12 months.

Chambers frames the moment competitively, predicting the top three enterprise B2B software companies of the next decade will be AI-native from the ground up.

Pam Maynard (Chief AI Transformation Officer, Microsoft) and John Chambers (Founder & CEO, JC2 Ventures) discuss enterprise AI adoption with Lari Hämäläinen (Senior Partner, McKinsey & Company)

Hämäläinen opens with a provocative briefing: despite explosive model improvement, 95% of enterprises report no measurable P&L impact from AI, and initiative cancellations more than doubled in 2025. He argues this changed by late 2025, when developer productivity flipped from a 20% real-world decline to a 20% gain as new model generations crossed a five-hour automation threshold.

Chambers adds that the internet raised U.S. productivity 1.4% over a decade, while Q4 2025 productivity hit 2.7%. He believes AI can push that to 5% or more.

Jim Martineau (Co-Founder & CEO, Nosis Bio), Jacob Rubens (Co-Founder, Tessera Therapeutics), and Chun-Hao Huang (Co-Founder & CEO, Algen Bio) discuss the convergence of biotech and AI with moderator Pradeep Khosla (Chancellor, UC San Diego).

All three companies build directly on Nobel Prize-winning science, including CRISPR gene editing and RNA biology, and argue that the time between discovery and clinical application has compressed dramatically. Martineau contrasts traditional drug development's roughly 95% failure rate with the 70% success rate of targeted gene-silencing medicines already on the market.

Rubens describes Tessera's first gene-writing drug now in patients for alpha-1 antitrypsin deficiency, with a second targeting sickle cell disease entering clinic in 2026.

Mel van Londen (Co-Founder & CEO, Develop Health), Anan Kashyap (CFO, Hippocratic AI), and Andre Esteva (Co-Founder & CEO, ArteraAI) discuss healthcare AI adoption with moderator Katya Andresen (SVP & Chief Digital & Analytics Officer, The Cigna Group).

The panel moves beyond efficiency narratives to genuinely new capabilities. ArteraAI predicts patient outcomes to personalize cancer treatment, Hippocratic AI's voice agents conduct clinical outreach at scale, and Develop Health automates prior authorization, addressing $500 billion in payer-provider administrative waste.

Esteva frames adoption through five moats: data, evidence, regulation, reimbursement, and clinical workflow. Hippocratic AI reports 99.9% safety across 11 million patient calls, while van Londen's "fully subtractive" design philosophy avoids retraining staff or upending existing workflows.

Dimitrios Iliopoulos (Co-Founder & CEO, Athos Therapeutics), Mike Nally (CEO, Generate:Biomedicines), and Jorge Cardoso (Co-Founder & Chief AI Officer, Hologen) discuss AI's role in drug development with moderator Dr. Narimon Honarpour (SVP, Global Development, Amgen)

The panel reaches broad consensus that AI has genuinely transformed early-stage discovery, including protein design, target identification, and hypothesis generation, while its translation into successful late-stage clinical outcomes remains largely unproven.

Nally frames the opportunity through "Arrow's Law," the $2.9 billion cost to develop a new drug, predicting the industry will see its first downward cost inflection in 50 years within five years.

Scott Sandell (Executive Chairman & CIO, NEA) and Peter Chung (Managing Director & CEO, Summit Partners) discuss the future of venture investing with Sumant Mandal (Co-Founder & Managing Partner, March Capital)

Sandell and Chung, each with over 30 years in venture, argue that AI is transforming the economy far faster than the internet did, compressing what took 25 years into a fraction of the time. They point to $600 billion currently being invested in U.S. AI infrastructure by the world's most profitable companies, roughly 10x the peak cloud buildout spend a decade ago.

Both argue that AI-native companies play by entirely different rules, forcing firms built during the dot-com and mobile eras to rethink how they evaluate businesses.

Michelle Gonzalez (Corporate VP & Global Head, M12), Jai Das (Co-Founder & President, Sapphire Ventures), and Matt McIlwain (Managing Director, Madrona Venture Group) discuss the private capital landscape with moderator Ben Bergman (Senior Correspondent, Business Insider).

McIlwain warns of a tension between disruption speed and durability, while Das focuses on infrastructure "picks and shovels" investments as less exposed to model displacement. Gonzalez highlights a striking reversal: 90% of subscription AI revenue went to OpenAI in February 2025; a year later, 70% was going to Anthropic. McIlwain discloses that Madrona passed on OpenAI's $25 billion round in 2022, before ChatGPT launched.

The panel also compares AI-native company economics with the SaaS benchmark and debates the likely order of anticipated SpaceX, Anthropic, and OpenAI IPOs.

Charles Black (Head of North America Technology, Equity Capital Markets Investment Banking, Citi), Emma Norchet (Partner, T. Rowe Price), Carly Roddy (Head of West Coast Private Capital Markets, J.P. Morgan), and Jack Cassel (Head of New Listings, Nasdaq) discuss the 2026 IPO market with moderator Michael Brown (Partner, Fenwick)

The panel describes a "very selective" market. Only 23 tech IPOs priced in 2025, and 2026 opened with just three before geopolitical disruption paused momentum. There is a sharp bifurcation between AI-adjacent companies attracting strong investor appetite and traditional software facing valuation compression.

The panel highlights the rise of "triple-track" processes --- IPO, M&A, or private continuation vehicle --- as the new standard exit path, and debates whether SpaceX, OpenAI, and Anthropic will go public in 2026.

Zane Mountcastle (Co-Founder & CEO, Picogrid), Ian Kalin (Co-Founder & CEO, TurbineOne), Grant Demaree (Co-Founder & CEO, OneBrief), and Katie Arrington (CIO, IonQ) discuss the defense industry's transformation with moderator General John Allen (Ret.), Strategic Advisor to Microsoft.

The founders describe a new generation of software-native defense companies with SaaS-style margins and rapid deployment timelines, incompatible with traditional cost-plus prime contractor economics. Arrington says the department is now run with a venture mindset, with EBITDA and innovation replacing cost-plus as the metrics that matter.

OneBrief's Demaree describes reducing multi-week operational planning cycles to about 10 minutes using agentic AI, with users productive on classified networks within an hour versus 17 months under legacy systems. The panel identifies the Pentagon's Risk Management Framework as the single biggest obstacle to faster AI deployment.

Joe Lonsdale (Founder & Managing Partner, 8VC) and Geoffrey von Maltzahn (Co-Founder & CEO, LILA Sciences) discuss national AI policy with moderator Martin Giles (Policy Fellow & Executive Director, Hoover Technology Policy Accelerator, The Hoover Institution).

As AI becomes a driver of national competitiveness, this session examines the policy decisions positioning the United States to lead. From the DOE’s Genesis Mission to accelerate science through AI to the U.S. approach to open-source models and AI infrastructure, the conversation explores how government and private builders are aligning around compute, talent, and capital.

The discussion focuses on what it will take to secure durable American leadership in the next wave of AI.

Steve Westly (Founder & Managing Partner, The Westly Group) in conversation with Rick Caruso (Chairman, Caruso) and Mayor Matt Mahan (Mayor of San Jose)

Westly frames the discussion around California's paradox: the world's fourth-largest economy, larger than Germany despite less than half its population, yet ranked 51st in business climate and near the bottom nationally in education. Mahan details how San Jose cut fees representing 10--20% of housing costs, driving 2,000 additional homes above baseline and moving 4,000 people off the streets. Caruso contrasts Los Angeles's fewest housing starts in a decade and $1 million-per-person homelessness spending that yielded only 1,500 people housed.

The conversation turns to accountability as the throughline for fixing schools, safety, and governance. Both argue California's fix lies in the pragmatic middle, not the partisan extremes, and close by promoting civic organizations built to sustain private-sector engagement between election cycles.

General David Petraeus (Ret.) (Partner, KKR), General Amos Yadlin (Ret.) (President & Founder, MIND Israel), and General John Allen (Ret.) (Strategic Advisor to Microsoft)

Three retired four-star generals deliver a frank assessment of the Middle East after Israel's campaign against Iran and its proxies. Petraeus walks through Iranian force sizes, missile penetration rates, and the strategic implications of Iran's nuclear program, while Yadlin gives the Israeli military's targeting rationale and the distinction between wars of necessity and wars of choice.

The numbers are stark: Iran's 3,000--4,000 air defense units allowed only 5--7% of Israeli-targeted missiles through, what Petraeus calls a catastrophic failure of Iran's deterrence posture. Allen raises a less-discussed risk: U.S. precision-weapons stockpiles are depleting faster than they can be rebuilt.

Marko Papic (Macro-Geopolitical Chief Strategist, BCA Research); K.T. zu Guttenberg (Chairman, Guttenberg Ventures; former German Defense and Economy Minister)

Marko Papic (Macro-Geopolitical Chief Strategist, BCA Research) and K.T. zu Guttenberg (Chairman, Guttenberg Ventures; former German Defense and Economy Minister) Zu Guttenberg describes Europe as a "Sleeping Beauty" kissed awake by Trump, Putin, and Xi after 30 years of complacency, pointing to Germany's deregulation push, the abolition of its Supply Chain Due Diligence Act, and a historic unlocking of up to 500 billion euros, potentially a trillion, for defense spending.

Papic backs the thesis with markets: European banks have outperformed NASDAQ over the past two years. Zu Guttenberg argues Germany needs 400,000 well-educated immigrants a year just to sustain its economy and military, and advocates a UK-inspired 40-40-20 defense spending model that favors technology over legacy platforms. Both speakers agree Europe's chokepoints in critical tech supply chains give the continent real leverage in negotiations with Washington.

Marko Papic (Macro-Geopolitical Chief Strategist, BCA Research) and Michael J. Mazarr (Senior Political Scientist, RAND Corporation)

Papic frames the US-China relationship through capital allocation — up to four Xi-Trump summits are expected this year, an extraordinary level of engagement, and even a modest diplomatic thaw could unlock major capital flows back into Chinese assets. Mazarr brings the security counterweight, arguing that nobody currently has a coherent theory for a stable US-China equilibrium over the next two to five years.

China outperformed US equities by double in the prior year, yet private market investors face a structural problem public investors don't: seven-year lockups that make it hard to trade in and out as diplomatic winds shift. Mazarr adds that the US national security community is quietly far more hawkish on China than official strategy documents suggest, with landmines around investment, semiconductors, and export controls sitting underneath any surface-level truce.