Podcast Summary: This Week in Startups – “It’s a Boat, It’s a Plane, It’s REGENT’s Seafarer” | E2205 (Nov. 8, 2025)

Episode Overview

Host Alex Wilhelm (filling in for Jason Calacanis) interviews three innovative founders whose startups are reshaping transportation, biotechnology, and cloud infrastructure:

1. Billy Talheimer (REGENT): Electric “seaglider” vehicles for ultra-fast regional coastal transit and defense logistics.
2. Ayaan Parikh (Convexia): AI agents to mine and revive shelved pharmaceutical IP and accelerate drug discovery, especially for rare diseases.
3. Hunter Leith (Arkil): Seamless, high-performance cloud storage caching, making remote data feel instantly local for AI and enterprise applications.

The episode centers on radical approaches for seemingly intractable problems—reinventing regional transit, unlocking the value of existing drugs with AI, and making cloud data instantly accessible.

REGENT: Sea, Air, and Transit Reinvented (00:00–29:29)

The Seaglider: Not a Boat, Not a Plane – Something New

• Hybrid Innovation: REGENT is building "seagliders": electric vehicles that combine hydrofoil and ground effect wing technology to travel 180 miles on a single charge—e.g., from Providence to New York in under an hour, for under $100.
• Three Modes: "We float, foil, fly," explains Billy Talheimer.

  “We are creating a fundamentally new vehicle. It goes fast, it flies low. It's sort of operating in this different way.” – Billy Talheimer (00:06)

• Operation: Boards like a boat at city-center docks, hydrofoils through congested harbor zones at 50 mph, then takes off using ground effect, flying ~30 feet above the water for maximum efficiency and minimal regulatory/airport hassle.

Key Technology Breakthroughs

• Hydrofoil Technology: Leveraging America's Cup innovations for greater wave tolerance—"operate in all weather up to five-foot waves" (03:38).

• Electric Propulsion: Enables efficient distributed power (multiple small propellers), critical for takeoff and cost savings (09:15).

• Digital Flight Control: Advanced software turns the vehicle into a "fly-by-wire boat"—simpler to pilot, needing only a few weeks’ training for experienced mariners.

  "So much of your pilot training is how to take off and land…In a Seaglider you don't even have control of that. You're driving a boat and if something goes wrong, you land, press the land button." – Billy Talheimer (17:02)

Regulatory Advantage

• Not an Aircraft: Because seagliders stay below a certain altitude and above open water, they’re regulated by the Coast Guard as vessels, not the FAA, greatly streamlining deployments (11:13).

Business Strategy and Progress

• Commercialization Timeline: First Viceroy prototypes (12-seat, cargo-capable) are in sea trials. Full-scale deliveries expected in the second half of 2027 (28:28).
• Order Book: $10B+ across airlines, ferry operators, new “seaglider-only” carriers, defense (Marine Corps), oil and gas, and luxury buyers (21:27).
• Sales over Operation Model: REGENT focuses on vehicle sales rather than operating its own lines—citing brand trust, economic, and value-accretion reasons (18:14).

Defense & Rhode Island as HQ

• Defense Applications: High-speed logistics in island/contested zones; medevac; rapid distributed transport for Marine Corps and DoD (23:23).

  "We're like a boat. We can operate from the water… but we have the speed of a plane. That's a huge gap we're filling." – Billy Talheimer (23:23)

• Rhode Island Factor: Maritime expertise, legislative support, military synergy, and a tight manufacturing cluster (25:29).

Convexia: AI-Driven Drug Discovery and Repurposing (31:10–50:30)

Mission & Approach

• AI Maximalist Pharma: Convexia deploys a suite of AI “agents” to identify overlooked drug candidates, analyze their scientific, clinical, and commercial viability, and help bring them to market—especially for rare/ultra-rare diseases.

  "We're looking at all types of drugs. Early stage... shelved assets... or things overlooked. We evaluate science, clinical, and commercial viability." – Ayaan Parikh (31:40)

Solving the Drug Discovery Bottleneck

• The Problem: Only 5–10% of drugs that enter clinical trials ever reach market; thousands more are shelved for non-scientific reasons (33:24).
• AI Sourcing: Models mine patents, reports, manufacturing, and public data to surface dormant candidates, even when formal access is relationship-based (35:29).
• Scientific Agents: ML models now rival traditional physics-based drug prediction in some areas (e.g., binding), with rapid improvements underway (37:10).
• Clinical Agents: Early digital twin work focuses on practical chunks—manufacturing, trial design—not full-patient simulation yet (39:21).

Commercial Model and Market Impact

• Current Market: Selling technology to mid-size pharma, biotech VCs, hedge funds. Not big pharma yet (41:46).
• Endgame: Run vertical “AI-first” pharma operations, potentially within a year if a strong candidate emerges (42:18).

Rare Disease Focus & Repurposing

• Strategic Shift: Rare disease market is attractive due to regulatory tailwinds (priority vouchers) and less competition—Convexia’s smaller scale can do what Pfizer cannot (44:01).

• Repurposing: Graph neural networks and pharmacological databases are used to match existing compounds to new indications—thousands of rare diseases could be addressed (47:07, 45:55).

  "Out of all the rares and ultra rares that are known, only 5% have cures. So 95%... is still up for grabs." – Ayaan Parikh (45:55)

• Team/Culture: Hiring AI/ML engineers interested in health and drug discovery.

Arkil: Making Cloud Data Instantly Local (51:08–75:12)

The Vision: Cloud Like Local Storage

• Instant Access: Arkil builds an intelligent caching layer that makes cloud storage (e.g., Amazon S3, Google Cloud Storage) immediately accessible on demand—as if the data were a local disk.

  "We've cut out that step completely and just made cloud storage appear local." – Hunter Leith (51:24)

• No Infrastructure Hassle: SaaS model with Arkil managing the heavy lifting (cache infrastructure, prediction/modeling of what to load).

Technology & Competitive Moat

• Advanced Caching: Massive, shared, SSD-based cache runs near cloud regions. Intelligent prediction of usage patterns so users are rarely waiting for data (53:29).
• POSIX Compatibility: Applications can treat Arkil storage as normal file systems, enabling non-cloud-native tools (databases, FFmpeg, bioinformatics pipelines) to work seamlessly without complex rewrites.

  “We take that cloud storage and turn it into this… POSIX… compatible format that can run with any application that works on a machine today.” – Hunter Leith (65:08)

Business Model & Market

• Transparent Pricing: Customers pay only for data actively cached, not for idle storage—addresses a core frustration with cloud pricing (68:50).
• Mutually Beneficial for Cloud Vendors: Arkil rides on major clouds' infrastructure, making their object storage more attractive while not directly competing with high-margin storage offerings (59:22, 60:08).
• Market Pull: Strong demand from enterprise R&D, geospatial, bioinformatics, and AI companies dealing with huge remote datasets (64:14).
• AI Enablement: Critical for large-scale AI training (RAG, agentic workflows), moving massive datasets quickly between cloud and compute (61:41).

Startup Dynamics & Team

• Capital Needs: Moderate; expansion into new cloud regions requires provisioning SSDs, but no need to own datacenters or physical infrastructure (73:45).
• Go-To-Market: Currently engineering-focused, with plans to build out GTM as product matures.

  "If you get the best engineers, you can grow the product and then fill GTM in later." – Hunter Leith (74:42)

Notable Quotes & Moments

• On disrupting regional transit:

  "If you guys are cheaper and faster [than trains], I mean it's game over. It's absolutely game over." – Alex (15:42)

• On regulatory jujutsu:

  "Because we stay... about half a wingspan at that 30ft. We are regulated under the Coast Guard as a vessel." – Billy Talheimer (11:13)

• On the rare disease market:

  "Because Convexia is so lean... you can afford to go after these drugs that don't match the cost structure of pharma companies that we can name." – Alex (45:25)

• On cloud incentive misalignment:

  "If all of these clouds are making billions on effectively the unused space that's on those disks, it's very unlikely they would want to spend a tremendous amount of R&D to get rid of that extremely high margin income stream." – Hunter Leith (67:21)

Episode Structure & Key Timestamps

| Segment | Guest | Key Topics | Timestamps | |----------------------------------------------|--------------------|----------------------------------------------------------|---------------| | REGENT: Electric Seagliders | Billy Talheimer | Hybrid tech, city-to-city transit, defense, business | 00:00–29:29 | | Convexia: AI for Drug Discovery | Ayaan Parikh | IP mining, AI agents, rare diseases, repurposing | 31:10–50:30 | | Arkil: Cloud Data Caching | Hunter Leith | Instant local-like cloud storage, SaaS, infra innovation | 51:08–75:12 |

Takeaways

• REGENT could revolutionize regional travel (and military logistics) with practical, near-term tech that side-steps both infrastructure and regulatory obstacles.
• Convexia demonstrates the horizontal reach of AI/ML, resurrecting value from the pharma industry’s forgotten corners—especially for underserved and rare disease communities.
• Arkil is the kind of infrastructural innovation that quietly makes the bold work of other startups possible, especially as AI’s hunger for rapid, massive data access accelerates.

Where To Learn More / Careers

• REGENT (Seagliders): regentcraft.com/careers
• Convexia (AI Pharma): convexia.bio — hiring AI/ML for bio
• Arkil (Cloud caching): arkil.com | disk.new — hiring engineers

This episode powerfully illustrates how ambitious founders are attacking outdated realities in infrastructure, medicine, and transit by mixing world-class technology with sharp business models. Each founder’s candor, technical depth, and willingness to explain their moonshots make this an essential listen for founders, investors, and hopeful technologists.