It’s a boat, it’s a plane, it’s REGENT’s Seafarer | E2205

Key Concepts

• Sea Gliders: Hybrid vehicles combining boat and airplane functionalities, designed for efficient, low-altitude flight over water.
• Hydrofoils: Underwater wings that lift the hull of a vessel out of the water at speed, reducing drag and increasing efficiency.
• Ground Effect: An aerodynamic phenomenon where a wing flying close to a surface (like water) experiences increased lift and reduced drag due to trapped air.
• Float-Foil-Fly: The three operational modes of a sea glider: floating at rest, hydrofoiling at moderate speeds, and flying at higher speeds.
• Electric Propulsion: The use of electric motors and batteries to power the vehicle, enabling efficient and quiet operation.
• Digital Flight Control Systems: Advanced computer systems that manage the vehicle's flight, ensuring stability and safety, particularly at low altitudes.
• AI Agents: Autonomous software programs designed to perform specific tasks, used by Convexia for drug discovery and by Arill for data access.
• IP (Intellectual Property): In the context of Convexia, refers to unused or overlooked drug research and development.
• Digital Twins: Virtual replicas of physical systems or processes, used by Convexia for simulating drug performance.
• POSIX (Portable Operating System Interface): A standard that allows applications to interact with storage as if it were local, crucial for Arill's technology.
• Caching: Storing frequently accessed data in a faster, more accessible location (like Arill's system) to improve performance.
• Object Storage (e.g., S3): A method of storing data as discrete units (objects) with associated metadata, typically low-cost but with higher latency.
• Block Storage (e.g., EBS): Storage that is presented as raw blocks, often used for operating systems and databases, offering lower latency than object storage.

Regent: Revolutionizing Coastal Transportation with Sea Gliders

Regent is developing "sea gliders," a novel class of electric vehicles that are hybrids of boats and planes, aiming to transform coastal transportation. These vehicles combine the best of both worlds, offering speed, efficiency, and the ability to operate without traditional airports.

Main Topics and Key Points

• Vehicle Concept: Sea gliders are designed to "float, foil, and fly." They begin by floating at a dock, then use hydrofoils to lift out of the water and maneuver at speeds up to 50 mph through crowded waterways, leveraging technology from America's Cup and SailGP. Finally, they take off from these foils and fly at approximately 30 feet above the water's surface, utilizing the ground effect for aerodynamic efficiency and extended range.
• Operational Advantages:
  • No Airports Needed: Boarding occurs at city centers or docks, eliminating the need for travel to airports and the associated security and waiting times.
  • Wave Tolerance: Sea gliders can operate in all weather conditions up to five-foot waves, ensuring high utilization and reliability.
  • Speed and Cost: The Providence to New York route (180 miles) can be completed in about an hour for less than $100, offering a competitive alternative to Amtrak (3.5-4 hours) and flights.
  • Regulation: As they operate within a wingspan of the water and do not ascend to airport altitudes, sea gliders are regulated by the Coast Guard as vessels under maritime jurisdiction.
• Technical Innovations:
  • Hydrofoiling and Ground Effect: The core technology relies on advanced hydrofoil design and the exploitation of ground effect for efficient flight.
  • Electric Propulsion: Enables precise control over distributed propulsion systems, allowing for high lift at low speeds necessary for takeoff from foils. This also significantly reduces maintenance costs and environmental impact.
  • Digital Flight Control Systems: Sophisticated software manages the complex transitions between float, foil, and fly modes, ensuring safety and allowing for simplified pilot operation. The system governs flight envelope and altitude, with the pilot primarily controlling left/right and speed, akin to "super cruise" in cars.
• Progress and Timeline:
  • The Viceroy prototype, carrying 12 passengers or 3500 lbs of payload, is currently undergoing sea trials.
  • Manufacturing facility in Rhode Island broke ground in January 2025.
  • First sea gliders are expected to be delivered in the second half of 2027.
  • Training for mariners to operate sea gliders is estimated to be a 4-6 week course.
• Market and Business Model:
  • Orders and Backlog: Regent has secured significant orders, with a backlog exceeding $10 billion by the end of 2024.
  • Sales vs. Operation: Regent plans to sell sea gliders to airlines, ferry companies, energy companies, hotels, and defense contractors, rather than operating its own coastal airline. This model offers better margins and brand recognition through established operators.
  • Defense Applications: The US Marine Corps is a key partner, utilizing sea gliders for high-speed logistics in contested environments, particularly for island chain operations in the Indo-Pacific. Medevac capabilities are also a growing application.
  • Pricing: The upfront cost of a sea glider is comparable to a Cessna Caravan or Twin Otter, but operating costs are significantly lower.
• Rhode Island Ecosystem: Rhode Island is highlighted as an ideal location due to its maritime defense renaissance, skilled workforce in composite boat building, proximity to military decision-makers, and state support for manufacturing.

Notable Quotes

• "We are creating a fundamentally new vehicle. It it goes fast. It flies low. It's sort of operating in this different way." - Alex (Host)
• "Sea gliders are sort of hybrids of boats and planes. Sort of use the best of both worlds." - Billy Tallheimer (Regent)
• "So, it's float foil fly." - Billy Tallheimer (Regent)
• "The technical hard part of what we're doing and why we can only do this now with electric propulsion and digital flight control systems." - Billy Tallheimer (Regent)
• "We are regulated under the Coast Guard as a vessel. There is a special rule in US law that is also mirrored in international law and regulation that this kind of vehicle and a sea glider is a wing in ground craft is a vessel under maritime jurisdiction." - Billy Tallheimer (Regent)

Convexia: AI-Driven Drug Discovery and Commercialization

Convexia is an "AI maximalist pharma company" focused on identifying, evaluating, and bringing overlooked or shelved drug intellectual property (IP) to market. They leverage AI agents to mine data, assess scientific and commercial viability, and accelerate the drug development process.

Main Topics and Key Points

• Core Mission: To identify and commercialize underutilized drug IP, aiming to bring life-saving drugs to market faster and more efficiently.
• AI Agent Stack:
  1. Sourcing Agent: Mines structured and unstructured data (company reports, patents, manufacturing data) to predict which drugs are shelved, overlooked, or available for licensing. This agent is hypothesis-generating and thesis-aware, offering an advantage over traditional competitive intelligence tools.
  2. Scientific Agent: Utilizes AI and ML-based models to evaluate the scientific and clinical viability of potential drug candidates. While physics-based models have been the gold standard, ML-based models are rapidly advancing and are expected to surpass them due to their pace of innovation and computational capabilities.
  3. Commercial Agent: Assesses the market viability and commercial potential of drugs.
  4. Clinical Agent: Employs digital twin simulations and analysis of clinical trial components (manufacturing, patient stratification, trial design) to predict success rates. The focus is on building these simulations from smaller, manageable components.
• Drug Failure Reasons: Drugs typically fail to reach the market due to:
  • Lack of safety and efficacy in humans.
  • High clinical risk (e.g., manufacturing issues, difficult trial execution).
  • Lack of commercial viability (e.g., existing superior treatments, small market size, poor insurance coverage).
• Market Opportunity:
  • High Failure Rate: Only 5-10% of drugs entering Phase 1 make it to market, leaving a vast amount of pre-clinical and shelved assets.
  • AI's Role: As AI drug discovery increases, the number of novel drug molecules will grow, making the evaluation of their success critical. The future of pharma is seen as increasingly reliant on M&A and efficient evaluation.
  • Rare and Ultra-Rare Diseases: Convexia is particularly interested in these underserved markets, where smaller outcomes (under $1 billion) can be significant for their lean operation, unlike large pharma companies seeking multi-billion dollar outcomes.
    • Definition: Ultra-rare diseases affect under 200,000 people in the US annually; rare diseases affect under 50,000.
    • Market Size: Only 5% of known rare and ultra-rare diseases have cures, leaving a vast market opportunity.
• Business Model and Strategy:
  • Technology Sales: Convexia sells parts of its AI technology to small to midcap pharma companies and investment firms (biotech VCs, hedge funds).
  • Long-Term Goal: To become a fully vertically integrated "AI maximalist pharma company," running the entire drug development process internally. This transition is expected to take 9-12 years, similar to other companies in the space, but Convexia aims for a faster pace.
  • Repurposing Existing Drugs: A significant focus is on identifying approved drugs that can be repurposed for rare diseases, leveraging their AI agents to analyze mechanisms of action and potential off-target impacts.
• Technological Advancements:
  • ML vs. Physics-Based Models: ML models are rapidly advancing and are expected to become the gold standard in areas like binding and toxicity prediction, surpassing traditional physics-based models.
  • Digital Twins: While challenging, digital twin simulations are being developed by focusing on smaller components of clinical trials.
• Regulatory Tailwinds: Convexia is benefiting from regulatory incentives for rare disease drugs, such as Orphan Drug Designation (ODD) and Priority Review Vouchers (PRVs), which make pursuing these drugs more lucrative.
• Capital and Classification: Convexia recently closed its seed round, classifying itself primarily as an AI company applied to biopharma, with a future transition towards a biotech classification as they develop their own asset pipeline.

Notable Quotes

• "We want to be on that forefront." - Ayan Periq (Convexia)
• "The future of pharma is in M&A." - Ayan Periq (Convexia)
• "The goal is to be like as the tagline says, an AI maximalist pharma company." - Ayan Periq (Convexia)
• "I think like Sam Altman talks about this like as the model gets better our company will also get better. So we're not really competing against models but like it's very synergetic." - Ayan Periq (Convexia)
• "95% of the market is still up for grabs." - Ayan Periq (Convexia)

Arill: Universal Access Layer for Cloud Data

Arill is developing a technology that makes cloud data accessible as if it were local, offering significant improvements in speed and cost compared to traditional cloud storage methods. They aim to become the "universal access layer for data."

Main Topics and Key Points

• Core Problem: Accessing data stored in cloud object storage (like AWS S3) is slow and expensive. Traditional methods involve transferring large amounts of data to local servers, which is time-consuming and costly.
• Arill's Solution: Arill provides a Software-as-a-Service (SaaS) solution that acts as a caching layer. It makes cloud storage appear local to applications, eliminating the need for data transfer to local servers.
  • Mechanism: Arill runs a fleet of servers with high-performance SSDs. Data from cloud storage (e.g., S3) is cached on these SSDs.
  • Intelligent Pre-caching: Arill uses predictive algorithms to anticipate what data users will need, ensuring it's available in the cache before it's requested.
  • POSIX Compatibility: Arill's system is POSIX-compliant, allowing any application that works with local files and folders to interact with cloud data seamlessly, without needing to be rewritten for cloud APIs. This unlocks a vast range of existing applications (databases, video transcoding tools) for cloud data.
• Performance and Cost Benefits:
  • Speed: Offers up to 30x lower latencies than direct S3 access.
  • Cost: Claims up to 90% lower cost than using high-performance block storage like EBS.
  • Elasticity: Customers pay only for the active data they use in the cache, unlike traditional block storage where they pay for provisioned capacity, much of which might be unused.
• How it Works:
  • Data Access: Applications can start using data even if it's not fully cached. Arill asynchronously backfills the cache, allowing for earlier interaction with datasets.
  • No Infrastructure Needed: Arill's SaaS model means customers don't need to manage their own caching infrastructure.
• Technical Difficulty: The underlying technology is extremely complex, as evidenced by Amazon being the only major cloud provider to have built a similar product (Elastic File System). Arill's unlock comes from understanding how cloud primitives can be reimagined.
• Market and Demand:
  • High Demand: Tremendous demand from large enterprises and innovative startups, particularly in AI, who need fast and cost-effective data access.
  • Customer Segments:
    • Enterprises: For ML/AI research environments requiring collaboration and rapid data sharing.
    • Geospatial and Bioinformatics Companies: Bridging the gap for applications not natively designed for cloud storage APIs.
    • AI/ML Use Cases:
      • Training: Facilitating rapid data ingestion into GPUs for model training.
      • Inference/Agents: Enabling AI agents to quickly access specific data for tasks like RAG (Retrieval Augmented Generation).
• Business Model and Economics:
  • Cost of Goods Sold (COGS) Model: Arill operates on a model with significant COGS, as they rent servers and SSDs from cloud providers. They aim for margin on top of these costs.
  • Customer Pricing: Pricing is tied to active data usage in the cache. If data is not actively used, customers are not charged by Arill, and it remains on S3. This contrasts with the unused capacity charges of traditional cloud storage.
  • Cloud Provider Relationship: Arill is a customer of cloud providers like AWS and Google Cloud. They are seen as enabling more enterprises to use these clouds, making them a valuable partner rather than a competitor.
  • Multi-Cloud Support: Arill is currently available on Google Cloud and plans to expand to all major hyperscalers and GPU clouds based on customer demand.
• Future Vision:
  • Universal Access Layer: Arill aims to become the standard way for any cloud application to interface with its data, regardless of where it resides.
  • Higher-Level Services: Beyond storage, Arill plans to offer compute services tightly integrated with their caching layer to generate additional revenue.
  • Abstraction: The long-term goal is to abstract away the complexities of cloud infrastructure, allowing developers to focus on their applications.
• Capital and Funding: Arill has raised capital, with Felicis Ventures as a lead investor. They have capital requirements for expanding their presence in different cloud regions and for R&D into higher-level services.
• Hiring Focus: Arill is currently focused exclusively on engineering talent, believing that a strong engineering team can drive product development before scaling Go-To-Market (GTM) functions.

Notable Quotes

• "We've cut out that step completely and just made cloud storage appear local." - Hunter Leaf (Arill)
• "We effectively if you think about S3 or cloud storage in general as this low-cost slow storage there is also highcost high performance storage you can buy which are traditionally like the SSDs that you might have in your laptop." - Hunter Leaf (Arill)
• "It's easy to explain. Everyone understands caches, but it's extremely hard on the back end." - Hunter Leaf (Arill)
• "Amazon is the only company that has actually built a product in this space because it is so tremendously difficult to do." - Hunter Leaf (Arill)
• "Our goal over like the next three years is to become the universal access layer for data." - Hunter Leaf (Arill)
• "Storage is a place where we start and then we want to provide services on top of that that might be more valuable to our customers." - Hunter Leaf (Arill)
• "We are somewhere in the middle which is because we're offering storage as a service and storage is something that both the capacity and the performance and the safety actually scale with the number of servers you have." - Hunter Leaf (Arill)