The dawn of surgery bots + buy a home for $250 (w/ Andromeda & Mogul) | E2313

Andromeda Surgical co-founder Nick Damiano argues that surgery is technologically analogous to the horse-and-buggy era in terms of autonomy, but that AI and autonomous systems will be ubiquitous in surgery within a decade. His co-founder built the first autonomous truck to drive on a freeway before starting the company, and more than half of Andromeda's team came from the autonomous vehicles space rather than healthcare, with additional candidates recruited from Neuralink.

Andromeda uses an off-the-shelf KUKA robotic arm controlled via an iPad, with all core innovation residing in the software layer. The company launched its first robot in three years and spent approximately 15 million dollars total, compared to an industry norm of hundreds of millions, and recently raised a round in the tens of millions with a formal announcement pending. The company is starting with urology, specifically a procedure called HOLEP for benign prostatic hyperplasia, a condition affecting nearly every man at some point. Urologists were the first adopters of surgical robots historically, making them among the most technology-receptive surgeons. The HOLEP procedure has a dramatic skill gap where a novice surgeon takes around five hours while the best surgeons complete it in 30 minutes and can perform eight to ten cases in a day.

Andromeda has completed 45 surgical cases, which Damiano describes as representing 100 percent of the world's data for their specific procedure type. The company has built a Google Maps-style landmark tracking feature for the body and an auto-pivoting feature that controls the pivot point and minimizes forces applied during procedures to reduce tissue damage. Software updates are released every week, a pace that has shocked surgeons accustomed to medical technology that changes over years. Surgical data is scarce relative to domains like autonomous driving and becomes a competitive moat for companies that accumulate it. Andromeda has sourced a large dataset of surgical videos from YouTube but notes that robot kinematics, force data, and other procedural signals not captured on video must be collected directly through clinical cases. Jason Calacanis observed that reaching autonomous surgery will take longer than autonomous coding because surgical data is far harder to index and obtain than the coding data available freely online.

The AI approach currently used is closer to machine learning methods from autonomous vehicles than to generative AI, with generative AI deliberately avoided in the first product version due to concerns about hallucinations in a medical context. Andromeda uses LLMs across internal operations and documentation but applies strict guardrails to coding to prevent non-deterministic bugs in a safety-critical environment. The company is currently cleared to operate only in New Zealand but is close to regulatory clearance in several other countries. The FDA strategy involves submitting small incremental updates frequently so each regulatory review is manageable, building cumulatively toward autonomous surgery across many procedures. Having surgeons in the loop during procedures makes regulatory approval significantly easier than a fully autonomous system would. The long-term vision includes a single surgeon overseeing approximately ten simultaneous procedures from an air traffic control style center, with eventual deployment in remote regions and on other planets.

Alex Blackwood, co-founder of Mogul and a former Goldman Sachs real estate private equity professional, described a platform that allows individual investors to buy fractional shares in single-family rental homes with a minimum investment of 250 dollars. His co-founder scaled Goldman Sachs's single-family rental platform from zero to one billion dollars in under 12 months with a team of three to four people. Mogul has completed 130 properties totaling over 600 units, with individual property values ranging from 500,000 to 2.5 million dollars. The minimum investment is 250 dollars, though the average investor commits 15,000 to 20,000 dollars, and each deal typically involves 20 to 80 investors.

Mogul projects an 8 to 12 percent annual dividend yield paid monthly from rental income, with property appreciation of 3 to 4 percent amplified by leverage producing an additional 4 to 8 percent return, for a total projected return of 15 to 20 percent. The bear case underwriting targets a 12 percent IRR. Rental income is typically reported as a passive loss for tax purposes because depreciation offsets it without affecting cash flow, meaning investors generally owe no taxes on rental distributions. Each raise capitalizes a maintenance reserve based on the age of home components, and a 12-month vacancy reserve is funded so the asset can sit vacant for up to a year without requiring additional investor capital.

Mogul charges investors a 5 percent upfront platform fee on the purchase price and earns an additional 1.5 to 2 percent paid by the seller at closing, plus approximately 2.75 percent interest on reserves held. A title insurance joint venture is expected to launch by end of the current quarter, adding roughly 9,000 to 12,000 dollars per asset. A property insurance vertical is projected to generate roughly 4 to 5 million dollars in recurring revenue in its first year. The typical investment period is five to seven years, with Mogul targeting next year to launch a secondary market where it will act as market maker and earn a bid-ask spread. On market selection, Mogul focuses on areas where net new supply is insufficient to meet demand. The company currently views Austin as attractive partly because a supply surge from 2020 to 2022 caused roughly a 25 percent price decline and has since scared away developers, reducing future supply competition.