All-In's Best Ideas Pitch Competition: 4 Investors Present Their Top Trades Live

Source: User-provided authorized transcript. The file contains 3,844 lines of heavily duplicated rolling-caption text.

Context: The transcript covers a live “Best Ideas Pitch” segment associated with the “besties”/All-In event environment. It includes an introduction explaining the inspiration for the format, a short presenter montage, four investment pitches, audience and host Q&A, post-pitch evaluation, live investment comments, and audience/judging results.

1. Host(s)/All-In participants — individual attribution is often not identifiable from the raw captions.
2. Aaron [surname rendered variously as Cowan/Cowen; verify] — MGM pitch.
3. Daniel/Dan [surname later rendered “Drafus”; verify] — “Talon Energy” pitch.
4. Oleg Nelman [verify spelling] — founder/managing director of “Eco1/Eco R1 Capital”; Actis Oncology pitch.
5. Kyle Samani — former Multicoin Capital executive/founder; “Geonet/Geodet” crypto/RTK pitch.
6. Gavin, David, Brad, Sam, “Timoth” and others participate in questions or evaluation; several surnames are absent or unclear.

The transcript contains extensive duplicated caption fragments, stutters and false starts. Those have been removed where they add no information.

Several names and technical terms are unstable or apparently mistranscribed. They are not silently corrected here. Important examples include:

• “Ira,” “Iris,” and “sone foundation.”
• “Serreta Capital.”
• Barry “Dillard” / Barry Diller.
• “Aaka” before the location is identified as Osaka.
• “Ararjan.”
• “SACE.”
• “Talon Energy.”
• “3M Island.”
• “Actis/Axis/Activis Oncology/Therapeutics.”
• “Eco1/Eco R1 Capital.”
• “radium 233.”
• “Amgen Sophi.”
• “glip ones.”
• “Geonet,” “GeoNet,” “Geodet,” “Jonet.”
• “Salana.”
• “geode geo d,” “GOD,” and “G.”
• several price targets and market-cap figures.

This document records what the speakers said. No outside verification has been performed at this stage.

A host explains that the concept was inspired by an earlier investment event connected to a man referred to in the transcript as “Ira,” who died of cancer at a young age. His family created what the transcript renders as the “sone foundation.” That organization held an event beginning at Lincoln Center where investment managers presented ideas.

The host recalls attending the Lincoln Center event in 2015 as a young venture investor and predicting that Amazon would become a trillion-dollar company. He says the room laughed at the prediction. He recalls David Einhorn responding:

“I know trillion dollar companies. This is not a trillion dollar company.”

The host says the Amazon call ultimately worked.

He returned in 2016 and presented Tesla, specifically saying “we picked the converts.”

In 2017, he considered his presentation his “magnumopus” and declared:

“AI is the future.”

However, he picked Box rather than Nvidia. He jokes that if he had selected Nvidia, he would have become “an Irish legend” and could have retired.

The stated purpose of the new pitch event is to recreate the spirit of the earlier event: give exposure to strong investment managers who may generate returns for their limited partners but do not receive mainstream distribution or frequent appearances on outlets such as CNBC.

A promotional montage introduces the “best ideas pitch.” Because the montage captions do not reliably identify each speaker, the following lines should remain unattributed unless independently matched:

• One speaker argues that anyone should be able to trade any asset, anywhere, at any time, 24/7, with only an internet connection and a phone.
• Another says they are building “a new financial system from the ground up.”
• Another predicts people will want to own equities, companies will create new products and applications, and long-short managers may make significant money from the resulting opportunities.
• A biotech investor says his fund searches for “unfollowed, unloved, misunderstood biotech companies.”
• Another speaker says nuclear power has experienced a “structural and permanent perception shift” and predicts both US political parties will increasingly support it.
• Another says the resulting outlook leaves him “maximum risk on.”

The presenter says he runs a roughly $4 billion New York firm rendered by the transcript as “Serreta Capital.” He says that before founding it, he ran the equities business for George Soros and later served as CIO for Steve Cohen. He says he has spent 29 years working in hedge funds.

He describes himself as a generalist investor. Although his fund owns technology stocks, he considers pitching a technology stock to this audience “absolutely completely stupid” because of the audience’s existing technology expertise.

Because poker is another theme of the conference, he chooses MGM.

He describes MGM as commonly perceived as a Las Vegas casino company with 13 Vegas properties, and says MGM and Caesars are the two largest owners of casino assets in Las Vegas.

He also claims Vegas conditions are beginning to improve.

His central point, however, is that the MGM thesis is not primarily about Las Vegas.

The presenter says MGM shares had been aggressively accumulated by Barry “Dillard”/Diller, who he says owned 26% of the company.

He says he prepared the presentation two weeks earlier. At that time, MGM shares were about $37. By the time of the presentation, the shares were in the high $40s because Barry had made a $48 bid for the company.

The presenter says directly that he would not sell his shares at $48.

He adds that MGM itself had also aggressively repurchased stock. He claims the company bought back roughly half of its float in six years, which he describes as highly unusual.

He also says MGM had become roughly 80% of Barry Diller’s NAV [verify exact meaning and figure].

The presenter asks why someone known for media/internet assets such as Expedia would make MGM such a concentrated investment. His conclusion is that MGM contains hidden assets whose value is not reflected in the market price.

The presenter estimates that MGM’s existing Vegas plus China assets alone are worth something in the “low 60s” per share.

Therefore, even before assigning value to the two future opportunities he focuses on, he sees upside from either the original roughly $37 share price or the roughly $48 bid level.

The first major hidden asset is MGM’s Japanese casino opportunity.

The transcript initially renders the location as “Aaka, Japan,” but the presenter immediately identifies it as Osaka.

He says Japan went through a process involving its prefectures and that Osaka was the only one that chose to open a casino.

He says the project is expected to open in 2030.

He argues that MGM mentions the project in company presentations but does not emphasize it sufficiently.

The presenter says Japan already has a significant gambling economy through activities including pachinko and horse racing.

He gives the following market-size comparisons:

• Japan: approximately $40 billion.
• Macau: approximately $30 billion.
• Las Vegas: approximately $10 billion.

His conclusion is that Osaka could become a very large gaming opportunity.

He estimates the property could generate roughly $2 billion of EBITDA [rendered “IBIDA”].

He says MGM owns 40% of the property and will also receive a management fee.

The presenter argues that Osaka is strategically well located for Chinese gamblers:

• From Shanghai, he says Osaka is closer than Macau or Singapore.
• From Beijing, he says it is about the same distance as Macau and much closer than Singapore.

He argues that Osaka could therefore become an attractive weekend gambling destination for people in Shanghai and Beijing.

He also considers Japan a more attractive jurisdiction for investors than Macau, describing Japan as a “first world country” and Macau as having issues that cause it to trade at a lower multiple.

His interpretation is that Barry Diller understands the casino business and wants to acquire MGM partly because of the Osaka opportunity.

He argues that Japan alone could more than double MGM’s stock value.

The presenter describes a branded MGM development in Dubai that includes Aria, MGM and Bellagio components.

He acknowledges that gambling is currently illegal in Dubai.

However, he says the development contains 300,000 square feet of space that could potentially be converted into a casino if gambling becomes legal.

His reasoning is that the space creates an embedded option: the project can operate without a casino, but if Dubai legalizes gaming MGM already has a large purpose-suitable area.

He also says “Wynn” [rendered “win”] is expected to open a casino in a place the transcript renders as “Ararjan”, around 45 minutes from Dubai, approximately two years later.

He argues that people who want to gamble while visiting Dubai would prefer a casino located directly in Dubai rather than travelling roughly 45 minutes away.

The presenter additionally refers to “a possibility of the war” and Dubai wanting to “reestablish themselves,” but this passage is [transcript unclear]. The substantive point appears to be his speculation that broader regional circumstances could strengthen Dubai’s incentive to legalize gaming.

He treats Dubai as an option, not a required component of the base case.

The presenter gives a simple sum-of-the-parts framework:

• Existing Vegas-related assets: roughly $60 per share [the earlier wording includes Vegas plus China in the low $60s].
• Japan: roughly $50 per share.
• Dubai if legalized: another $40–$50 per share.

He therefore argues MGM could be worth more than $100 and potentially $150 per share, and describes the opportunity as potentially a triple.

He considers Barry Diller’s $48 offer a useful floor or validation point because Diller already owns 26% and is making a firm bid around the current market level.

He characterizes Diller as a financial rather than strategic buyer, saying Diller is trying to make money from the investment. From this he concludes MGM is “in play.”

The presenter says the risk/reward is unusually attractive and tells existing shareholders not to tender their shares.

His thesis is effectively:

1. Existing assets provide meaningful value.
2. Diller’s bid creates downside support.
3. Osaka represents a substantial visible future catalyst.
4. Dubai offers additional upside that need not occur for the investment to work.

An audience/host question says Barry Diller has reportedly been spending significant time trying to reinvent MGM’s entertainment offering.

The questioner says the entertainment side may be under-monetized and could generate substantially more revenue per property/customer.

The presenter says he does not know the answer to the entertainment question. He adds that if Diller can improve that side of the business, it will help, but reiterates that his own thesis does not depend on Vegas entertainment.

He states:

“This is not really a Vegas. This is an Asian casino play…”

The presenter says MGM’s own presentations barely emphasize the Asian opportunity.

He also discusses looking for a “catalyst path”, saying that he previously worked at an organization rendered as “SACE” [verify]. Initially, he expected an investor day or similar corporate event to reveal the value. Instead, he says Barry’s takeover bid effectively “showed his cards” and became the catalyst.

Another question asks how MGM will scale customer credit and customer relationships when expanding internationally.

The presenter says MGM has a large customer database.

He assumes the Vegas properties already attract customers from China and Japan, giving MGM a pool of known customers it can market the Osaka property to.

He emphasizes MGM’s loyalty program.

To illustrate the importance of loyalty, he recounts a bad investment in the Rio casino in Las Vegas.

He says he and friends bought the Rio after Caesars merged with Eldorado and had to divest an asset. They acquired it at roughly $200 per square foot.

The mistake, he says, was failing to appreciate that separation from Caesars meant losing access to the Caesars loyalty system.

The transcript cuts off or becomes disordered before he fully explains the financial consequence, but his substantive lesson is clear: casino loyalty databases materially affect economics and repeat visitation.

An audience member asks why the market would still ignore Osaka if the casino was already approved in 2023.

The presenter compares the situation with Wynn’s expansion from Las Vegas into Macau.

He says the market did not begin assigning serious value to Wynn’s Macau asset until approximately three years before opening.

His historical rule of thumb is therefore that the market starts focusing on major future casino developments roughly three years before launch.

Because Osaka is approaching that window, he considers the present period opportunistically attractive.

Another question notes that Caesars left Dubai while waiting for a gambling license and asks why MGM should be different.

The presenter again frames Dubai as an option rather than part of the required investment case.

He says someone else built the Dubai development and MGM is operating it, and that the project intelligently reserved 300,000 square feet in case a casino is eventually allowed.

If gambling is never legalized, he still thinks the investor can more than double their money based on existing assets and Japan.

If Dubai is legalized, he thinks the investment could triple.

A host summarizes this as a “free option,” and the presenter agrees.

Daniel begins with what he calls the anatomy of a power cycle.

His framework:

1. During normal periods, electricity demand grows roughly in line with GDP.
2. Major technological breakthroughs can create unusually power-intensive demand.
3. Power demand temporarily spikes.
4. Once the new technology is widely adopted, demand growth tends to normalize.
5. Economies then enter an efficiency phase in which users work to conserve power.
6. Eventually another technological shift starts the cycle again.

As historical examples, he says appliances and air conditioning drove a previous surge in electricity demand.

He says demand normalized through the 1970s, 1980s and 1990s.

He describes the 2000s as an efficiency era driven by:

• LED lighting.
• Smart HVAC.
• Tinted windows.
• Smarter electronics.

He adds that the US also removed or relocated power-intensive industrial infrastructure such as aluminium smelters, often moving production to China.

The combined result, in his telling, was roughly two decades with effectively no power-demand growth.

He argues that this period has now ended and the US is entering another technology-driven power cycle.

Daniel emphasizes that his power thesis does not require AI.

He says:

“We do not need AI demand to keep the power markets incredibly tight for the next 20 years.”

His argument is that underlying demand and supply constraints are already strong enough to tighten power markets.

AI merely “turbocharges” the process and creates deeper shortages.

Daniel recounts appearing early in his career on a panel with Sam “Zel” [verify spelling], discussing opportunities in Mongolia.

Daniel was looking at a copper mine; Zell was looking at real estate.

He says Zell gave him an investment principle that shaped the rest of his career:

• Buy a hard asset below replacement cost.
• Ensure that society will need additional capacity of that asset in the future.
• Hold the existing asset as new-build costs rise or scarcity becomes obvious.
• Eventually sell at a premium to replacement cost when the market recognizes its value.

He says Zell cited Equity Office Properties as an example: acquired below replacement value and ultimately sold near the peak of the market.

Daniel applies this framework directly to “Talon Energy.”

Daniel says the company owns:

• 2 gigawatts of nuclear power.
• 6 gigawatts of natural-gas baseload power.

He says investors can buy the company at approximately a $25 billion enterprise value, while replacing those assets would cost approximately $45 billion.

Because the company also has debt, he argues that simply rerating the enterprise to replacement cost would produce more than a doubling of equity value.

He further argues that a true power scarcity cycle could eventually cause such assets to trade above replacement cost, creating greater upside.

Daniel argues that US power strategy should imitate the scale of China’s buildout.

He says that around 20 years earlier the US had approximately twice China’s power-generation capacity.

He claims China now has approximately three times the US capacity.

He connects electricity capacity directly to AI competition:

• If AI drives scientific breakthroughs, countries without enough AI capacity lose those breakthroughs.
• If AI powers robotics, countries without enough compute and electricity lose that productivity.
• If AI is important to military affairs and national security, falling behind becomes potentially existential.

He therefore calls the power buildout mandatory, not optional.

His formulation is:

“Make America great again. Copy China.”

Daniel compares an AI data center with a hydrocarbon refinery.

A refinery:

• Takes crude oil as an input.
• Uses expensive capital equipment.
• Produces gasoline or jet fuel.

A data center:

• Takes electricity as an input.
• Uses expensive capital equipment.
• Produces what he variously calls photons, tokens or intelligence.

He argues that electricity is therefore to AI infrastructure what crude oil is to a refinery.

The transcript renders his estimate of a data center as a “$50 billion per gawatt” capital-intensive asset; this figure requires verification.

Daniel says “Jensen” was recently quoted saying the world needs 1,000 times more power than it currently has.

Daniel explicitly qualifies this with “if that’s remotely true.”

If it is even directionally correct, he argues the system needs every available energy source:

• Hundreds of gigawatts of nuclear.
• Solar.
• “Orbital.”
• Other generation technologies.

His thesis is not that one technology wins, but that demand becomes so large that every viable source is needed.

Daniel says power plants, data centers, rockets and orbital infrastructure compete for many of the same critical minerals and industrial inputs.

Examples he gives include:

• Nickel superalloys needed for rockets and power infrastructure.
• Silver used in photovoltaic cells.

He predicts:

• Shortages across many inputs.
• Infrastructure delays.
• Continued scarcity of existing electricity capacity.

His conclusion is that existing power plants become unusually valuable because new supply cannot be built quickly enough.

Daniel points to PJM, which he expands as “Pennsylvania, Jersey, Maryland.”

He says the grid operator forecasts a need for 106 gigawatts of new power over the next ten years in that region.

He emphasizes that ten years is extremely short in infrastructure terms.

Consumers are accustomed to “internet time,” where food, transportation and other services arrive almost instantly. Power plants and transmission infrastructure operate on “geological time.”

He compares the 106 GW requirement with roughly the entire electricity consumption of Japan, saying the US is being asked to add that scale in one region alone.

Because of these physical constraints, he rejects expectations that thermal coal plants will simply be retired on schedule.

He argues existing coal generation will remain needed because the system cannot replace roughly 100 GW of capacity quickly enough.

Daniel says data-center operators and hyperscalers are “in a panic” to secure electricity.

They are attempting to sign long-term power purchase agreements (PPAs) with fixed pricing for periods such as 20 years.

He uses Microsoft and Constellation Energy as an example.

The transcript says Constellation owns the “3M Island nuclear reactor,” described by the speaker as the reactor involved in the accident that damaged nuclear power’s reputation.

Daniel says Microsoft encouraged Constellation to restart it by effectively offering a long-term floor substantially above prevailing electricity prices.

His example is:

• Market power: around $50/MWh.
• Microsoft: approximately $100/MWh for 20 years.

The exact facility, reactor history, contract terms and prices require verification.

Daniel jokes:

“3M Island brought to you by Microsoft Azure.”

He says regulators are becoming concerned about hyperscalers taking power away from ordinary grid customers, asking how homes will be heated if data centers consume so much capacity.

He calls this approaching collision “crunch time.”

Daniel gives three broad scenarios.

He says the stock is in the “high 300s” [later transcript figure conflicts with this].

If management simply operates the existing business and allows its Amazon data-center contract to “roll up,” he estimates approximately:

• $50 per share of annual free cash flow.
• Approximately 7× free cash flow at the stated share price.

He says good US infrastructure assets normally trade around 15×.

On that basis, he believes merely rerating the existing cash flow could approximately double the shares.

If the company signs more premium data-center contracts or electricity prices rise, he estimates around:

• $70 per share of recurring annual free cash flow.

He applies a 15× multiple.

The transcript then renders his resulting valuation as “that’s,50” [transcript unclear].

If the company begins adding generating capacity and captures approximately 4 GW of an estimated roughly 100 GW regional requirement, Daniel says annual free cash flow could exceed:

• $100 per share.

The transcript then says the shares are in the “high30s”, conflicting with the earlier “high 300s.” Both figures must be checked.

Daniel explicitly calls the stock a “good speculation” and tells the audience to buy shares, after which a host says “Not financial advice.”

Gavin asks how Daniel accounts for regulatory risk.

The concern is that:

• Consumers dislike rising electricity prices.
• AI is becoming politically contentious.
• Regulators may prevent data centers from pushing power prices too high.

Daniel says AI is necessary and society will find ways to make the system work.

He compares the electricity grid with a highway:

• At 4 a.m., the highway appears overbuilt.
• At rush hour, the same highway appears severely undersized.

Electricity works similarly because the grid is only acutely stressed during a limited number of peak hours.

His proposed bridge solution:

1. Sign PPAs with data centers.
2. Require data centers to install substantial battery capacity.
3. Add peaker generation for the most stressed hours.
4. Use this as a temporary solution until more permanent generation is built.

He predicts consumer electricity bills will rise, but believes the increases can remain “relatively under control.”

He expresses confidence that:

“Human ingenuity is going to win here.”

An audience question asks whether the thesis depends on behind-the-meter co-location being approved, or simply on clean firm baseload power remaining scarce regardless of how electricity is delivered.

Daniel says it is the latter.

His valuation does not require a particular regulatory outcome for behind-the-meter arrangements.

He repeats his three stages:

• $50/share cash flow with essentially no change.
• $70/share with more attractive power arrangements.
• $100+/share if the company also adds generating capacity.

Brad asks about competition from:

• Fuel cells.
• Gas turbines.
• “aerodyit” turbines [transcript unclear].
• Orbital compute.
• Other independent power producers.

Daniel’s answer is that all of them are needed.

He describes fuel cells and Caterpillar/Solar turbines as useful bridge technologies even though their levelized cost of energy may be extremely high.

His reasoning is that if an operator has already spent $50 billion on a data center, allowing it to sit idle for three years while waiting for a baseload combined-cycle gas turbine is economically worse than temporarily paying very high electricity costs.

Therefore, developers will use expensive bridge technologies where necessary.

He again supports an “all of the above” strategy, including the possibility of orbital data centers.

An audience question asks the correct terminal multiple if the business changes from merchant electricity exposure toward contracted infrastructure.

A related question asks how much EBITDA must become contracted before the market should rerate the company.

Daniel says this is an important issue he could not fully cover in the six-minute pitch.

He explains why he uses 15× as a rough blended multiple:

• Contracted cash flows resemble bonds.
• Bond-like income deserves a higher multiple because it is predictable.
• If Treasuries yield around 5%, a pure bond-like stream mathematically corresponds to roughly a 20× baseline before adjustments for growth/risk.
• Merchant electricity revenue is less predictable and deserves a lower multiple.

His rule:

• More contracted revenue → higher multiple.
• More merchant/spot-market exposure → lower multiple.
• 15× is his simplifying blended assumption.

The presenter identifies himself as Oleg Nelman [verify], founder and managing director of “Eco1 Capital,” a San Francisco value-oriented biotech fund founded approximately 13 years earlier.

He says this is his 25th year investing in biotechnology.

He previously spent 11 years at another fund before launching “Eco R1” in 2013 with approximately $13 million.

He claims:

• Investors have made approximately 10× since inception.
• The fund compounded at approximately 20% annually.
• Current assets under management are approximately $2.5 billion.
• Many long-term investors are themselves biotech entrepreneurs.
• The fund recently reopened to investors for the first time in four years.

Oleg jokingly says investing in biotech is generally a horrible idea, ranking somewhere among movies, wineries and SPACs.

He describes biotech as resembling a casino, where most investors are playing slot machines.

Eco1 instead sees itself as a poker player.

The fund deliberately is not primarily managed by PhDs or MDs because Oleg does not want his team to become emotionally attached to the scientific story.

His philosophy is:

“We don't want to fall in love with the science. We fall in love with the riskreward.”

The fund searches for a margin of safety and wants to monetize “other kids science projects.”

Oleg frames the pitch as part of the “war on cancer.”

He says military language has been used for cancer since the early 1970s and President Nixon’s National Cancer Act.

He argues the analogy fits because both warfare and oncology seek to:

1. Find the enemy.
2. Identify the best weapon.
3. Kill the enemy.
4. Minimize collateral casualties.

He compares early surgery and radiation with a medieval siege:

• Destroy the castle.
• Burn the surrounding village.
• Hope the enemy was in the rubble.

He says chemotherapy evolved from observations that mustard gas killed rapidly dividing tissue.

Because tumor cells divide rapidly, chemotherapy attacks them.

However, hair, skin, gastrointestinal and bone-marrow cells also divide rapidly, so chemotherapy damages healthy tissue as well.

He compares targeted therapy to a GPS-guided munition.

Rather than attack every dividing cell, the drug identifies a particular molecular command-and-control system.

The problem is that cancer can adapt through resistant mutations.

He says modern immunotherapy recruits the body’s own T cells as “local allies.”

It can be spectacular when successful but depends heavily on the tumor microenvironment.

His preferred analogy for modern radiopharmaceuticals is a swarm of micro-drones:

• Navigate through the bloodstream.
• Recognize a molecular target.
• Deliver a radioactive payload.
• Produce a very limited blast radius.

He says the target blast radius is around 100 microns, which he describes as approximately the diameter of a single cell.

He calls the concept:

“An autonomous assassination with the force of a bunker buster and minimum collateral damage.”

Oleg names the company Actis Oncology, ticker AKTS [both require verification].

He gives the following figures:

• Market capitalization: approximately $1 billion.
• Enterprise value: approximately $500 million.
• Cash runway: more than three years.
• Company age: approximately five years.
• Recent IPO: approximately $300 million.
• IPO demand: reportedly 18× oversubscribed.
• Eli Lilly order/backstop: transcript renders amount as “hund00 million” [transcript unclear].

He says the company’s platform uses small or “mini” proteins capable of carrying radioactive payloads.

The stated design goals are:

• Carry different radioactive payloads.
• Target multiple biological markers.
• Remain small enough to clear the body rapidly.
• Reduce side effects.

A key advantage, according to Oleg, is that doctors can use imaging to see whether the therapy actually reaches the tumor in early clinical trials.

He considers this a major de-risking mechanism because investors can observe target engagement before waiting for later-stage efficacy data.

Oleg says Actis further reduces risk by initially pursuing already validated biological targets.

He names:

• Nectin-4, which he associates particularly with bladder cancer.
• B7-H3, which he describes as being expressed across many major solid tumors.

He specifically names:

• Prostate cancer.
• Colorectal cancer [transcript rendered “colurectyl”].
• Lung cancer.

He says clinical trials began the previous year.

The company is publicly guiding to initial data from both programs in 2027, with the Nectin-4 program potentially reporting as early as Q1 2027.

If either program demonstrates a positive signal, Oleg expects investors to assign value not only to that drug but also to the broader mini-protein platform.

He calls this the biotech “holy grail”: investors paying for the future promise of an entire platform rather than only one product.

Oleg says major pharmaceutical companies are building radiopharmaceutical capabilities and need additional assets.

He names:

• Bristol [verify full company].
• Novartis.
• Bayer.
• Eli Lilly.

He claims there has been roughly $15 billion of M&A and dealmaking in radiotherapy/radiopharmaceuticals during the previous few years.

He believes the field is still in its “early innings.”

Oleg argues radiopharmaceuticals are particularly difficult to copy.

He says:

• Traditional generic manufacturers generally do not operate in this area.
• Radioisotope supply creates additional manufacturing complexity.
• The relevant class is effectively “off limits to China.”

He therefore sees a stronger competitive moat than in many other biotechnology categories.

This China/radioisotope claim becomes a major subject of the Q&A and requires careful verification.

Oleg delivers a parody of a pharmaceutical safety advertisement.

Although several lines are garbled, the substantive points include:

• The investment is not suitable for everyone.
• Biotech ownership can create anxiety and loss of sleep.
• Biotech stocks can suffer sudden and severe declines.
• Competitor developments can create volatility.
• Safety problems have occurred in other development programs.
• No safety issue may be visible today but one could emerge later.
• The mini-protein delivery technology is not yet proven commercially.
• Actis has no marketed product and therefore lacks product revenue.
• Equity offerings/dilution may occur.
• His tongue-in-cheek response to declines or secondary offerings is to call the broker/company and consider buying more.

The section is intentionally comedic but acknowledges real clinical, financing and volatility risks.

Oleg says conventional biotech valuation is difficult because risk-adjusting and discounting early-stage drug programs can rapidly reduce present value toward zero.

For early-stage opportunities, his fund instead triangulates across possible outcomes.

He estimates Actis could be worth approximately:

• $10 billion, or
• $200 per share

if even one of its programs reaches the market.

He says the company has “a lot of outs,” meaning multiple programs, platform value and possible strategic interest could produce a successful outcome.

A questioner asks why China cannot compete in radioisotopes.

The response says Actis uses actinium.

The speaker claims actinium is manufactured from “radium 233” generated by US nuclear programs in the 1950s and 1960s.

He says this material is effectively a waste product from those programs and is not similarly available in China because China used a different nuclear pathway involving enriched uranium and plutonium.

These technical claims should be independently checked before publication.

Another participant raises what the transcript renders as the “Amgen Sophi Supreme Court case”, saying it enabled competitors to make small amino-acid changes around composition-of-matter patents and arguing China has copied many US and European biologics through similar strategies.

Oleg responds that radiopharmaceuticals require locally sourced manufacturing and says his team has not seen meaningful Chinese competition in the US market.

When asked whether China could simply acquire the necessary isotopes after a positive Actis readout, Oleg says China could likely produce something for its domestic market, but his team has not seen a pathway for such products to transfer into the US market.

A questioner asks why the market has discounted Actis since its IPO.

Oleg describes the explanation as classic biotech behavior.

He says the stock has traded roughly flat since the IPO because biotech investors are extremely short-term oriented.

Even though the company is only around eight or nine months from data, he says that is still too distant for many biotech traders.

His expectation is that investors will begin accumulating shares during the second half of the year as the first-quarter data release approaches.

Gavin says he previously ran a biopharmaceutical fund shortly after sequencing of the human genome.

At that time, investors expected genome sequencing to quickly trigger an explosion of personalized therapies.

He believes progress over the following roughly 25 years fell short of those early expectations.

His hypothesis is that the genome represents a search/problem space too large for either unaided human reasoning or traditional human-written software.

He expects AI to unlock more revolutionary therapies.

He asks Oleg whether people currently alive could eventually see average lifespan in developed countries move beyond:

• 100 years.
• 125 years.
• Potentially 150 years.

Oleg says he would “take the over.”

One reason he gives is the rise of obesity drugs, rendered by the transcript as “glip ones.”

Oleg argues that caloric restriction is one of the few interventions supported by data as extending life and that obesity drugs effectively produce caloric restriction.

He further says these drugs have normalized the idea that people can regularly inject pharmaceuticals as part of healthy living.

He views this as only the beginning of a broader pharmaceutical longevity trend.

An audience member asks whether falling launch costs could make microgravity itself a therapeutic variable because cancer cells may behave differently in low-gravity environments.

Oleg says it is an interesting question but probably not applicable to this particular Actis investment.

Another audience question asks what AI-driven technological development might disrupt precision radiotherapy or alter drug development and early cancer screening.

Oleg says Actis has a small internal “skunks work” AI project.

He says many biotech companies own proprietary datasets and hope that AI can find additional insights from them.

Actis has a small group of PhD/data-science researchers examining whether its mini-protein datasets can be exploited computationally.

He does not claim this AI work is currently central to the valuation.

A question notes that CAR-T has struggled in solid tumors and asks whether personalized peptide-based immunotherapies are making progress.

Oleg says a promising recent pancreatic-cancer drug from a company rendered as “Revmed” is still essentially another targeted therapy.

He says there has not yet been a huge amount of progress from peptides.

Another participant asks about “Droins” [transcript unclear], described as right-handed proteins capable of penetrating solid tumors.

Oleg responds that a central advantage of Actis’s mini proteins is precisely their proposed size: small enough to penetrate tumors and deliver their radioactive payload internally.

The pitch then ends.

A host remarks that someone appears to have bought the stock while Oleg was speaking and that it had risen approximately 6% during the presentation.

Kyle Samani says he founded Multicoin Capital approximately eight and a half years earlier and stepped down a few months before this event.

He says he is particularly known for leading all three pre-launch investment rounds in “Salana” [verify] before the network launched in 2020.

He describes himself as a long-time crypto investor.

He says the pitch sits at the intersection of crypto and AI.

He also thanks David “Sax” [verify], saying David seeded Multicoin early in its history.

Kyle begins by describing ordinary GPS positioning errors, such as a phone displaying a user on the wrong side of a road or facing the wrong direction.

He says Geonet uses RTK — real-time kinematics.

His stated accuracy comparison is:

• Ordinary GPS: approximately 2 metres.
• RTK/Geonet: approximately 2 centimetres.

That represents roughly 100× greater location accuracy.

He argues this precision is particularly useful for robotics, with drones as an obvious example.

Kyle calls Geonet the world’s largest and fastest-growing RTK network.

He compares it with three incumbents:

• Trimble.
• Hexagon.
• Topcon.

He says these companies have spent roughly 20–30 years developing RTK networks and collectively operate approximately 12,000 base stations.

Geonet, by contrast:

• Was founded in 2021.
• Began building the network in 2022.
• Has roughly 22,000 nodes.
• Is approximately twice the size of the three incumbents combined.
• Is active in roughly 150 countries.
• Operates in more than 11,000 cities.
• Covers approximately 80% of the global population, excluding some sanctioned countries.

All of these figures require verification.

Kyle says Geonet grew much faster than incumbent RTK networks because it uses crypto incentives.

Instead of a telecom company such as AT&T or Verizon financing and installing every base station, individual participants can do it.

A participant can:

1. Buy a base station for a few hundred dollars.
2. Put it on a house or business roof.
3. Broadcast RTK correction data/radio signals.
4. Receive token rewards.

Kyle says operators are paid in “geode” tokens.

He sees token incentives as the mechanism that enabled the network to bootstrap physical infrastructure rapidly without Geonet financing the entire capex itself.

He claims the network now has coverage not only across major US metros but across much of rural America as well.

Kyle says the USDA launched a program encouraging farmers and ranchers to adopt precision-agriculture technologies, including RTK.

He claims the USDA now subsidizes farmers and ranchers adopting high-precision agriculture and that much of this technology is powered by Geonet.

He shows a machine made by a company rendered as Burrow.

The robot transports grapes and could theoretically move other agricultural material.

Kyle expects these machines to proliferate as computer vision, processors and batteries improve.

Kyle discusses a John Deere system rendered as:

Global Unmanned Spraying Systems (GUS).

He says these autonomous machines drive around agricultural properties spraying pesticides and other treatments.

He says he confirmed that morning that wineries in Napa were using these vehicles.

He jokes that wine consumed at the event may have come from grapes serviced by a GUS machine using Geonet.

Kyle identifies TomTom as a Geonet customer.

He claims TomTom supplies mapping/navigation technology to almost every autonomous-vehicle program worldwide, with a few possible exceptions.

He says TomTom uses Geonet data to continuously increase the accuracy and precision of its global maps.

Kyle sees robotic lawnmowers as an emerging mass-market example of physical AI.

He claims approximately one million robotic lawnmowers are expected to be sold during the year.

He names:

• Yarbo.
• Sunseeker.
• Other manufacturers.

He claims these systems use Geonet.

Kyle says DJI, which he calls the world’s largest drone manufacturer, is a Geonet customer.

He qualifies this by saying Geonet is not used in every DJI product but is present across many models.

He says DJI is already producing substantial Geonet traffic.

Kyle predicts that if DJI’s US position declines and American drone manufacturers grow, most or all of those companies are likely to join Geonet as well.

This is explicitly his prediction, not an established fact.

Kyle describes Geonet as a natural network-effect business.

His analogy is telecommunications:

• The service becomes more valuable as geographic coverage expands.
• Customers prefer networks available everywhere.
• Large telecom infrastructure networks historically tend toward concentration.

He therefore predicts Geonet could become a natural monopoly.

He believes it already benefits from:

• The largest network.
• The fastest growth.
• The lowest cost structure.

He attributes the low cost to the decentralized deployment model: individual operators, rather than the company itself, finance many base stations.

Kyle initially says the business had just crossed approximately $1 million in annualized run-rate revenue and was growing more than 3× year over year.

He predicts it will more than triple again in the next twelve months.

However, moments later he refers to approximately $11 million in revenue, saying 80% equals roughly $8.8 million of token purchases.

These figures cannot both describe the same annualized revenue base without additional explanation.

Do not reconcile them silently. Verify whether the first figure should be $11 million or whether the two figures refer to different revenue measures.

Kyle says approximately 80% of Geonet revenue is used for open-market purchases of the Geonet token.

He says the relevant blockchain addresses are public, making purchases visible in real time on the network rendered as “Salana.”

Using the transcript’s $11 million revenue figure, he says approximately $8.8 million per year is being used for token purchases.

He says the remaining 20% is enough to pay:

• R&D.
• Engineers.
• Operating costs.
• Expansion of the business-development team.

This capital efficiency is a central part of his thesis.

Kyle says the customer base itself is tightly interconnected.

People at John Deere know people at DJI, who know people at TomTom, allowing adoption to spread through a relatively small professional ecosystem.

He also claims individual customers dramatically increase spending after adoption.

His figures:

• First year: roughly $60,000 annual spend.
• After two years: roughly $170,000 annual spend.

He describes that as roughly a 3× increase.

He also says the number of new customers increased approximately 5× in the previous year.

Combining new-customer growth with increased spending by existing customers, he sees a clear path to revenue growing more than 3× again.

Kyle summarizes the thesis as:

• World’s largest RTK network.
• Fastest-growing network.
• Strong network effects.
• Potential natural monopoly.
• Very low decentralized infrastructure costs.
• Approximately 3× year-over-year growth.
• Recognizable customers.
• Large tailwind from robotics and physical AI.
• Revenue-linked token purchases.

He says the token trades 24/7.

The transcript renders its ticker as “geode geo d.”

He tells interested investors to obtain a crypto wallet on the network rendered as “Salana” and buy the token.

Asked for the token valuation, Kyle says it trades at approximately:

• $150 million fully diluted.

He says crypto-price websites may show only:

• $60–$70 million

because not all tokens are circulating.

He then says:

“the fully diluted number is about 1 million.”

This directly conflicts with the earlier $150 million statement and is almost certainly a transcription or speaking error, but must be verified rather than silently corrected.

A host asks whether there is a real company behind the project or merely an opaque offshore token structure.

Kyle says:

• The operating team is US-based.
• The company/team is in San Francisco.
• The CEO is rendered as Mike Horton.
• The CEO has prior IoT/smart-device experience.

Asked whether investors should own corporate equity or the token, Kyle says:

“you should own the token because I own a lot of the token…”

He discloses that he owns substantial tokens but no company equity.

He says the operating company deals directly with customers such as John Deere and DJI and has a contractual relationship with the Geonet Foundation requiring 80% of revenue to be used to buy tokens in the open market.

He also says Multicoin previously led a venture investment round in Geonet.

Asked whether he likes Helium — described by the questioner as something like Geonet for 5G — Kyle says yes.

He says he led Multicoin’s investment in Helium roughly six or seven years earlier and remains a long-term believer.

He also says Helium had significant news that morning.

An audience question notes that many decentralized physical-infrastructure projects have failed because token incentives eventually lose value.

Kyle says Geonet differs because it has real customer revenue and uses that revenue to purchase tokens.

He cites:

• Approximately $8.8 million in annualized token purchases.
• Approximately $150 million valuation.
• Expected approximately 3× growth.

He calls it:

“an unbelievably cheap asset.”

He attributes the low valuation partly to a crypto bear market and lack of attention.

A host characterizes the token as effectively a securitized interest in customer cash flows.

Kyle largely agrees.

The hosts summarize it as a revenue-share token, and Kyle agrees that the 80% mechanism is the relevant value-accrual path.

The more customers such as John Deere pay the operating company, the greater the open-market token purchases.

A question asks whether value accrues to the equity or token and how the arrangement fits US securities law and the contemplated CLARITY Act.

Kyle says value accrues mainly to the token because 80% of revenue is directed toward purchases.

The remaining 20% finances company operations.

On legal classification, Kyle explicitly says:

“I'm not a lawyer…”

He does not claim the structure definitely satisfies the CLARITY Act.

He says he has been heavily involved with the Act, believes its passage would be favorable for Geonet, is optimistic, and is “not too worried about it.”

This is Kyle’s opinion and not a legal conclusion.

A questioner with agricultural-technology experience notes that John Deere, Trimble, Leica, Topcon and others already operate or sell RTK infrastructure.

He asks why John Deere would outsource the service instead of using its own network.

Kyle answers primarily in terms of capex versus opex.

He claims Geonet costs only roughly one-third to one-quarter as much as deploying proprietary infrastructure.

It is also already geographically available.

Therefore, John Deere can sell a tractor or autonomous system without separately deploying new local RTK infrastructure, reducing friction and shortening the sales cycle.

A questioner describes a well-funded effort involving “microats” [likely microsatellites; verify] that could potentially provide sub-centimetre location accuracy and eventually replace both GPS and ground-based RTK.

Kyle dismisses the threat primarily on cost.

His argument:

• Ground-based Geonet stations cost only a few hundred dollars.
• Launching satellites is much more expensive.
• He therefore sees little chance satellites can match Geonet economically.

Another speaker disagrees, arguing a sufficiently dense low-Earth-orbit network might eventually reach RTK precision and replace terrestrial systems.

The discussion suggests a possible hybrid architecture involving:

• GEO satellites.
• LEO satellites.
• Ground base stations.

Another participant believes LEO alone could eventually replace terrestrial RTK if enough satellites are deployed.

Kyle does not accept that as the likely economic outcome.

He adds one more argument: satellite communication consumes more device energy than talking to a nearby terrestrial base station.

For power-sensitive devices such as drones, he therefore expects ground infrastructure to remain preferable.

Kyle is asked why he prefers this project over crypto tokens tied to distributed AI training or inference.

He says his interest partly comes from having known and followed the Geonet founder for years.

Regarding distributed training:

• Many projects are attempting it.
• Kyle is “pretty skeptical.”
• He does not think most will work.

Regarding distributed inference:

• He considers it technically possible.
• He invested money in the idea several years earlier.
• The results have been weaker than hoped.
• He describes the outcome as working, but “not A+.”

This distinction is important: Kyle is bullish on decentralized physical location infrastructure while being much more cautious about decentralized AI computation.

An evaluator rendered as “Timoth” says he follows what the transcript renders as the “Stan Ducken Miller” school of investing [verify name].

His basic principle is that people care more about an investment thesis when they have money at risk.

He says:

“if you don't have any skin in the game, you don't care.”

He likes all four pitches partly because they are actionable investment ideas rather than abstract commentary.

The evaluator says all four may offer asymmetric returns, but their downside risk and appropriate position size differ dramatically.

He likes Kyle’s thesis but says the token’s liquidity is too low for him to deploy large amounts.

He doubts he could invest even $1 million without moving the market.

He suggests beginning with perhaps $10,000, $20,000 or $30,000 and accumulating gradually.

He says Talen/Talon shares could absorb tens of millions of dollars without materially moving the market.

He sees biotech as having a discontinuous outcome:

• Potentially close to zero.
• Potentially 10×.

A participant interjects that Lilly could bid for it.

He regards MGM and Talon as the ideas that could support much larger position sizes.

He characterizes Actis and Geonet more like lottery-ticket allocations because of their higher probability of extreme loss.

Gavin says the pitch format was excellent and should be expanded the following year.

He suggests the hosts could potentially create an All-In basket or ETF containing the ideas.

He agrees to return the next year and pitch a silicon and memory supercycle.

From a pure risk/reward standpoint, Gavin considers MGM the best.

His reasoning:

• Barry Diller’s bid caps some downside.
• Japan offers substantial upside.
• Dubai offers an additional option.

Gavin also finds Talon highly compelling.

His main concern is regulatory risk around AI and electricity.

Potential threats include:

• Government price caps.
• A different administration.
• A different Congress.
• New laws changing the economics of electricity supply to data centers.

He emphasizes that this risk may have nothing to do with Talon’s own operational performance.

Outside such policy intervention, he considers the pitch very strong.

Gavin considers Actis compelling because a successful biotech platform can be far more valuable than a single drug.

If a company has a broadly reusable targeting or drug-delivery mechanism, he says it can create rare outcomes exceeding $100 billion.

He sees this platform possibility as Actis’s most attractive feature.

Gavin says he found the Geonet discussion fascinating but does not invest in crypto.

He compares crypto with snowboarding.

He has spent decades learning public-equity investing, accumulating experience, mistakes, pain and “scars,” just as someone might spend thousands of hours learning to ski.

Starting crypto would be like abandoning skiing to learn snowboarding from scratch.

He is not condemning crypto; he simply does not want to rebuild his investment expertise in an unfamiliar market.

His joke is that snowboarders are welcome as long as they do not “go sideways down the mountain and ruin the powder.”

David broadly agrees with the positive MGM thesis.

He says MGM might offer roughly 3× upside.

He adds a factor that Aaron did not include: better monetization of Vegas entertainment.

David recounts hearing from a Vegas contact that the Venetian earns approximately $1 million of incremental EBITDA per day when there is a show at the Sphere.

His inference is that compelling entertainment materially increases gambling revenue.

He also says he has separately heard that Barry Diller is spending considerable time redesigning entertainment at MGM properties.

If true, this could create another source of upside even without Japan or Dubai.

David likes the downside floor created by the bid and sees a possible 3× outcome in roughly two years, even if the takeover itself fails and MGM remains independent.

David thinks Talon may offer 3× to 5× upside, but over approximately eight years, making it a longer-duration thesis than MGM.

He also believes Talon is more interest-rate sensitive.

His reasoning is that contracted power-purchase agreements resemble long-duration financial assets, so their valuation multiple depends partly on interest rates.

If rates rise sharply, Talon’s assumed 15× multiple could compress.

He presents this as an additional downside scenario beyond the regulatory risks already discussed.

David says he invests in another company with a protein-conjugate technology that shows strong solid-tumor delivery.

Because other therapeutic modalities are developing, he worries Actis could face technological competition.

He also considers the China copying risk legitimate based on his experience in biotech.

However, he sees major upside if Actis produces strong clinical data.

He speculates that Lilly could bid for the company within six months after a good readout.

His downside estimate is severe:

• Approximately 50%–75% loss after a bad readout, emergence of a superior modality, or relevant China competition.

His ranking is:

1. MGM.
2. Talon.
3. Actis.
4. Geonet/Geodet.

For Geonet, David believes satellite systems are likely eventually to replace both RTK and conventional GPS during the next decade. This directly conflicts with Kyle’s terrestrial-network thesis.

Another evaluator divides the four ideas into two groups.

• AKTS/Actis.
• “GOD”/Geodet.

These could deliver enormous returns but, in his view, also have meaningful probability of going to or near zero if the underlying technology fails.

• MGM.
• Talon.

He summarizes their durability humorously:

“people will always gamble and leave the lights on.”

He says he put $200,000 into each in real time.

Another participant says he bought half of that person’s action because he did not have a Robinhood account and otherwise would have needed to call his office.

One speaker says the portfolio was already up approximately 7% during the session.

Another ranking is stated as:

1. MGM.
2. Talon.
3. “G.”
4. AKTS.

The precise speaker responsible for that ranking is unclear from the captions.

The hosts bring the four presenters back to the stage.

There is extended joking around a 3D-modeled trophy depicting two men hugging, repeatedly called an “extremely heterosexual alpha male” award.

The hosts joke about deliberately extending a hug and oxytocin release.

This material is entertainment rather than part of the investment analysis, but it is part of the event record.

The audience vote is said to include approximately 150 votes.

Results:

1. Daniel/Dan — Talon: 50%.
2. Aaron — MGM: 24%.
3. Oleg — Actis: 21%.
4. Kyle Samani — Geonet: 5%.

The audience therefore selects the Talon pitch as its winner.

The hosts then announce their own ranking:

1. Aaron / MGM — first.
2. Dan / Talon — second.
3. Oleg / Actis — third.
4. Kyle / Geonet — fourth.

This reverses the top two positions from the audience vote.

The segment therefore concludes:

“MGM wins.”

The Stage 1 record above includes:

• Event origin and presenter-selection rationale.
• Introductory montage.
• MGM pitch and all substantive MGM Q&A.
• Talon Energy pitch and all substantive Talon Q&A.
• Actis Oncology pitch and all substantive Actis Q&A.
• Geonet/Geodet pitch and all substantive Geonet Q&A.
• Host/evaluator comparisons.
• Liquidity and sizing discussion.
• Regulatory, technological and competitive counterarguments.
• Live-purchase remarks.
• Audience voting.
• Final judge ranking.

No substantive section of the supplied transcript has intentionally been omitted.

Speaker: Host/presenter [identity unclear]
Exact quote: “I know trillion dollar companies. This is not a trillion dollar company.”
Context: Recalled statement attributed to David Einhorn in response to the speaker’s 2015 Amazon trillion-dollar prediction.
Topic: Contrarian investing / Amazon.

Speaker: Host/presenter [identity unclear]
Exact quote: “AI is the future.”
Context: Recalling his 2017 investment presentation, after which he says he selected Box rather than Nvidia.
Topic: AI investing.

Speaker: Aaron [surname verify]
Exact quote: “I would not sell my shares to him.”
Context: Rejecting Barry Diller/Dillard’s reported $48 MGM bid as too low.
Topic: MGM valuation.

Speaker: Aaron
Exact quote: “So we think the stock is a triple.”
Context: After combining his estimates for MGM’s existing assets, Osaka and potential Dubai casino option.
Topic: MGM upside.

Speaker: Aaron
Exact quote: “This is not really a Vegas. This is an Asian casino play…”
Context: Responding to a question about improving entertainment at MGM’s Vegas properties.
Topic: Core MGM thesis.

Speaker: Aaron
Exact quote: “it's free option.”
Context: Agreeing with a host’s description of the Dubai casino possibility as optional upside rather than necessary to the base case.
Topic: Dubai optionality.

Speaker: Daniel/Dan
Exact quote: “We do not need AI demand to keep the power markets incredibly tight for the next 20 years.”
Context: Arguing that structural electricity scarcity exists independently of AI.
Topic: Power-market thesis.

Speaker: Daniel/Dan
Exact quote: “AI demand just turbocharges.”
Context: His description of AI as an accelerator rather than the foundation of the electricity thesis.
Topic: AI electricity demand.

Speaker: Daniel/Dan, quoting Sam “Zel”
Exact quote: “If you can buy an asset, a hard asset at below replacement cost for an asset that's going to be needed in the future where we're going to need to build new capacity of that asset. Then you buy that asset at the discount to replacement cost.”
Context: The hard-asset investment framework Daniel applies to Talon.
Topic: Replacement-cost investing.

Speaker: Daniel/Dan
Exact quote: “Make America great again. Copy China.”
Context: Calling for a much larger US electricity-generation buildout.
Topic: US-China power competition.

Speaker: Daniel/Dan
Exact quote: “I look at a data center as the exact same thing as a refinery.”
Context: Comparing electricity-to-intelligence conversion with crude-oil refining.
Topic: AI infrastructure economics.

Speaker: Daniel/Dan
Exact quote: “we need every single source of power that you can imagine.”
Context: Responding to the scale of future electricity demand.
Topic: All-of-the-above energy strategy.

Speaker: Daniel/Dan
Exact quote: “Human ingenuity is going to win here.”
Context: Answering concerns about regulation, peak demand and consumer power bills.
Topic: Regulatory/power constraints.

Speaker: Daniel/Dan
Exact quote: “We need all of it. We need all of it.”
Context: Responding to questions about fuel cells, turbines, orbital compute and competing power solutions.
Topic: Energy scarcity.

Speaker: Oleg Nelman [verify]
Exact quote: “Generally speaking, investing in biotech companies is a horrible idea sandwiched somewhere between movies, wineries, and spaxs.”
Context: Opening his biotech pitch with a warning about sector risk.
Topic: Biotech investing.

Speaker: Oleg
Exact quote: “We don't want to fall in love with the science. We fall in love with the riskreward.”
Context: Explaining Eco1’s deliberately value-oriented approach.
Topic: Investment discipline.

Speaker: Oleg
Exact quote: “find the enemy, figure out the best weapon to kill them, and have minimal unwanted casualties along the way.”
Context: Explaining his warfare analogy for cancer treatment.
Topic: Oncology.

Speaker: Oleg
Exact quote: “An autonomous assassination with the force of a bunker buster and minimum collateral damage.”
Context: Describing targeted radiopharmaceutical delivery.
Topic: Radiopharmaceuticals.

Speaker: Oleg
Exact quote: “This is the holy grail in biotech, getting value simply for the promise of what might come.”
Context: Explaining how successful lead data could cause investors to value Actis’s entire platform.
Topic: Biotech platform valuation.

Speaker: Oleg
Exact quote: “We think Actis could be worth 10 billion or $200 per share if even one of their programs makes it to market.”
Context: Oleg’s high-case valuation framework.
Topic: Actis valuation.

Speaker: Oleg
Exact quote: “I would take the over on that”
Context: Responding to whether people alive today could see average lifespan exceed 100–150 years.
Topic: Longevity.

Speaker: Kyle Samani
Exact quote: “you can localize your location down to about 2 centimeters”
Context: Comparing RTK precision with ordinary GPS.
Topic: RTK / robotics.

Speaker: Kyle Samani
Exact quote: “telecoms naturally form monopolies historically. I think the same is likely to be true here.”
Context: Explaining why he believes Geonet has natural-monopoly characteristics.
Topic: Network effects.

Speaker: Kyle Samani
Exact quote: “80% of it is being used to make open market purchases of Geo tokens.”
Context: Explaining the token value-accrual mechanism.
Topic: Token economics.

Speaker: Kyle Samani
Exact quote: “It's an unbelievably cheap asset.”
Context: Comparing the stated token valuation with the amount being spent on token purchases and expected growth.
Topic: Geonet valuation.

Speaker: Kyle Samani
Exact quote: “I'm not a lawyer”
Context: Qualifying his answer about the CLARITY Act and securities-law treatment.
Topic: Crypto regulation.

Speaker: Kyle Samani
Exact quote: “The distributed training stuff, there's a whole bunch of people trying it. I I'm pretty skeptical. I don't think any of it's going to work.”
Context: Explaining why he prefers the physical-infrastructure thesis to many decentralized AI-compute tokens.
Topic: Decentralized AI.

Speaker: “Timoth” [verify]
Exact quote: “if you don't have any skin in the game, you don't care.”
Context: Explaining why he values actionable investment pitches.
Topic: Investment conviction.

Speaker: Gavin
Exact quote: “I feel about crypto exactly the way I do about snowboarding.”
Context: Explaining why decades of equity-investing experience make him reluctant to learn crypto from scratch.
Topic: Circle of competence.

Speaker: Host/evaluator [identity unclear]
Exact quote: “people will always gamble and leave the lights on.”
Context: Grouping MGM and Talon as the two ideas with stronger downside protection.
Topic: MGM/Talon durability.