The Quest for the Holy Grail

In 221 BC, Qin Shi Huang, China’s first emperor, funded expeditions and alchemists to go searching through his kingdom for the elixir of immortality. The search was fruitless and he reportedly ended up dying prematurely from mercury pills his alchemists made for him. Egyptian pharaohs devoted enormous state resources to defeat death in the next world. European monarchs funded alchemists chasing the ‘philosopher’s stone’, which was tied as much to the elixir of life as it was to turning lead into gold.

With the latest advances in science and the prospect of using AI, it’s no surprise that America’s billionaires are attempting their own ‘elixir of life’ search. Altos Labs launched with $3 billion in 2022, with Bezos reportedly among its backers. Sam Altman backs Retro Biosciences. Coinbase’s Brian Armstrong co-founded NewLimit. Peter Thiel is a major backer of SENS Research Foundation, whose original science officer at one point claimed that the first person to live to 1,000 was probably already alive.

About $3.74 bn was raised in biotech longevity research across 49 deals in Q1. Projections are for $8 to $9 bn in funding for 2026. Eli Lilly is starting to write checks alongside the venture capital firms.

Life Biosciences, co-founded by David Sinclair, got FDA Investigational New Drug (IND) clearance and dosed the first human patient with partial epigenetic reprogramming therapy in June 2026. They are targeting glaucoma, but the goal is to reverse the ageing in eyes so people can regain their sight. They are in a Phase 1 safety trial because epigenetic reprogramming carries cancer risks. But their research in mice showed restored visual function because of the treatment, and on a recent podcast Sinclair became less hedged than earlier in the year on whether sight can be restored to humans like what happened in mice. It’s easy to test, a blind man can see.

The root of the longevity movement stems from a series of tool breakthroughs and a couple of high-profile results. Shinya Yamanaka discovered four transcription factors in 2006 that could reset a mature cell back to a stem-like cell state. He won the Nobel Prize for it in 2012. This is the mechanism behind Life Biosciences glaucoma treatment. For years, this was thought of as a stem-cell biology tool, but only recently has been attributable as an anti-ageing tool.

In 2012, a paper on CRISPR turned gene editing from a slow, expensive, specialist procedure into something any well-equipped lab could do cheaply and precisely. This accelerated the rate at which ageing mechanisms could be tested in animal models. You could tag a gene of interest in months rather than years. In 2013, the “Hallmarks of Ageing” paper organized ageing into a list of discrete, nameable mechanisms (genomic instability, telomere attrition, epigenetic alterations, senescence) that provided a roadmap of study for researchers involved in the field. There have been several other innovations including epigenetic clocks, the ‘young blood’ experiments, senolytic drug combinations, but the tools and papers made ageing measurable and manipulable for the first time.

Over the next year we should hear more from Life Biosciences. If the safety profile works, something researchers have worried about, that would support the modality and potentially more trials in other areas. If people start seeing, plausibly 2027 onward, then the whole longevity platform would re-rate. There would be significant valuation and fund-raising tailwinds. Big pharma would go into M&A action. There would be start-ups to de-age all our body parts. And there will be dozens of Qing Shi Huangs funding alchemists.

The Case for Tesla

It’s easy to make a bear case on Tesla’s self-driving cars. Elon Musk promised a million Cybercabs and self-driving cars by 2019-2020. He cut his Austin fleet from 500 cars to about 60. Waymo launched in 2018 and now has 4,000 vehicles, pulling ahead of Tesla. Others are attempting to enter with Uber as an aggregator. Tesla’s operating margin was only 1.4% and gross margins excluding regulatory credits are only 16.3%. Morgan Stanley’s free cash flow burn prediction is $14 billion for 2027. There’s still an open question whether relying on cameras is enough to avoid a serious accident at real scale.

Our internal analysis puts the probability that Tesla overcomes its manufacturing and safety challenges and ramps full self-driving by 2030 at 40%, with a wide margin of error.

However, Tesla is a dogged company. We think they get there eventually. By 2033, our probabilities go up significantly. And this is when the auto revolution happens.

Ride-hailing today costs about $2.00-$3.00 a mile, with most of the cost being the driver. Owning a new car costs about $0.70-$1.00 a mile all-in. If a driverless fleet can get below the cost of ownership, then many urban and suburban households stop needing a car, or at least a second car. Then behaviour shifts. People could buy cars and then rent them into Tesla’s fleet and earn money, dramatically reducing the cost of the car. Ron Baron, a Tesla bull, thought a $30,000 car could make $50,000-$60,000 of cash flow a year. This number probably does not take into account of many costs such as insurance and the fees to Tesla, but the direction of travel of the math is right. A car becomes an asset rather than an expense.

Because your car could drive around trying to collect revenues, there would be no need for parking. Commuting time becomes usable time. Vehicle interiors could become workspaces.

Traditional auto manufacturers would be in trouble. The model would change to cost per mile rather than one-time margins. Validated safety miles end up becoming a moat, and Waymo and Tesla are the only ones with a credible path to that data at scale. Also, when you integrate the system by owning the software, vehicle, fleet operations and data loop, it let’s you optimise the whole system. This lowers cost per mile. Building the software and fleet is capital intensive, requiring tens of billions.

Nvidia is attempting to build a horizontal stack signing up Hyundai, Nissan, BYD and Geely to its Drive Hyperion robotaxi platform. Wayve, Nuro, Mobileye and Momenta are attempting similar licensing models, gathering test-mile data. Uber has committed more than $10 bn to its autonomous vehicle strategy, partnering with over 30 players. China has its own champions and, if Western markets allow them in, they may export their software.

But given the needs of integration, fleet optimization and scale effects, alongside the high safety barrier to entry, Tesla and Waymo have huge advantages. Only Tesla is well positioned for a fleet optimization strategy because it’s a true car manufacturer as opposed to Alphabet’s Waymo. And it’s unlikely Chinese competitors will be given license to operate in many Western countries, particularly the U.S.

The total addressable market of new cars is $800 billion a year in the U.S. The used car market is roughly similar. About 290 million registered vehicles drive 3.3 trillion miles per year. At ownership costs of $0.70 to $1.00 a mile, that is a $2 to $3 trillion mobility market. Tesla’s fleet strategy crosses over all these markets.

The bears have some good arguments, but the bulls do too. The market is modelling the probability of success very high but given Musk’s track record of eventually getting it right at SpaceX and Tesla, the market may be right.

French Bonds Wish They Were Greek. Bring out the champagne. The OAT-bund (French – German yield) crossed 100 bps for the first time in 14 years. The immediate catalyst was the government’s Thursday admission that the 2026 deficit will be 5.4% of GDP. The entire narrative since 2025 was that France was at least directionally consolidating. The 2025 target was 5.4% deficit, and the actual deficit was 5.1%. The target for 2026 was 5.0% and the projections are for 5.4%. An entire year of fiscal consolidation produced nothing.

The last time the OAT-bund spread was here was during the eurozone debt crisis. Greece can fund 10-year debt 26 bps cheaper. France is back at square one with a 0.2% quarter on quarter contraction in Q1 and stagnation in Q2, higher borrowing costs, and a dysfunctional political system.

On September 30th, France will present the budget. Prime Minister Lecornu is targeting a 2027 deficit of 5%, which will include a €54 bn savings plan. There is a 45%-50% chance that we see tighter financial conditions with no systemic break. However, a chaotic reception would likely push spreads towards 120 bps. This creates a 35%-40% chance of broader contagion including French bank stress and the ECB forced to engage under political pressure. There is also a tail risk that we have a full-blown systemic crisis triggered by a government collapse and a Mélenchon second-round confirmation to become the base case.

Bessent is the House… of Cards. The BOJ delivered its widely expected rate hike on Friday but BOJ Governor Ueda’s guidance disappointed hawks. He could not set down a timetable for future hikes. And two board members dissented on the hike. The yen sold off 1.2% to 158 looking like it will re-test the 160-intervention level. The yen selling dissipated after a ‘rate check’, which is the BOJ signalling it may intervene. The yen weakness helped push the 10-year above 5% on concerns of more U.S. Treasury bond selling from Japan. France’s deficit debacle and a hawkish Fed didn’t help either.

Japan goes on holiday next week. I expect the Takaichi reflationists and Treasury to continue to fight it out, but don’t see an imminent crisis event. But France’s September 30th budget will impact the yen and Treasury market as well as the Western sovereign bond markets are all interconnected.

End Note

David Wallace’s Infinite Jest has numerous plots and sub-plots, but one overriding theme is a weapon where people watch a VCR tape and have a permanent smile on their face such that they go into a state of permanent pleasure and can’t function anymore. In the spirit of Wallace, fal.live was created, an infinite “livestream” based on the MiniMax H3 model. It suffers from all the pitfalls of contemporary AI videos, like a lack of coherence over long time horizons, inability to create satisfying narratives, and so on. But it certainly gives a taste of the strange ‘infinite jest’ future we may be entering.

I bought my first car in 2006 in Tokyo, a time when engineers made cars and not CFOs. It’s so well made, I still drive it. My kids and wife tease me endlessly about changing to a newer car, but I like my car. So, I was quite edified to see a Bloomberg article saying that people are paying $30,000 for cars with 90,000 miles or more. Now my kids think I’m a genius for keeping the car so long. We’ll keep the lessons on inflation adjustment around for another time.

Omar Sayed