A hedge fund manager, Dinakar Singh, spent $150 million of his own money to develop three drugs for his daughter's rare disease (spinal muscular atrophy) after his diagnosis. He used a 'parallel processing' approach (trying all paths at once) instead of the usual slow, step-by-step method. He regrets not capping prices—now some drugs cost $730,000 a year. Key players: Biogen (its SMA drug makes $4-5 billion yearly but is criticized for high cost), Novartis (bought a gene therapy for $8.7 billion, but only works for very young kids), and Roche (his daughter uses their daily injectable pill).
This episode of Invest Like the Best features a conversation with Dinakar Singh, founder and CEO of Axon (formerly TPG-Axon Global Long/Short Hedge Fund). After his son was diagnosed with spinal muscular atrophy (SMA), Singh combined his financial background with pharmaceutical R&D, using a "virtual
Dinakar Singh, former founder and CEO of TPG-Axon hedge fund, combined his financial background with drug development after his daughter was diagnosed with spinal muscular atrophy (SMA). He drove the approval of three SMA treatments through a "virtual company" model. Singh argues that in biotech investing, time is the scarcest capital, and "parallel processing" rather than "sequential advancement" is key to breaking the deadlock of rare disease drug development — he self-funded approximately $150 million, compressing a development cycle that might have taken decades into the fastest approval timeline in FDA history.
Singh believes that for parents, "seeing their child suffer while realizing they could have done something but didn't act in time" is more terrifying than losing the child—this fear drives him to act "at maximum speed and with no effort spared."
Singh explains that traditional drug development is "serial"—do one study, then decide the next step, taking over a decade. He advocates for "parallel" processing: advancing five paths simultaneously, exposing failures quickly and accelerating successes.
Singh attributes his success to "luck," but a more accurate description is that by systematically covering all possible mechanisms, he increased the probability of "winning the lottery."
| Drug Type | Main Driver | Key Milestone | Current Status |
|---|---|---|---|
| Antisense Oligonucleotide (ASO) | Ionis → Biogen | First to be approved, fastest FDA review | Annual fee ~$730,000; now one of Biogen's three top-selling drugs, annual revenue ~$4-5 billion |
| Small Molecule | PTC Therapeutics → Roche | Drug used by his daughter | Daily injection, effective and safe |
| Gene Therapy | Avexis → Novartis (acquired for $8.7 billion) | First approved gene therapy drug | Only suitable for very young children (viral vector toxicity is too high for older children) |
Singh points out that the success of the SMA drug market has changed the narrative of rare disease R&D—but extreme pricing and international regulatory fragmentation are creating new tragedies.
| Position | Guest Attitude | Key Data |
|---|---|---|
| Biogen | Neutral (mentions collaboration) | Its SMA drug costs approximately $730,000 per year, with annual revenue of about $4–5 billion |
| Novartis | Neutral (mentions collaboration and acquisition) | Acquired Avexis gene therapy for $8.7 billion; early on, it offered free testing of compound libraries |
| Roche | Neutral (mentions collaboration and that his daughter is using its drug) | The drug is a daily injectable small molecule; Singh's father previously worked at Roche |
| Ionis Pharmaceuticals | Neutral (partner) | Developed the first approved ASO drug |
| Avexis | Neutral (acquired entity) | Developed the first approved gene therapy drug |
| PTC Therapeutics | Neutral (R&D partner) | Identified an effective small molecule drug through extensive screening |
| Jackson Laboratory | Positive (Singh currently serves on the board of directors) | Core supplier of mouse models; Singh plans to establish a "Rare Disease Think Tank" there |
1. “The worst thing is not that the child suffers, but that you realize you could have done something but didn’t have time.” — Singh used this “personal risk formula” to define action priorities: time is scarcer than money, and parallel execution is safer than serial.
2. “Neuroscience is a graveyard for drug development—but SMA has a ‘backup gene’ and an extremely wide therapeutic window, which makes it more ‘solvable’ than most neurological diseases.” — This shows that sound investment judgment requires identifying problems that are “structurally easier to solve,” rather than focusing only on scale.
3. “We open up all testing resources, and the company retains the IP—we only ask that someone discovers something and tells us.” — This is a “platform” approach to R&D: giving up ownership in exchange for information flow, reducing the cost of trial and error for all participants.
4. “If you had told me 20 years ago that SMA drugs would sell for a million dollars a year, I would have thought you were crazy—now it’s reality, and I regret not adding a price cap when I invested in the beginning.” — Singh candidly reveals the “narrative reversal” of rare disease drug pricing: from no one caring to no one being able to afford it, the system needs new constraints.
5. “SMA is becoming the ‘standard model’ for muscle research—because the hole in the boat is patched, and now we can really drain the water.” — This means that a breakthrough in the “infrastructure” of a field (such as an effective base treatment) can unleash downstream innovation potential, and investors should focus on the second wave of opportunities after “patching the hole.”
6. “Gene therapy can only save very young children—for older people, the viral vector itself will kill you.” — This is a critical technical time window constraint: not all therapies are applicable to all patients, and investing requires a precise understanding of the technology’s applicable boundaries.
7. “My ‘Ghostbusters’ plan: set up a team at Jackson Lab to help rare disease families create a ‘game plan’—no fees, no IP ownership, just be the ‘Switzerland.’” — Singh attempts to replicate the “model” from SMA’s success (parallel execution, neutral coordination, open resources) to other rare disease areas, with the goal of lowering the entry barrier for the “first attempt.”