TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
$ANTHROPIC·$MA····$INTC····$BLUE-ORIGIN·$SPCX····$CRWV····$CRM····$MSFT····$NVDA····$ORCL····$CURSOR·$AAPL····$OPENAI·$AMZN····$UBER····$GOOGL····$META····$TSLA····$DATABRICKS·$PERPLEXITY·$LYFT····$NBIS····$TSM····$LITE····$ANDURIL·
←

mala gaonkar

T2 · manager / operator

Founded SurgoCap in 2021 with $1.8B, now ~$6B AUM. Spent 23 years as founding partner at Lone Pine Capital, one of the most successful hedge funds. Focuses on concentrated, long-duration moats across enterprise data/tech, financial services, healthcare services, and industrial technologies. Uses data-science debiasing and private-market access.

6 calls·6 names·67% bull·last heard 8 months ago·In Good Company with Nicolai Tangen
track recordleaderboard →
hit rate
50%
avg alpha
-5.9pp
scored
4

top calls

best measured alpha vs SPY, then highest conviction · one per company
1st+20.2pp vs SPY
$TSMTaiwan Semiconductor Manufacturingposition

TSMC's process-engineering moat cited as systemic, multi-layered advantage

TSMC owns not just the leading-edge fab but the entire process-engineering ecosystem around chip manufacturing, creating a systemic moat that is hard to replicate.

In Good Company with Nicolai Tangen2026-01episode →
2nd+6.2pp vs SPY
$NVDANvidiaposition

Gaonkar details Nvidia inference moat and admits sunk-cost bias from 2015 exit

Inference is the annuity stream of AI; application-specific chips (ASICs/TPUs) built for inference workloads create very long-duration moats. Gaonkar missed the full move after selling in 2015 due to crypto/gaming headwinds and failed to revisit due to sunk-cost bias.

In Good Company with Nicolai Tangen2026-01episode →
3rdmedium conviction
$ISRGIntuitive Surgicalposition

Robotic surgery penetration of 300M annual procedures seen as massive runway

Only early innings of robotic surgery adoption across ~300M global surgeries; combination of haptic feedback, mapping software, and surgical workflow integration drives lower error rates and easier training.

In Good Company with Nicolai Tangen2026-01episode →

most discussed · click a bar to filter

  • $ISRG
  • $OPENAI
  • $GOOGL
  • $NVDA
  • $TSM

recurring themes

  • AI Hardware & Chip Architecture1
  • AI in Healthcare1
  • Robotics & Physical AI1
  • AI Economics & Business Models1
  • Venture Capital1
6 total
$ISRG
···
Intuitive Surgical
MEDmala gaonkar·In Good Company with Nicolai Tangen·8 months ago·Mala Gaonkar - Founder of SurgoCap Partners | Podcast | In Good Company· position
Robotic surgery penetration of 300M annual procedures seen as massive runway
Only early innings of robotic surgery adoption across ~300M global surgeries; combination of haptic feedback, mapping software, and surgical workflow integration drives lower error rates and easier training.
"there about 300 million surgeries conducted globally. Uh and they're just beginning to be penetrated uh by you know what's happening with robotic surgery specifically you know com…"
7:33
$OPENAI
OpenAI
LOWmala gaonkar·In Good Company with Nicolai Tangen·8 months ago·Mala Gaonkar - Founder of SurgoCap Partners | Podcast | In Good Company
OpenAI building proprietary inference stack alongside Google and Anthropic
Major frontier model providers are each developing their own inference infrastructure, reinforcing the view that inference is a distinct, defensible layer.
"Google will have its own inference stack um open AAI uh anthropic and others"
22:26
$GOOGL
···
Alphabet
MEDmala gaonkar·In Good Company with Nicolai Tangen·8 months ago·Mala Gaonkar - Founder of SurgoCap Partners | Podcast | In Good Company· position
Google's custom TPU inference stack highlighted as key AI infrastructure play
Google's vertically integrated TPU stack for inference workloads reduces cost and creates a durable moat in the AI inference layer.
"Google will have its own inference stack um open AAI uh anthropic and others and I think what is happening around Google's TPU is really interesting in terms of reducing that"
22:26
$NVDA
···
Nvidia
HIGHmala gaonkar·In Good Company with Nicolai Tangen·8 months ago·Mala Gaonkar - Founder of SurgoCap Partners | Podcast | In Good Company· position
Gaonkar details Nvidia inference moat and admits sunk-cost bias from 2015 exit
Inference is the annuity stream of AI; application-specific chips (ASICs/TPUs) built for inference workloads create very long-duration moats. Gaonkar missed the full move after selling in 2015 due to crypto/gaming headwinds and failed to revisit due to sunk-cost bias.
"I think what is happening around Google's TPU is really interesting in terms of reducing that so I I think um AS6 or ASIC chips application specific chips that are designed for sp…"
22:09
$TSM
···
Taiwan Semiconductor Manufacturing
MEDmala gaonkar·In Good Company with Nicolai Tangen·8 months ago·Mala Gaonkar - Founder of SurgoCap Partners | Podcast | In Good Company· position
TSMC's process-engineering moat cited as systemic, multi-layered advantage
TSMC owns not just the leading-edge fab but the entire process-engineering ecosystem around chip manufacturing, creating a systemic moat that is hard to replicate.
"businesses like TSM that are well known that have really kind of corner in Taiwan kind of the process engineering aspects of that. So it's not really just about one thing it's all…"
21:37
$ANTHROPIC
Anthropic
LOWmala gaonkar·In Good Company with Nicolai Tangen·8 months ago·Mala Gaonkar - Founder of SurgoCap Partners | Podcast | In Good Company
Anthropic building proprietary inference stack alongside Google and OpenAI
Major frontier model providers are each developing their own inference infrastructure, reinforcing the view that inference is a distinct, defensible layer.
"Google will have its own inference stack um open AAI uh anthropic and others"
22:26
9
AI Hardware & Chip Architecturetailwind
Inference ASICs/TPUs create long-duration moats as training demand proves spiky
Training demand is volatile but inference is a recurring annuity stream; custom silicon (Google TPU, merchant ASICs) purpose-built for inference workloads will compound advantages over years.
8
AI in Healthcaretailwind
Medical imaging accuracy and speed improving ~70% via AI, transforming preventive care
AI-driven imaging (MRI/CT) advances are accelerating early detection of chronic diseases in aging populations, making care more cost-effective and shifting focus from therapeutic to preventive.
8
Robotics & Physical AItailwind
Robotic surgery at inflection: 300M annual procedures barely penetrated
Da Vinci-style systems combined with haptic feedback, mapping software, and workflow integration are reducing errors and training time, creating a multi-decade adoption curve across global surgical volumes.
8
AI Economics & Business Modelstailwind
Inference is the annuity stream; training is spiky and capital-intensive
The recurring revenue of model inference calls creates predictable, high-margin economics versus the lumpy, capex-heavy training phase, favoring companies that own the inference stack.
7
Venture Capitaltailwind
Concentrated four-vertical portfolio with thematic hedges enables offense during factor rotations
Focusing on enterprise data/tech, financial services, healthcare services, and industrial technologies provides diverse long drivers; thematic overlays allow adding risk when passive flows create dislocations.
7
AI Safety & Alignmenttailwind
System-2 thinking and data-science debiasing are critical to avoid sunk-cost and confirmation traps
Explicit checklists, automated alternative-data tracking, and 'thoughtfully missing out' (TOMO) on hype cycles (crypto, quantum) protect against behavioral biases that caused past errors (Nokia short, Altice leverage, Nvidia re-entry failure).
7
Private Marketstailwind
Change happens at the edges: private markets offer earliest exposure to disruption
Most disruptive companies start small and private; maintaining global networks in focus verticals (enterprise data, fintech, healthcare services, industrial tech) captures compounding optionality before public markets.
7
Data Center Infrastructuretailwind
Real-time data providers with system-of-record + workflow moats resist LLM disruption
Proprietary real-time data (FX pricing, fixed income, equities) embedded in trading/compliance workflows is structurally hard for LLMs to displace due to latency, accuracy, and regulatory requirements.