Study a Lot, Tread Lightly
Build conviction before committing capital to an emerging technology wave
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 97%
A major technology shift can be real while most early companies built around it still fail. The method separates conviction in the underlying wave from conviction in any individual asset. Begin by studying the technology, historical analogues, enabling infrastructure, competitive dynamics, and adoption evidence. Then evaluate whether a candidate possesses durable value despite falling prices, rapid imitation, and uncertain market structure. Compare speculative private companies with public businesses already producing measurable revenue, productivity, or margin gains. Passing on an investment is an active decision, not a failure to participate. The output is a concentrated set of evidence-backed opportunities—or no investment at all—while the investor continues learning and waits for clearer winners.
Origin
Brad Gerstner described this approach on Masters in Business while explaining why Altimeter passed on more than fifty AI companies despite believing AI could become bigger than the Internet.
Core principles
- 01A transformative technology can still produce many failed investments.
- 02Research creates readiness, but readiness does not require immediate action.
- 03Durable value matters more than exposure to a fashionable category.
- 04Public incumbents may offer clearer evidence than speculative startups.
- 05Patience preserves capital until winners become distinguishable.
How to run it
- 1
Separate the wave from its participants
Assess whether the underlying technology is structurally important without assuming that today's companies will capture its value. Hold both the long-term opportunity and the short-term speculation in view.
Pro tip Use prior platform shifts to distinguish category success from company success.
Watch out A correct macro thesis can still produce poor company-level investments.
- 2
Study the emerging stack
Research the technology's primitives, economics, adoption patterns, competitors, and likely sources of durable advantage. Keep building knowledge before forcing a decision.
Pro tip Track changes in pricing and capabilities because both can invalidate a business model quickly.
Watch out Do not mistake a polished product demonstration for a defensible company.
- 3
Look for visible fruits
Identify businesses already receiving measurable revenue, productivity, margin, or strategic benefits from the shift. Compare those proven beneficiaries with private companies whose value remains hypothetical.
Pro tip Include incumbents and infrastructure providers rather than limiting the search to startups.
Watch out Category labels such as “AI company” are weaker evidence than operating results.
- 4
Test durability and valuation
Ask whether the asset can retain value as underlying technology becomes cheaper, competitors proliferate, and incumbents respond. Reject opportunities whose price assumes an unproven winner.
Pro tip Model adverse scenarios such as a major supplier cutting prices by ninety percent.
Watch out Scarcity premiums can disappear faster than private valuations adjust.
- 5
Act selectively or wait
Invest only where research produces deep conviction and an attractive risk-reward relationship. Otherwise continue studying without committing capital merely to avoid missing out.
Pro tip Treat a well-supported pass as a successful decision.
Watch out Doing extensive work can create psychological pressure to invest even when the evidence says not to.
In the wild
In 1998, venture investors knew Internet search would become enormous and funded AltaVista, Infoseek, Lycos, Dogpile, and Ask Jeeves. Those early category bets largely lost their value. An investor could instead have continued studying and bought Google at its 2004 IPO, capturing most of the industry's subsequent profit creation without selecting an early logo.
→ Patience preserved capital and still allowed participation after a durable winner emerged.
Altimeter passed on more than fifty AI companies while making selected private investments. Gerstner said the firm found public markets easier because Microsoft, Snowflake, and Nvidia were already showing tangible benefits from AI demand and adoption.
→ The firm gained AI exposure through businesses with observable operating evidence.
Common mistakes
Buying a category label
Believing AI will transform the economy does not establish that a particular AI startup has durable value.
Letting research force a trade
Investors may feel obligated to act after doing substantial work, even when waiting remains the best decision.
Ignoring incumbent advantages
Restricting the search to startups can overlook scaled companies already monetizing the transition.
Is it for you?
Best for
It is best for investors evaluating crowded, fast-changing categories where enthusiasm exceeds reliable company-level evidence.
Not ideal for
It is not ideal for traders whose strategy depends on short-term momentum rather than fundamental durability.
From the transcript
“But as Warren Buffett has said, the hardest thing investing is to do all the work and then do nothing right right”
“The private markets, it's hard to know which assets are going to have durable value.”
“So I think you know, study a lot, but tread lightly.”
From the episode
Brad Gerstner on Tech Investing
Brad Gerstner