When VC consensus forms around a playbook, its returns begin to fade
As VCs crowd into passive shareholding in consensus picks like Databricks and Anthropic, Naimi warns: history shows almost no strategy that everyone finds simple sustains long-term outperformance.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Second-tier allocations in the same round now cost 3 to 5 times as much
VCs that miss the main allocation get a second tier in the same round, priced 3–5x the main. Silicon Valley calls this the "cuck tranche." Naimi reframes it not as market failure but as these companies being capital-intensive enough that founders won't close in one draw. Most VC shops have been frozen out of one or two allocation tiers at least once.
— Ramtin NaimiMega-funds deploy 2 to 3 times faster than their budgets assume
VCs size funds on 18–24 months of planned deployment, but mega-platform funds have actually deployed in 12–18 months—two to three times faster than projected. This forces a binary choice: maintain position size but write far fewer checks, or write the same number of checks and accept dilution. The first tests stock-picking skill; the second bets that diversification hedges not being a top-tier picker.
— Ramtin NaimiLiquidity for most LPs depends on success in just four companies
Some LPs put 35–50% of assets into private equity. Drill down and 30–50% of that sits in just four names: SpaceX, Anthropic, OpenAI, and Databricks. Real LP cash doesn't flow from sector health—it flows from exposure to these four. LPs without these bets are drawing almost no liquidity this cycle.
— Ramtin NaimiBurning billions to reach 85% margin demands new finance tools
Anthropic's S-1 revealed $140 billion cumulative fundraising, $10 billion burned, $132 billion on the balance sheet, and 85% gross margin. Extreme burn married to high margin forces innovation: General Catalyst's customer-value fund, lending on "every $10M deployed unlocks $30M revenue," has proven viable. But the industry still needs purpose-built infrastructure finance instead of pure equity.
— Ramtin NaimiNearly 100 model labs burn annual compute budgets in single training runs
The NeoLab landscape exploded from two or three players to 80–90 companies at $1–5 billion valuations. Naimi's example: one firm is raising a second $1 billion for a single training run next year, then expects another $5 billion the year after. Real constraint isn't equity—it's securing 5–10x leverage as non-dilutive debt at rational terms. Recent stress in public infrastructure debt has spiked that cost.
— Ramtin NaimiEndpoint assumptions jumped from $100 billion to $1 trillion; seed valuations followed
In 2016, best-case endpoint was $10 billion. By 2021, Stripe and Databricks pushed it to $100 billion. Now SpaceX, Anthropic, and OpenAI have proven $1 trillion is real, so every early company gets priced as if it can reach $1 trillion—even entering at $100 million.
— Ramtin NaimiBacking unfashionable bets early separates top 5 percent from top 25 percent
Passive plays on hot sectors top out around 25th percentile. Top 5% requires early conviction in neglected bets. Example: Conduit, a brain-computer interface company. Nine months ago, BCI was ignored. The team built a headset reading neural activity at 80% accuracy. Eight to ten weeks later, BCI went hot—Conduit's round drew 13 term sheets and massively oversubscribed. Smarter investors see the wave before the crowd.
— Ramtin NaimiCrowding into "safe" mega-rounds is itself the largest risk signal
When early markets are chaotic, one hedge is mega-sized positions in proven winners. Stripe, valued at $100 billion in 2021, is now worth $179 billion, yielding 11–12% annualized over five years—still placing in the top 25% for VC returns. But over the past 18 months, VCs have embraced this same playbook: huge checks into a handful of known bets, stay passive. History shows almost no strategy that everyone finds simple sustains long-term market outperformance.
— Ramtin NaimiWhat looked reckless ($100M in 2018) now looks prudent ($800M in 2026)
Abstract's Fund I, $100 million in 2018, drew LP complaints for being oversized. The latest, $800 million, is praised as disciplined. In retrospect, Naimi's regret: undersizing each fund. The same returns could have been generated with twice the capital, and he'd have had more dry powder for leads instead of being pushed into follow-on checks.
— Ramtin NaimiIn their own words · checked verbatim
I think anybody who tells you they've never gotten cucked is completely full of shit. Everyone's gotten cucked once or twice at least.
Ramtin Naimi3:10
Every single one of the mega firms has dramatically underestimated how much capital they need to actually get that level of ownership in those companies over that window of time
Ramtin Naimi11:32
It doesn't actually have to do with all of these companies you're hearing about. There's just a few companies that really moved the needle for all of these LPs.
Ramtin Naimi12:35
But $10 billion of cash to get to the revenue numbers that they're at with 85% gross margins is unbelievable.
Ramtin Naimi15:43
there's probably 80 or 90 NeoLabs that have raised somewhere between one to $5 billion.
Ramtin Naimi21:53
I think reactive, if you pick correctly, allows you to be a top quartile manager. And I think being proactive allows you to be a top 5% manager.
Ramtin Naimi26:05
I can't really think of many times in history where like an easy strategy turned out to also be a good strategy long term.
Ramtin Naimi34:23
100 million was seen as a large fund in 2018 and 800 million is seen as a small fund in 2026.
Ramtin Naimi39:32
Figures
| Anthropic cumulative fundraising | $140 billion | 15:43 |
| Anthropic cash burned | $10 billion | 15:43 |
| Anthropic cash on hand | $132 billion | 15:43 |
| Anthropic gross margin | 85% | 6:20 |
| Abstract Fund I (2018) | $100 million | 39:32 |
| Abstract latest fund | $800 million | 39:32 |
| Naimi's current opportunity/bubble assessment | 10/10 opportunity; 6.5–7/10 bubble risk (vs. 10/10 in 2021) | 42:44 |
Glossary
- cuck tranche
- A higher-priced allocation tier in the same round, typically 3–5x the lead price, for investors who missed the main allocation.
- NeoLab
- An early-stage AI model company funded at $1–5 billion valuation.
- BCI
- Technology that reads brain activity to infer user intent.
- benchmark maxing
- Optimizing a model specifically for test performance rather than general-purpose capability.
How to listen
VCs and LPs working in primary markets, and founders wanting to understand the real mechanics of infrastructure finance and fund-raising cycles.
Around 44 minutes a transcription error loops—a segment on prediction markets; skip it.