Summary

Abstract’s founding thesis: strip out Uber and Roblox (both seeded by First Round) and it was “close to impossible” to find a $5B+ power-law company whose seed round was led by a seed fund — Sequoia seeded Stripe, Airbnb, Dropbox and Nubank; Andreessen led first institutional rounds for Okta, Databricks and Slack; Khosla seeded Instacart and DoorDash; Index seeded Robinhood and Figma; Lightspeed seeded Snap and (he thinks) AppDynamics. Seed funds claiming proprietary deal flow were “mostly kidding themselves. They’re a little delusional” — so Naimi built a seed firm aligned with the multi-stage platforms rather than against them.
The repricing math behind everything: roughly 1,000 seed deals got done 2008–2011, and an equal check into every one would have returned 3,000x on Uber alone — a 3x net fund for blanketing the entire market. “There is no world in which you could blanket an asset class and generate a 3, 5, 7x” — proof seed was too cheap, and why entry floors should be 3–5x higher. At today’s ~$25M average entries, blanketing merely breaks even: “now you actually need to be better at picking than you were historically.”
Ownership as a relative, not absolute, metric: his 5% from a $100M fund versus a co-lead’s 15% from $1.5B is 5x the look-through exposure — “people ask me if I had an index approach. I’m like, it’s the opposite of that.” In 94% of fund-one companies, no venture fund anywhere offered more look-through ownership. Abstract now leads 80–90% of its deals and claims the highest seed-to-Series-A graduation rate to a tier-1 firm “by a pretty wide margin.”
The product is lowering founders’ future cost of capital: Naimi runs the entire Series A raise as the sole node — deck, data room, mock pitches, direct GP intros, all meetings in a 3-day window, nightly feedback. Result per his data: highest-decile A valuations, lowest average dilution, and at a $2B exit an extra ~$200M in the founder’s pocket — “I can’t personally think of a single value-add that a venture capitalist brings to the table that translates to more.”
The 2021 warning: in the last six months he’s funded seven companies that raised Series As within weeks at 4–5x his entry price with “not a whole lot of fundamental business progress,” and his latest funds are printing 30–40% IRR within months of deployment when they should be in the J-curve. “The industry is getting a little drunk on IRR.” On AI valuations themselves he refuses a verdict for 3–4 years: 20x growth decaying to 10x has no established fair multiple, and application-layer companies can become obsolete overnight when they become a feature of a foundation-model company.
Staying private is a deliberate productivity throttle: “early liquidity was the bug and not the feature of crypto — people got rich too quickly and stopped building things.” When Google went public, over 100 people made millions and “92 of them were never heard of again.” Today’s 1–5% tender offers buy a down payment, not generational wealth — great for company building, “problematic from a venture returns perspective,” with continuation vehicles as the emerging fix.
The art world runs on venture mechanics: the big four galleries (Hauser, Gagosian, Zwirner, Pace) are the platform funds, smaller galleries are the seed spotters, and graduation to a blue-chip gallery quadruples prices — “nicer to pay 25% of the cost.” Masterpiece is a power law inside a single artist’s output: one painting in a 12-work show is the one everyone fights over, which is why one Picasso sells for $5M and another for $100M.
The biography is the pitch: bankrupt at 24 (Chapter 13) after self-funding a marketplace-lending exchange straight into the sector’s collapse, then 47 deals in 10 months via AngelList — first syndicate (Ripple) filled $470k in four hours; SPVs have returned ~$100M — before selling a stake in his management company to a consortium likely including Andreessen, Ovitz, Ackman and Sacks with a seven-year sunset. Abstract is now just shy of $2B AUM. “I don’t think there is any excuse saying that you can’t start a venture capital firm with no money.”

Deep dive

1. The art market is venture with paint: same structure, same graduation trade

Naimi’s route in was social, not academic: mentors Michael Ovitz and Stuart Peterson would spend “two, three hours walking you around their home” narrating their collections, and he wanted a hobby he and his wife could share — “art collecting seemed like it could be golf for both me and my wife.” He collects living contemporary artists in three generational buckets (mid-60s/70s, late-40s/50s, and his own 28–38 cohort), used an adviser only briefly, and concluded “you’re better off doing the majority of the work yourself.”
His structural read: at any moment ~50 important practicing artists are represented by the big four — Hauser, Gagosian, Zwirner, Pace — and once signed they’re blue chip, with prices that sustain because those galleries have the clients and institutions to “control the markets for those artists.” Beneath them sit the same recurring smaller galleries who discovered the talent first — exactly like the seed firms the big platforms cultivate for early deal flow.
The trade is the graduation: find galleries with a high “graduation rate” to the mega-galleries, buy before the jump, and “as soon as that happened, their prices basically quadruple … it’s nicer to pay 25% of the cost of something versus full value.” But not every artist graduates — “that’s when you have to start applying some of your own judgment and hope you have the right taste.”
The market has repriced like venture did: 20–25 years ago entry-level art was $10,000 and you could take a painting home for two weeks and bid 20–25% under ask; now big galleries carry 75–80 artists (versus 10–15 historically), he gets ~30 exhibition previews a day, and one mentor’s rule stings — he’s “never bought a painting for less than $100,000 that he’s made any money on.”

2. Masterpiece is a power law inside one artist’s output

The highest-scoring collection on the list that ranks collections by the means at which they were built wasn’t Cohen’s or Griffin’s — Patrick’s quip, which Naimi endorsed: “it’s like a low IRR collection” when you can buy anything. It was a postman and a librarian who spent 1960s weekends buying Pollocks and Rothkos on pure eye, before any value was assigned, and gave it all to museums.
Within any new show of 12 paintings, “there’s probably one that’s incredible that everyone’s going to be fighting over, two that are really good, and the rest is kind of just… the gallery was like, we could sell 12 of them.” The dramatic appreciation accrues to the best examples — why one Picasso clears $5M and another $100M. Ovitz’s edge was discipline: maybe only three Picassos, “but the three best Picassos ever,” rather than accepting consolation prizes.
His value screen: art cycles predictably — “things that were iconic 20 years ago tend to be undervalued 20 years later, then get very highly valued 30 to 40 years later.” The undisputed iconic bodies of the ’90s and 2000s (Richard Prince’s Cowboys and nurses, Tracey Emin) skew white-male and out of fashion — “not a lot of demand for their work, but a really good buying opportunity.” Counterweight: check collector-base depth — if everyone collecting an artist is 30–40 years older, “am I going to be the only person collecting this artist in 20 to 30 years?”
The unifying skill is Ovitz’s phrase, frame of reference: “the more companies you meet, the easier it gets to discern what the better companies are. The more art you look at by a specific artist, the easier it becomes to identify what a great example is.” The tell of a masterpiece: you know who painted it “within the first second of looking at it.”

3. Galleries make the most money; status and museums are the access layer

Who wins economically? “The galleries.” They take 50% on the primary sale, then steer resales back to themselves for another 15–25% — “on multiple round trips of a painting, they probably made more money on the painting than the initial cost of the painting.” Cost of goods “basically zero,” very large tickets, 50% rev share. Patrick’s verdict: “Good business.”
Auction houses look better than they are: 15–26% headline commissions, but estates (which must sell fast for taxes) trigger guarantee wars — Paul Allen’s collection cleared well over $1B at auction — and competing guarantees can compress net margins to 2–3% or losses. Sotheby’s at a ~$3B market cap after 100+ years: “good businesses, not phenomenal businesses.”
Status isn’t vanity, it’s plumbing: “artists actually don’t want their art in people’s homes. They want it hung in the MoMA where millions will see it.” Galleries favor museum-board members because those collections eventually land in institutions — “the boards of these museums don’t look like boards of any other industry in the world because the boards unlock access.” Naimi’s own reputational lesson from Ovitz: refuse the consolation prize, or “you will always be the guy they give second and third tier work to.” And the best entry point is between exhibitions, when an artist quietly gives the gallery one painting shared with “you and two or three other people” — not the fair previews sent to 5,000 collectors.

4. The founding thesis: multi-stage firms were the better seed investors

Naimi defined power law as a private value or exit north of $5 billion, then found that “if you eliminate Uber and Roblox, whose seed rounds were led by First Round Capital, it’s close to impossible to identify power law companies in which the seed round was led by a seed-stage venture capital firm.” The receipts: Sequoia led seeds for Stripe, Airbnb, Dropbox and Nubank; Andreessen the first institutional rounds of Okta, Databricks and Slack; Khosla seeded Instacart and DoorDash; Index seeded Robinhood and Figma; Lightspeed seeded Snap and (he thinks) AppDynamics.
The uncomfortable corollary: “seed funds that claim they had proprietary deal flow were mostly kidding themselves. They’re a little delusional. It’s hard to believe a seed firm with two or three people has more coverage at early stage than a multi-stage fund with 30 or 40.” Head-to-head, multi-stages won on brand and on terms seed funds couldn’t match — so he built Abstract “aligning my interest with multi-stage funds as opposed to aligning my interest with seed funds.”
The pricing proof that the multi-stages, not the seed purists, “had a bit more of the right idea”: between roughly 2008 and 2011 about 1,000 seed deals were announced. An equal check into every one captures Uber’s 3,000x — a 3x net portfolio for blanketing the entire market, before adding Airbnb, Dropbox and Instagram. “There is no world in which you could blanket an asset class and generate a 3, 5, 7x.” Conclusion: seed deals needed to be 3–5x more expensive as a floor — and even at a $25M average entry, blanketing should only break even. “Now we’re in a market where you actually need to be better at picking than you were historically.”

5. The tracker: 6–7,000 LinkedIn profiles and 47 checks in ten months

In 2016 he reverse-engineered the last few hundred tier-1-backed founders — “venture is a pattern matching business, for better or for worse” — into a profile (these schools, degrees, companies, roles, at these moments in the company’s inflection), tracked roughly 6–7,000 matching people on LinkedIn, and got a push notification whenever one changed their title to founder. His discovery: “an unfunded seed-stage founder might be like the easiest person in the world to get a meeting with” — useful, since “I was a nobody at this point in Silicon Valley.”
The output, August 2016 to June 2017: 47 deals — an angel check into Rippling, “a first dollar check into Solana at four cents a token,” seeds in Clay, Cherry, Newfront, the management-company round of Polychain (briefly the world’s largest crypto hedge fund), then Avalanche and dYdX. Scoreboard: two positions with $100B+ coin market caps (Ripple and Solana), Rippling approaching $20B, and eight or nine unicorns — with AngelList SPVs that “have returned close to $100 million.”
AngelList was the entire cold-start solution: he sourced the deal, wrote the memo, and the platform sourced every dollar. His first syndicate — Ripple — had “$470,000 subscribed” when he refreshed four hours after posting: “Holy shit… I can’t believe this actually works.” Economics: 0-and-20 to LPs, AngelList takes 5 points of carry for sourcing capital, he keeps 15. For a six-month stretch he was one-third of all volume on AngelList.
How he got in: “Very aggressive… just being relentless,” but deliberately likable and infinitely flexible on size — “I don’t care if it’s 25,000 or 500,000… I made myself flexible enough that it became hard to say no to me. If you give founders hard constraints on allocation or ownership targeting, you make it very easy for them to say no.” Plus weekly catch-ups with every junior VC in the Valley: when multiple people named the same company in one week, he called it.

6. Flat broke at 24: the résumé before the firm

The pre-history is pure hustle: Iranian immigrant parents (“America revolves around money — I heard it a little too often as a kid”), $2,000 borrowed at 13 to trade — quickly derivatives, since $2,000 wasn’t enough for stocks. At West Elm at 16 he gamed the non-clawback incentive structure by upselling staging (“have you ever seen a coffee table that’s higher than a sectional?”) and once single-handedly unloaded four UPS trucks of furniture, 7:30am to 7:30pm, when every other stock guy called in sick. His sweet-16 party business — he financed the venue and DJ, charged cover, hired “roided-out meatheads” for $50 a night — netted ~$2,000 a weekend, cash stuffed in a garbage bag when the pencil box overflowed.
Senior year of high school he made a few hundred thousand dollars trading out-of-the-money options on triple-levered bank ETFs through the financial crisis — “the good thing was it gave me a ton of confidence and I convinced myself I was a genius. The bad thing was it convinced me I was a genius.” He skipped college, got a Series 65, and launched a hedge fund in January 2009 with ~$3M from 45 small checks. Months down 37% and 51% followed; August 2011 — Greece, 13 consecutive down days, and he was short vol (likely; garbled in source) — was “one of the worst months of my life.” Everyone still made multiples, but he hated it: “the feedback loops in hedge funds are too tight — you can convince yourself you’re a good or bad investor too quickly.”
Then the wipeout: told by VCs (including GV’s David Crane, after a cringeworthy suit-and-tie meeting) to start a company — “terrible advice to give anybody” — he self-funded a secondary exchange for marketplace-lending loans in mid-2014. Fourteen months later, product ready, the sector had collapsed: Sequoia was marking Prosper toward zero, LendingClub was down ~85% from its IPO peak. “I had spent my entire net worth building a supplemental product to a collapsed industry” — ending in Chapter 13 bankruptcy at 24, cleared within two years.
The reframe he now carries: Arjan Schütte (likely; “Aron Chute” in audio) of Core Innovation Capital gave him a job when “no one else in the world would have” — his answer to the kindest-thing question. And Jerry Yang, as an LP, refused his embarrassment: “wear that as an entrepreneurial badge of honor… Venture capital is you put everything somebody else has into whatever you believe in. And if it doesn’t work out, you get a job at Apple or Google.”

7. Selling a stake to the consortium — with a sunset — and rewriting portfolio construction

Cyan Banister (likely), then at Founders Fund, noticed him and introduced Kevin Hartz — Xoom founder (sold to PayPal for over $1B), Eventbrite co-founder, $3M-post Airbnb investor, first dollar into Pinterest. Hartz’s line after their first meeting: “If you were this good at this, not knowing anybody, I wonder how much better you’ll get if you know all the right people.” Within roughly three weeks the chain ran Hartz → Chris Dixon and Keith Rabois → Marc Andreessen → Michael Ovitz → likely Bill Ackman, Kevin Warsh, and likely David Sacks.
That consortium bought an equity stake in his management company when he was 26, priced “just shy of 50” — but with a six-to-seven-year sunset: “everything I do in seven years, and one day they’re not entitled to economics.” Since venture’s excitement “starts to happen in years 9 and 10,” a seven-year timer “is nothing” — it lapsed a couple of years ago and “I’m proud to say I own 100% of my business again.”
Fund one: $100M closed end of 2018, ~$50M anchored by the consortium, the rest from Josh Kushner, Matt Cohler, Neil Mehta, Leif Abraham (likely), Dan Rose, Chase Coleman, plus operators like Jerry Yang and Okta’s Frederic Kerrest (likely) — and only four institutions (two college endowments, two fund-of-funds).
Those four grasped the heresy: institutions had been stuck on “VCs must own 15%” for 25 years while the funds grew 10–30x. Naimi’s reframe — ownership as a relative metric: his consistent 5% from a $100M fund versus a co-lead’s 15% from $1.5B is one-third the ownership from 1/15th the fund — “I actually had 5x the exposure… people ask me if I had an index approach. It’s the opposite of that. I’m the most concentrated exposure you can get.” By full deployment, in 94% of portfolio companies “there was not a venture fund anywhere in the world that would have gotten you more look-through ownership.”

8. From the 5% tax to leading: the multi-stage floor is 10%

The original playbook ran 20-for-20: introduce his best companies to the relevant tier-1 partner — “You take your 15, I’ll take my five” — every time getting his five, a tier-1 co-lead, and happy LPs. Then the irritation: “I felt like I was doing most of the leg work and getting this 5% tax.”
The fix was leading while breeding the multi-stage in as co-lead, and stress-testing “how far can I push a multi-stage firm down on ownership before they walk but are still happy”: the answer is 10% — “if we push them down to six or seven, it’s just not worth it. They’d rather wait for the Series A.” Validation came fast: of his first four led deals, two raised follow-ons from Benchmark, one from Sequoia, one from Andreessen — proof his leads weren’t adverse selection.
Today Abstract leads 80–90% of its portfolio, its led deals “have dramatically outperformed” its co-invests, and by his data it has “the highest likelihood of getting a follow-on Series A by a tier-1 VC firm… by a pretty wide margin” — at a pace of ~14 net new seed deals a year, “a little over one deal per month.”
The machine is built for speed and coverage: 3–5 meetings plus back-channels, compressing “what might take another firm two to three weeks into two to three days”; 18–30 pitches a week (“it’s not a flex to get a meeting with me — I pretty much don’t say no to meetings”); a daily 30-minute all-hands on every company met. His creed, via Doug Leone through Roelof Botha: “Dumbo ears — you have to hear everything or see everything.” And a direct pushback on the prior day’s guest: coverage may not matter at growth, but “at seed stage, coverage is absolutely key… most of these companies don’t even have websites yet.”

9. Picking: three salesmanships, dilution sensitivity, and the non-local pivot

“Investing in founders works a lot better for me than investing in markets.” The genetic makeup: commercial plus technical, where commercial means salesmanship in three verticals — fundraising; hiring (“hiring in early-stage startup companies is sales” — convincing a $400k Meta-caliber engineer that a $150k salary plus equity in “a concept of an idea” is rational); and selling V1s that are “half-ass broken glitchy things that no one would pay for.” Technical bar: an engineer other engineers would actually work for, with shipping velocity — “six weeks or six months?”
His favorite screen: “prove to me that you’re exceptional” — because of the old Quora question about founders over 35, answered by Reed Hoffman, Marc Benioff and Reed Hastings, all already successful when they started. “It’s hard for me to believe that the first impressive thing they’re ever going to do is this company they’re asking me to invest in.”
The counterintuitive preference: dilution-sensitive founders — not the “standard” seed seller of 20–25%, but the one who says “I need $3 million and I want to sell as little of my company as humanly possible.” Those founders hold the same bar on every point of equity, fire misses before the 12-month cliff, and build teams whose smallness shocks people relative to what they’ve achieved — “incredibly high-quality talent has a multiplier effect on company efficiency.”
On resilience: “the local pivot is what kills companies.” Poparazzi (seeded as TTYL, five or six pivots earlier) hit #1 on the App Store on launch day, took a Benchmark term sheet — then, six weeks in, no retention, and the founder returned the cash rather than grind out another consumer-social idea. Vapi (originally Superpowered) went zero to double-digit millions of ARR within 14 months of launch — four years and seven pivots after Naimi seeded it. Krea (likely; originally Genverse) is the same story.

10. Winning: run the whole fundraise as the sole node — that’s the product

The broken thing he chose to fix: the founder’s Google-doc fundraise, where nine investors intro 30 VCs — too many misaligned incentives, too many leaks (“VCs are just trying to gather information”). His alternative: “I will be the sole node. I’ll help with the deck, the data room, mock pitches; I’ll make introductions to the relevant top GP at every firm; we’ll line up all meetings over a 3-day window; at the end of every day I’ll get feedback on what’s resonating.” Choking off information forces VCs to “truly build conviction” — and they move faster.
Leverage is the deliverable: more term sheets, less dilution. His data: highest seed-to-A graduation rate, highest-decile Series A valuations, lowest average dilution. One founder’s reference call: the odds you own 10% more at exit are “10x higher with Abstract on your cap table” — 5% saved at the A, 3% at the B, 2% at the C, worth an extra $200M on a $2B exit. “I can’t personally think of a single value-add that translates to more than an extra $200 million in your pocket.” The ultimate validation: tier-1 co-leads now tell founders “we’re more than happy to let Ramtin run this process — they’ll do a better job than anybody.” Patrick’s tag: “you’re sort of the czar now for this process.”
Who wins the competitive A? “Sequoia, Benchmark and Andreessen… when one of those three extends an offer, they tend to win unless competing with one of the other two.” A LinkedIn stat he cites: the firms leading the highest number of financings in companies before they became unicorns since 2015 — Andreessen first, Sequoia second, Benchmark fifth, “shocking when you consider how much smaller their funds are.” The moat is board-member density: founders are told to optimize for “the best board director for the next 10 years,” and those three have the storied ones.
The missing piece at Abstract is brand — and the gap between personal and firm brand: as one VC told him, “I hear the name Ramtin multiple times a week; I hear the name Abstract once a month.” Very few firms have more than one to three partners who can independently source, compete and close — “which is why venture is hard to scale as an asset class,” and why LPs wish they could back just the four GPs, not sixteen. The criticism he accepts historically: “heat seekers, signal chasers… which wouldn’t have been inaccurate six or seven years ago — but over time, I think we became the signal. People tend to hold on to old narratives.”

11. The health check: AI multiples undetermined, seed-to-A drunk on IRR, private-market liquidity as throttle

On AI valuations he refuses the easy call: “I actually don’t know if it’s true [that they’re too high], and I don’t think anybody will know for another three or four years.” A company growing 20x, then 15x, then maybe 10x year-over-year has no established fair multiple — but there’s also “no shortage of application-layer companies that become obsolete overnight when it becomes a feature of a foundation model company.” Verdict: “undetermined” — and since venture is about capturing outliers, “you’ll do whatever it takes to get into one of those companies.”
The part he calls unhealthy, “very reminiscent of 2021”: in the last six months, seven of his seed companies raised Series As within weeks at 4–5x his entry — “basically just paying 10 times the price for the same exact company I invested in a couple months ago, plus or minus a few hires.” His latest funds are exiting the J-curve “way too quickly” at 30–40% IRR within months of deployment: “there should be negative IRR during the deployment period… the industry is getting a little drunk on IRR.”
The staying-private flywheel: mega-funds can only scale because company scale scaled — now a dozen to 15 private companies can absorb $500M–$1B checks while still out-growing their public comps. The unintended benefit is enforced hunger: “early liquidity was the bug and not the feature of crypto — people got rich too quickly and they stopped building things.” Tender offers of 1–5% buy “a down payment on a house, not $40 million and generational wealth.” The Google cautionary stat as he tells it: over 100 people made millions, “and 92 of them were never heard of again — and the other eight became VCs.” Good for company building; “problematic from a venture returns perspective” since no LP wants liquidity in 3–5% installments — hence the new wave of continuation vehicles.
On LPs, his taxonomy: “the ones with imagination and the ones without.” The imaginative ones — fund-one backers of non-obvious managers — “have done dramatically better” than those who only back spinouts with underwritable track records (which likely overstate future access anyway). His exemplar: Paula Volent (likely) — Bowdoin’s endowment chief, now Rockefeller, day-one LP to Chase Coleman, early to Josh Kushner, with Druckenmiller chairing the IC — was Abstract’s first institutional LP, and her kind of add “was a signal to the market: pay attention to this fund.”

Meta’s 15 Billion Dollar AI Bet & The Race To ASI w/ Salim Ismail & Dave Blundin The Rise Of Crypto Treasury Companies