Ramtin Naimi on the State of Venture Investing - [Invest Like the Best, EP.499]

Conduit

• Ramtin Naimi · Abstract VenturesBullish

  Conduit's brain-computer interface shows scaling data, a potential trillion-dollar category.

  Abstract seeded Conduit nine months ago, when brain-computer interfaces were not a category investors cared about. Conduit's model improved with more data and could infer thoughts with about 80% accuracy before a person typed them. It has about 30,000 hours of human training data and plans a million-hour run. Naimi says that if scaling holds, a wearable device could understand thoughts, and a company in this category could reach a trillion dollars. The round was heavily oversubscribed, with 13 term sheets.

  “you could see a world where five, ten years from now, you and I have a little device that we wear over our ear that understands your thoughts and communicates with our agent.”at 28:00

Etched.ai

• Ramtin Naimi · Abstract VenturesBullish

  Startups like Etched can beat incumbents in AI chips; the market only saw it later.

  Most investors assumed a startup could not beat incumbents at AI chip fabrication. Sequoia's $10B valuation of Etched changed that view, and other chip companies then raised very high valuations. Naimi argues that the founder saw the opportunity before the broader market did, and that the people who met Gavin Uberti thought he was brilliant even though the market doubted the business model.

  “a startup company cannot outcompete incumbents on, like, AI chip fabrication. That's just never gonna happen. And it wasn't until Sequoia blessed them with a $10,000,000,000 valuation that everyone was like, oh, wait a second.”at 28:50

Stripe

Across the shows: 1 speaker in other episodes, 0 bullish and 1 bearish calls, past 30 days

• Ramtin Naimi · Abstract VenturesNeutral

  Stripe at $179B is a great company, but its price implies only about an 11-12% five-year IRR

  Ramtin contrasts Stripe's 2021 valuation of $100B with today's $179B. He calls Stripe arguably one of the best companies, but says buying it at that price gives roughly an 11-12% IRR over five years. He notes that is still better than the bottom 75% of venture investments, so backing the best companies at the highest prices can still earn top-quartile returns. The point is that the price, not the business, now limits upside.

  “Stripe was a great company. It still is a great company, arguably one of the best. But today, Stripe is a $179,000,000,000, which is great. But that's what, like an 11, 12% IRR over five years”at 33:39

Chai Discovery

• Ramtin Naimi · Abstract VenturesBullish

  Chai Discovery is a next-generation AI-native drug company worth exposure, with a $500B-$1T opportunity

  Ramtin names Chai Discovery as one of the companies he ranks very high among the next-generation companies he expects to matter about five years out. He sees it as a candidate to build an AI-native next-generation Genentech, with a $500B to $1T market opportunity that grows if it executes on its ambitions with strong teams, investors and capital.

  “I would put, like, New Limit and Chai Discovery really high up on there, Isomorphic really high up on there in terms of IO and in terms of can you build AI native next gen Genentech. And the market opportunity for those companies is 500,000,000,000 to trillion dollar outcomes.”at 44:18

Isomorphic Labs

Across the shows: 2 speakers in other episodes, 1 bullish and 1 bearish calls, past 30 days

• Ramtin Naimi · Abstract VenturesBullish

  Isomorphic is a next-generation AI-native drug company worth exposure, with a $500B-$1T opportunity

  Ramtin ranks Isomorphic very high among the next-generation companies he expects to be important about five years from now, and as a candidate to build an AI-native next-generation Genentech. He puts the market opportunity at $500B to $1T, larger if it executes on its ambitions with the best teams, investors and capital.

  “Isomorphic really high up on there in terms of IO and in terms of can you build AI native next gen Genentech. And the market opportunity for those companies is 500,000,000,000 to trillion dollar outcomes.”at 44:18

NewLimit

• Ramtin Naimi · Abstract VenturesBullish

  New Limit is a next-generation AI-native drug company worth exposure, with a $500B-$1T opportunity

  Ramtin lists New Limit among the next-generation companies he expects to matter about five years from now. He ranks it highly because it is trying to build an AI-native next-generation Genentech. He puts the market opportunity at $500B to $1T, and says it is bigger if the company executes on its ambitions with the best teams and investors.

  “I would put, like, New Limit and Chai Discovery really high up on there, Isomorphic really high up on there in terms of IO and in terms of can you build AI native next gen Genentech. And the market opportunity for those companies is 500,000,000,000 to trillion dollar outcomes.”at 44:18

Applied Compute

• Ramtin Naimi · Abstract VenturesBullish

  Applied Compute pioneered owning your own knowledge intelligence and is the leader in that category

  Ramtin sees owning your own data and knowledge intelligence as a unique idea. He says Applied Compute probably pioneered it and is the category leader, so he will be watching its execution closely. The view is that it has an early lead in a differentiated category.

  “I think the idea of owning your own data, owning your own knowledge intelligence is a really unique idea. And I think the company that probably pioneered that as the leader in the category is applied compute.”at 44:41

Physical Intelligence

Across the shows: 2 speakers in other episodes, 1 bullish and 0 bearish calls, past 30 days

• Ramtin Naimi · Abstract VenturesBullish

  Physical Intelligence is super interesting and could become a multi-hundred-billion to trillion-dollar company

  Ramtin calls Physical Intelligence super interesting and says it could become a multi-hundred-billion to trillion-dollar company. He cites an exceptional team and a strong ability to attract and retain top talent. Talent retention is the factor he weighs most in venture, so a company with that density gets extra bullish from him.

  “I think physical intelligence is super interesting. I think that could be a multi $100,000,000,000, trillion dollar company, exceptional team, a huge magnet for talent, huge ability to retain exceptional talent.”at 44:55

Venture capital & startups

• Ramtin Naimi · Abstract Ventures

  Follow-on tranches priced at 3-5x are a necessity for capital-intensive startups, not an unhealthy sign.

  Investors who lose a hot primary round now buy into a second, pricier tranche at 3-5x, running alongside the primary. In 2021 the equivalent fast-follow was about 2x. Naimi argues this is not a sign of market unhealth because the companies are so capital-intensive that a primary tranche cannot cover their needs and founders do not want to raise it all at once. He notes that everyone gets 'cocked' sometimes, whether they are in the early tranche or the second.

  “I don't think that tranching is a thing that's an unhealthy sign of the market. I think it's actually a necessity of these types of companies and the environment that we're in.”at 3:11

• Ramtin Naimi · Abstract Ventures

  Today's fast deal pace is healthier than 2021 because it reflects real value creation.

  Naimi says a similar deal pace occurred in 2021, but he judges today's pace more healthy because the companies are creating real value and doing impressive things. He hedges that the market may be 2019 rather than 2021, and he distinguishes the enthusiasm now, where companies are reaching product-market fit faster, from 2021, when investors did not expect category-defining companies within 18-24 months.

  “I think the pace is more healthy today because there's real value creation. These companies are doing truly amazing things.”at 4:42

• Ramtin Naimi · Abstract Ventures

  Term sheets are being renegotiated at the highest rate since 2021, a sign of froth.

  Naimi says the market shows 2021-like behaviour: term sheets are issued very quickly and then pulled, reneged on or renegotiated. In the last six to eight weeks he has seen more offers renegotiated after being made than at any point since 2021. He also says the pace of deals is not unprecedented and the market may be closer to 2019 than 2021, so the evidence is mixed rather than conclusive.

  “And in the last six, seven, eight weeks, I've probably seen a larger amount of deals get renegotiated after being offered than I have since 2021.”at 8:54

• Ramtin Naimi · Abstract Ventures

  Mega venture funds are raising 2-3x faster and in larger sizes, underestimating the capital needed for ownership.

  Mega platform funds are raising two to three times as much capital in an 18-24 month window, and deploying about three times faster than planned. Naimi says these firms underestimated how many companies they wanted to back, how much those companies would cost, and how much they would need to invest to reach their target ownership. Examples: one $2.5B fund started deploying in July and plans to return in January; another raised $3B nine months ago and is now seeking $5B. LPs face a choice between less ownership and fewer shots on goal. Some firms are committing to deals before their capital is in place, with funding delayed to mid-January or paid out in tranches over eight to twelve weeks.

  “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.”at 11:07

• Ramtin Naimi · Abstract Ventures

  Only two or three more trillion-dollar companies beyond SpaceX, Anthropic and OpenAI, not forty.

  Best-case startup valuations have risen from about $10B in 2016 to $100B in 2021 and now to a trillion dollars. Only a few companies from the 2021 cohort reached $100B, and most of the others were priced as if they could. With SpaceX, Anthropic and OpenAI already at a trillion or more, Naimi bets only two or three more will follow. Most other companies priced at trillion-dollar outcomes will not get there.

  “Then now between SpaceX and Anthropic and OpenAI, you have three trillion dollar companies. My bet is there might be two to three more and not 40 more.”at 24:52

• Ramtin Naimi · Abstract Ventures

  Startup round prices are erratic; term sheets can differ fourfold and founders anchor to peers' valuations.

  In a more normal market, a founder could be told roughly what valuation a good partner would offer. Now the range of term sheets is about four times wide. Investors may pay several times the next bid to avoid losing a deal, and founders often anchor on the valuation of a friend's company with similar traction. Naimi says he cannot predict where deals will price.

  “I think you can get a term sheet as low as here and as high as here, and that's like a four x difference in price.”at 31:47

• Ramtin Naimi · Abstract Ventures

  Consensus strategy of passively piling large checks into a few top AI companies is likely not a good bet

  Ramtin sees a record number of VC firms in the last 18 months converging on the idea that a handful of obviously great companies can absorb huge capital, and that firms with access to these names can raise large funds and act as passive investors. He argues the strategy has become consensus and easy money, and he can't recall easy strategies that were also good long-term strategies at this scale, so crowded consensus of this kind is probably not a great way to make money.

  “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. So I think that's kind of the interesting thing about this market is if everybody believes that's an easy way to make money, it's probably not a great way to make money.”at 34:45

• Ramtin Naimi · Abstract Ventures

  Silicon Valley's density of angels writing large personal checks keeps it the best place to start companies

  Ramtin argues Silicon Valley has an order-of-magnitude or greater density of individuals writing $100K to $1M personal checks before companies are institutional, and that no other ecosystem has matched this. Because these angels don't mind losing the money, the mentality supports innovation. He expects wealth created by early employees and big IPO outcomes to keep feeding more angel investing, which keeps the region the top place to start a company.

  “I think it just continues making Silicon Valley the best place in the world to start a company. I think the reason that no other ecosystem has ever taken off the way Silicon Valley has is because Silicon Valley has the largest density”at 38:20

• Ramtin Naimi · Abstract Ventures

  Current venture opportunity set rates 10 out of 10 versus his history, the most companies he wants to back

  Ramtin rates the opportunity set a 10 relative to his roughly ten years in venture. He says this is the largest number of companies his firm wants to invest in of any vintage, and the excitement extends beyond the best companies to the broader set of companies people want to fund.

  “It's the largest number of companies we found that we want to be investors in relative to any other vintage. In terms of opportunity set, it's a 10.”at 43:25

• Ramtin Naimi · Abstract Ventures

  Venture market frothiness is about 6.5-7 out of 10, below 2021's 10, though growth justifies some of it

  Asked to rate frothiness on a 1-10 scale, Ramtin gives 2021 a 10 and the current market a 6.5 to 7. He says frothiness is present, but a lot justifies it: companies are growing and reaching product-market fit at an unprecedented rate, which warrants a premium. The market is therefore frothy but not at the 2021 extreme.

  “I would call 2021 a 10, and I would probably call this market a 6.5 to seven. Because while the frothiness is there, there is also a lot to justify the frothiness”at 43:46

AI: overall outlook

• Ramtin Naimi · Abstract Ventures

  AI products are reaching product-market fit faster than ever, spreading even to non-adopters.

  Naimi says companies are reaching product-market fit faster than ever. He uses AI products daily from his own portfolio, which he could not say about his 2021 investments. He says family members who would never have been early adopters of tech now use these products as part of their daily life, which he takes as evidence of a real paradigm shift.

  “You're having companies find product market fit at a faster rate than ever.”at 9:22

• Ramtin Naimi · Abstract Ventures

  Demand for AI training data is underestimated and will keep growing, supporting a tens-of-billions data business

  Ramtin argues people have dramatically underestimated the data needs of model builders, and that those needs are only increasing. Some data companies are ahead of the curve on what data they need, can deliver it at the required quality, and charge accordingly for harder-to-collect data. He sees no end to these needs, so data can keep being sold for 10 to 20 years and generate tens of billions of revenue. He also notes a split camp that sees data as a means to an end and sells it to fund its own robotics research.

  “I just believe data is a gigantic business, and we can continue selling data for the next ten, fifteen, twenty years and collect tens of billions of dollars of revenue as a result of doing it.”at 36:01

AI infrastructure & data centers

• Ramtin Naimi · Abstract Ventures

  Compute is scarce and demand is strong; portfolio companies struggle to secure it.

  Naimi's portfolio companies need help finding compute. Large neoclouds are not interested in deals below about $200M, so smaller companies can only lease small amounts of capacity. The main drivers of compute demand are the hyperscalers, the frontier labs and well-funded private companies, including neolabs raising $1B or more. Demand would only fall if those companies' compute needs shrink or if they can no longer raise capital.

  “The thing that we're seeing in our portfolio companies is that we are routinely helping them find compute because finding compute for them is not easy.”at 20:45

• Ramtin Naimi · Abstract Ventures

  Debt financing for AI compute will get more expensive after neocloud stock weakness.

  Many companies building AI systems need debt worth five to ten times the equity they raise. Naimi says the binding constraint is more likely the debt side than the equity side. Recent public-market weakness in some neoclouds should push up the cost of that debt. The question is whether they can borrow on terms that keep their businesses viable.

  “I think a lot of these guys that are building these things, they need debt in excess of five to 10 x of the equity that they raised. Given what happened in public markets with some [of] the [neoclouds] in the past couple of weeks, I think their cost for debt is going to increase.”at 23:22

Enterprise AI adoption & AI software

• Ramtin Naimi · Abstract Ventures

  Headline AI startup revenue growth flatters quality: low margins, concentration and annualized monthly revenue.

  Naimi says fast-growing AI companies rarely have the clean, high-margin, recurring revenue of classic SaaS. Many have very high growth with low or negative margins, heavy customer concentration, consumer or usage-based revenue, or annualised monthly revenue, as in data sales. Data-labelling companies grow quickly at 15-20% margins, and companies reliant on foundation models hand most of their margin to those providers. He advises founders to compare growth against companies with similar business models rather than the headline top-line figures.

  “It's not a clear, clean, pure play SaaS revenue, 90% plus gross margins, annualized recurring revenue contracts. They all have something that's either ridiculously high growth but very low margins. Either low margins or negative margins.”at 4:58

AI pricing & unit economics

• Ramtin Naimi · Abstract Ventures

  AI startups' gross margins are improving quickly, with some moving from negative to 35% positive.

  Naimi says margins for AI companies are improving rapidly. Anthropic's margin was reported at 85% in the first half of the year. Portfolio companies expected to grow 200-700% this year are moving from margins of negative 15% to positive 35%, so revenue growth and margin expansion are happening together.

  “But the cool thing that you are seeing is that these margins are improving at really rapid clips... We have companies in our portfolio that are gonna grow 200 to 700 this year, but their margin profiles are going from negative 15% to positive 35%.”at 6:06

AI & jobs

• Ramtin Naimi · Abstract Ventures

  AI software could capture large labor budgets like paralegal spending, though labor replacement is not yet visible.

  Vertical AI companies that sell to law firms look too small when measured against law firms' software budgets. Law firms spend far more on paralegals, and Naimi argues that capturing even 20% of that spending makes the scale work. He admits he has not yet seen these companies replace labour, and that some investors do not accept the labour-replacement thesis. Their revenue is still growing five to ten times a year at large scale, which is why he believes the math can work.

  “But you look at how much money that these law firms spend on paralegals every year, and you're like, oh, that's actually a really big number. So if these companies can capture 20% of that, the scale of it doesn't work.”at 6:32

Private equity

• Ramtin Naimi · Abstract Ventures

  Most LP venture liquidity depends on just three or four names like SpaceX, Anthropic and OpenAI.

  LPs with 30-50% of their assets in venture are in fact exposed to three or four companies, such as SpaceX, Anthropic, OpenAI and Databricks. Those few names are what will generate liquidity. LPs whose funds lack exposure to them have received little liquidity, so funds without those holdings must look for new capital sources. Naimi says LPs expected much more liquidity at this point in the cycle than they have received.

  “Those are the names that need to generate liquidity for these guys to get their exposure levels to the asset class down.”at 13:02

Frontier AI labs & model race

• Ramtin Naimi · Abstract Ventures

  Neolab capital needs are huge, and benchmark results may overstate commercial readiness.

  Naimi estimates there are now 80-90 neolabs that have raised $1-5B each, up from two or three expected at first. One neolab he met was raising $1B for a single training run next year and expects to need about $5B more. It is hard to tell whether impressive launches are generalizable or tuned to benchmarks, and many of these companies will not be commercial for the near future. He wonders how many will keep raising.

  “And they were gonna spend the entire billion next year on compute for one training run, and they said they would probably need to raise another 5,000,000,000 throughout next year.”at 22:17

Sports, betting & prediction markets

• Ramtin Naimi · Abstract Ventures

  Prediction market companies are growing insanely fast as true consumer businesses, not AI companies

  Ramtin says it will be interesting to see how prediction markets play out over a few years. But he points to how fast prediction market companies are growing today, and says they are not AI companies but true consumer businesses. The current growth rate is the evidence he cites for their strength.

  “I think it'll be very interesting to see how the prediction markets play out in a few years. But currently, I mean, the rate at which those companies are growing is insane, and they're not even AI companies. They're just true consumer businesses.”at 44:55