The Twenty Minute VCSaam Motamedi: Why Series B Won’t Make Money & Why $1M ARR is a BS Milestone for Series A | E1177
CHAPTERS
- 0:00 – 1:03
AI investing is in a bigger bubble than 2021: valuation dislocation and shaky retention
Saam argues AI investing is in an exuberant bubble, with seed and growth-stage revenue multiples far out of line with public comps. He explains why explosive early growth can mask uncertain long-term retention and defensibility, especially in prosumer AI apps.
- •Seed rounds reaching $10s of millions to $100M+ post-money for very early teams
- •Private AI multiples (100–200x revenue) vs public leaders closer to ~15–20x forward revenue
- •Corporate/strategic investors distort pricing because they’re not purely return-motivated
- •Key risk: unclear persistence of user value and retention despite fast initial adoption
- 1:03 – 2:25
How childhood shaped Saam: debate-driven competitiveness and research lessons on small teams
Saam describes growing up in Houston, moving to California, and how policy debate instilled a deep competitive drive. He also shares how biomedical research taught him the leverage of small, high-performing teams—an idea that carries into his investing style.
- •Policy debate fostered comfort with high-stakes competition
- •Competition maps directly to winning founders and deals in venture
- •Biomedical research experience reinforced the power of small teams
- •Early formative experiences influence his investor mindset today
- 2:25 – 6:33
Is venture a young person’s game? Wiring, stamina, and constant competition
Harry and Saam debate whether venture favors youth. Saam largely agrees but emphasizes the real determinant is personal wiring—relentless work ethic and responsiveness to opportunities—citing his partner Ashim as proof.
- •Venture is intensely competitive; lagging feedback hides underperformance
- •Some older investors remain elite due to pace and obsession
- •Being ‘wired for it’ matters more than age
- •High-quality founder opportunities demand immediate action
- 6:33 – 7:58
When to ignore hype: combining explosive data with fundamentals and market dynamics
Saam explains Greylock’s two-lens approach: respect breakout metrics, but pressure-test core fundamentals like retention, market structure, and defensibility. He notes early AI app darlings have already seen growth cool, validating this discipline.
- •Lens 1: strong data shifts question from ‘why invest’ to ‘why not?’
- •Lens 2: fundamentals—PMF, retention, defensibility—still decide
- •Comfort with walking away even after ‘0 to $20M’ traction if dynamics are poor
- •Early AI apps have already shown post-hype deceleration
- 7:58 – 10:30
Differentiating in crowded AI categories: the same SaaS questions still win
Addressing the proliferation of similar AI tools, Saam argues the differentiation problem isn’t new—it’s classic SaaS. He returns to basics: founder quality, product depth and stickiness, and unique distribution that compounds.
- •Crowding is familiar; AI may increase competitor count but not the core problem
- •Back the best founder/management team with a clear product point-of-view
- •Defensibility comes from workflow depth, stickiness, and pricing power
- •Distribution advantages matter as much as technology
- 10:30 – 12:55
‘OpenAI could kill my startup’: foundational primitives vs vertical workflows
Saam acknowledges OpenAI’s ruthless execution but segments risk by product type. OpenAI will likely dominate foundational primitives (writing, code generation), while durable value can accrue to specialized, workflow-heavy tools in professions and enterprise contexts.
- •OpenAI will aggressively expand into foundational capabilities
- •High risk for startups competing on core primitives like generic code generation
- •Lower risk (and more opportunity) in vertical copilots and end-to-end workflows
- •Large outcomes can still be built on top of foundation models
- 12:55 – 16:03
Where venture can win in the foundation model layer: products over ‘models as a service’
Saam is uncertain about long-term economics of selling models via APIs due to potential commoditization and limited separation. He’s more bullish when the model must be tightly integrated with the product (agents, code, certain enterprise agent use cases).
- •Hard to predict AI’s trajectory; place ‘high error bars’ on forecasts
- •Skepticism about durable margins in pure model/API businesses
- •More confidence in first-party products requiring deep model-app coupling
- •Examples of tight coupling: personal agents, horizontal enterprise agents, code generation
- 16:03 – 21:30
Is SaaS ending? Why AI could enable the next Salesforce (data model, delivery, interface)
Saam strongly disagrees with the ‘end of SaaS’ thesis, arguing it’s one of the best times to build SaaS. He explains that platform-sized outcomes emerge when the data model, delivery model, or interface changes—conditions he believes AI is now creating, especially via agent-driven interfaces.
- •Largest SaaS wins are deep systems of record (Salesforce, Workday, ServiceNow)
- •New horizontal systems have been rare recently because disruption vectors were absent
- •Three disruption levers: data model, delivery/pricing model, and interface
- •AI agents can reshape CRM by inferring truth from raw interactions vs manual fields
- •Agent-driven interfaces could make legacy UIs irrelevant over time
- 21:30 – 23:21
AI monetization and the end of per-seat: pricing power comes from replacing work
They discuss whether AI features lead to customers paying more and whether per-seat pricing dies. Saam expects hybrid models and believes meaningful monetization comes when AI replaces real labor, not when it adds minor copilots to existing tools.
- •Per-seat may persist but will be augmented by usage/work-based pricing
- •Public markets have tempered near-term AI monetization expectations
- •Pricing power improves when AI eliminates roles/tasks (e.g., ‘AI BDR’)
- •Too early to know winners, but replacement value is the key
- 23:21 – 31:05
Seed and Series A pricing today: ranges, power-law logic, and portfolio-level discipline
Saam shares current pricing bands and defends high seed valuations using power-law reasoning: small price differences matter less than getting into the rare iconic outcomes. He also agrees that portfolio-level average entry prices still matter and stresses ownership targets and selective exceptions.
- •Typical seed pricing for strong teams: ~20–40M post; Series A: ~80–200M post
- •Power-law lens: intermediate outcomes don’t justify obsessing over small price deltas
- •Two truths: exceptions can be fine, but portfolio rules must hold overall
- •Greylock optimizes for ownership within a range; aims ~20–25%+ in core positions
- •Large early capital can accelerate great founders but can reduce discipline for others
- 31:05 – 33:00
High standards with ‘irrelevant’ check sizes: time is the real constraint
Harry challenges how a billion-dollar fund keeps a high bar when early checks feel small. Saam responds that Greylock makes very few investments and commits to long-term accountability; diligence is time-intensive and often starts before the company exists via early relationship-building.
- •Greylock partners do ~1–2 investments per year; constraint is time, not capital
- •Writing a check implies long-term responsibility to founders
- •90-day ‘work alongside founders’ diligence can still happen with early relationships
- •Many Greylock positions start small and scale to large capital deployment over time
- 33:00 – 38:07
Signaling is ‘bullshit’: when (and why) follow-on participation shouldn’t matter
Saam dismisses the idea that a prior investor not leading the next round kills a company. He argues good investors evaluate on merits, and large seed platforms do too many deals for non-participation to be a reliable negative signal; what matters more is adding real board value.
- •Saam’s view of signaling: absence of prior lead doesn’t meaningfully deter new leads
- •Example: Cresta—seed, A, and B led by different top-tier firms
- •Founders should prioritize adding truly impactful board members (a short list)
- •Many VCs are not additive; capital-only investors may not be worth adding
- 38:07 – 39:01
Reserves strategy: staying dangerous on offense—and prepared for air pockets on defense
Saam describes a non-formulaic but generally high-reserve approach because early bets can require large pro-rata follow-ons and occasional inside rounds. Reserves aren’t just for doubling down on winners; they’re also essential when strong companies hit temporary setbacks.
- •Greylock reserves heavily due to early entry and long company lifecycles
- •Initial $5M positions can grow into very large total exposure over time
- •Reserves support pro-rata in winners and emergency financing in downturns
- •Inside rounds can preserve outcomes when great companies hit air pockets
- 39:01 – 43:46
Why $1M ARR is a bad Series A filter: founder + market first, ARR second
Saam criticizes using $1M ARR as a primary Series A milestone because many companies reach $1–2M and never scale to meaningful outcomes. He argues Series A returns require hundreds of millions in ARR, so investors should focus first on founder caliber and market dynamics, using ARR as only a supporting PMF signal.
- •Series A economics require companies that can reach ‘hundreds of millions’ in ARR
- •Most companies that hit $1M ARR never reach $10M/$50M/$100M
- •ARR is a proxy for PMF, but not a sufficient predictor of scale
- •Example: Wiz raised an A at a high valuation with ~0 ARR, driven by founder/market timing
- •Contrarian-right can mean paying more for an exceptional founder in a massive market
- 43:46 – 47:15
Great founder, poor market: the ‘zip code vs street’ heuristic and painful misses
Saam explains how Greylock debates founder-market tradeoffs constantly. He shares heuristics: only bend the rule for truly iconic founders, prefer situations where the ‘zip code’ (macro area) is right even if the initial angle is wrong, and be more willing at seed than later—then recounts misses like Glean and Codium that changed his default toward ‘yes’ on exceptional founders.
- •Markets have gravity; late-stage pivots are harder once the ‘plane is flying’
- •Would invest in a bad ‘street’ within a good ‘zip code’ more readily than a bad zip code
- •More flexibility at seed, but still a common source of major mistakes
- •Miss: Glean—overweighted historical enterprise search failures; underestimated founder + upcoming genAI shift
- •Miss: Codium—passed despite exceptional founder; later realized founder-first orientation matters
- 47:15 – 54:18
Series B may not make money: why the ‘game on the field’ can be a trap in frothy markets
Saam argues growth isn’t dead for top companies—if anything, Series B can be frothier than 2021 due to concentrated demand and abundant capital at multi-stage platforms. He warns the vintage-wide Series B basket may underperform unless public multiples re-expand or growth durability proves far stronger than prior software eras.
- •Median-market data misleads; top-tier companies still raise easily
- •New capital sources: multi-stage early platforms with large growth funds and new partners
- •Fewer great companies + more money chasing them = exuberant pricing pressure
- •Concern: Series B asset class may not generate returns as a whole this vintage
- •Rejects ‘play the game on the field’—independent thinking is required
- 54:18 – 59:29
Mastering sourcing, selecting, and servicing: building a durable lane as an investor
Saam believes sourcing is the hardest and most important skill, especially for small teams competing against larger platforms. He advises young investors to focus less on personal deal count and more on creating a genuine reason founders would choose them—often by developing deep domain expertise that compounds into early access and differentiated help.
- •Great investors need all three, but sourcing is the primary bottleneck
- •Small teams must stay relevant and find emerging founder pockets early
- •Young investors hurt themselves by pitching everything as ‘great’ to partners
- •Career advice: develop a clear ‘why you’ value proposition for founders
- •Domain mapping and relationship-building can create proprietary access over time
- 59:29 – 1:07:39
Quick-fire: PMF help, incubation pitfalls, changing views on small models, and choosing VCs
In rapid Q&A, Saam argues investors can meaningfully help pre-PMF through ICP focus and customer learning. He critiques most incubations for cap table issues and negative selection, shares a major view change (focused models can win by moving up-stack), and emphasizes founders should reference-check how VCs actually work.
- •Investors can help pre-PMF: ICP, segmentation, customer meetings, feedback loops
- •Respects Elad Gil’s cross-stage investing and company-building ability
- •Incubations often fail due to taking too much equity and not attracting top founders
- •Changed mind: focused models + fast move into workflow/app layer can be durable (e.g., ElevenLabs)
- •Founders should ask: ‘What would be different without this VC?’ and do references
