Sep 17, 2019 · 28m · mad
Fireside Chat: Sarah Guo, General Partner at Greylock (FirstMark's Data Driven NYC)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this Fireside Chat hosted by Matt Turck at Data Driven NYC, Greylock General Partner Sarah Guo discusses venture capital dynamics, SaaS market shifts, and investment strategies across distributed engineering hubs. She delves into the growing impact of artificial intelligence, evolving cybersecurity architectures, and regulatory tailwinds shaping early-stage enterprise software.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 12.1% of the talking time here. How this is scored →
Matt as informed peer 3.1 Guest teaching 3.0 Guest disagreement 1.1 Matt pushing back 1.1
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 8:01 Reframing the SMB SaaS premise
Sarah politely pushes back on Matt's premise that her thesis is about moving from enterprise to SMB, clarifying that it is really about width of user adoption.
Hardest push from Matt ▶ 13:49 Probing zero trust cloud migrationMatt pushes the technical discussion forward by directly challenging whether cloud shifts require moving from traditional perimeter defense to zero trust architectures.
Biggest teaching moment ▶ 8:20 Explaining market width vs SMB focusSarah educates the host on how user adoption within organizations defines modern SaaS width far more accurately than customer firm size.
Matt holds his own ▶ 13:49 Demonstrating cybersecurity architecture expertiseMatt displays clear industry knowledge by introducing multi-cloud architecture concepts and contrasting perimeter models with zero trust.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Investing Across Distributed Ecosystems and New York Tech | 4 | 3 | 1 | 2 | Matt demonstrates knowledge of Greylock's investments by referencing specific portfolio companies like Screen and asking if it was their first European deal. Sarah clarifies her portfolio distribution across Paris, Dublin, and SoCal while elaborating on the global dispersion of tech talent. | |
| Venture Capital Market Dynamics and Valuation Multiples | 2 | 4 | 1 | 1 | Matt asks a broad question about overall VC market sentiment amid recent tech IPOs. Sarah turns the question around briefly to ask Matt's venture tenure before providing a detailed breakdown of sky-high valuation multiples at growth stages and Greylock's early-stage focus. | |
| Evolution of SaaS: Market Width and User-Driven Adoption | 4 | 4 | 2 | 1 | Matt cites a topic he saw on Twitter regarding SaaS expanding from enterprise to SMB. Sarah gently reframes his premise, explaining that the shift is less about SMBs and more about market width and user-driven adoption. | |
| Growth Drivers and Architectural Shifts in Cybersecurity | 5 | 3 | 1 | 3 | Matt shows strong domain knowledge by asking about specific architectural shifts like perimeter security versus zero trust in multi-cloud environments. Sarah agrees enthusiastically and expands on the need to re-architect security software. | |
| Machine Learning Innovations in Cybersecurity | 4 | 3 | 1 | 1 | Matt notes that Sarah has multiple portfolio companies at the intersection of cybersecurity and machine learning. Sarah confirms this core thesis and explains how data explainability drives value in security products like Obsidian. | |
| Broad Applications of AI Across Software and Consumer Tech | 4 | 3 | 1 | 1 | Matt asks whether machine learning has passed its primary investment window or is now embedded in everything. Sarah responds with a structured analysis distinguishing ML developer tooling from value-creating AI applications across consumer and enterprise software. | |
| Audience Q&A: Adversarial Machine Learning and Model Defense | 1 | 2 | 1 | 0 | Matt moderates an audience Q&A session where audience members inquire about adversarial machine learning and geopolitical investment footprints. Sarah fields questions directly with very little host intervention. | |
| Audience Q&A: Cybersecurity Pricing Models and Value Measurement | 1 | 2 | 1 | 0 | Audience members ask about pricing models when security prevents breaches successfully and the impact of regulations like GDPR. Sarah provides detailed industry insights while Matt handles facilitation. |
Statements from this episode (21)
Sarah Guo: Greylock operates with under 10 investors and under 40 total staff
“Greylock is a fifty-plus-year-old venture firm. We have less than 10 investors, an entire team of less than 40, actually based entirely in the Bay Area”
Sarah Guo: Greylock invests out of a $1B fund focused on early-stage startups
“We have billion dollar fund, and we largely invest in early stage companies and then put more and more capital into them over time as they progress.”
Guo: Half of her personal startup portfolio is outside the Bay Area
“Half my personal portfolio is not in the Bay Area entirely”
Sarah Guo: SaaS market is in early innings of market cap creation
“I think we're just gonna see a lot of market cap creation, so I fundamentally think we're, like, in the early innings of that.”
Sarah Guo: Growth-stage startup valuation multiples have skyrocketed
“Everything is a multiple more expensive than it used to be but at the mid-stage and at the growth stage, when you begin to see more and more data about how a company is doing, and these models are better and better understood, that multiple has skyrocketed, be…”
Guo: Greylock maintains a single fund rather than multi-billion opportunity vehicles
“Unlike many of our you know, colleagues in the venture ecosystem, we have not raised, An extra ten billion dollars and, like, spread it into different funds. We have one fund.”
Guo: Shopify is an extraordinary example of SaaS expanding market width
“Shopify's an extraordinary example of this, but you can have companies that enable entirely new customer bases to do new things, and they've figured out, like, web distribution, Basically, to make a different price point work for different customers”
Guo estimates the 2019 cybersecurity market at $75B to $80B
“Like, let's say in 2019, you can call it a 75 or eighty billion dollar industry, half of which is software and half of which is services.”
Sarah Guo: Cybersecurity software will outgrow cybersecurity services
“And I think the other interesting thing for us to think about is, like, I think the software piece of that is going to grow faster than the services piece.”
Guo: Most of 2019's $40B security software spend remains on-premise
“Most of the, call it, 35 or forty billion dollars of security spend today, it really is still, like, on-premise tooling in a few known categories, the largest of which happen to be, like, network and endpoint and some areas like that.”
Sarah Guo: Greylock's cybersecurity investment thesis centers on hybrid cloud and AI
“I think Matt can probably summarize our, like, investing thesis around security very simply in, like, two things. It's basically like hybrid cloud and AI, right?”
Sarah Guo: Cybersecurity is fundamentally a data inspection problem
“I think of security as a data problem. Right? I think of it as, like, basically, like, you can either re-architect something, or you can constantly inspect mass amounts of data to understand what is actually happening in your environment.”
Guo: Cybersecurity leads machine learning adoption due to massive data volume
“Security is actually like a, Sort of forefront industry for machine learning in some ways because you have a massive amount of different types of data where you have too much data to go inspect manually. And you want to be able to infer a bunch of different be…”
Guo: Five of her six active portfolio companies rely on ML
“I think five of the six companies that I am involved with actively today would say, like, machine learning is a core tenet of, like, what makes your product work.”
Sarah Guo predicts AI value will come from automating labor-heavy services
“I think a lot of the business value is gonna come from Traditionally very, like, manual labor heavy service industry.”
Sarah Guo considers Uber a machine learning company
“I'd argue Uber is an ML company because they use it for demand management and pricing and everything else.”
Guo: Greylock portfolio companies account for adversarial threats to AI models
“I definitely think that there are companies in the portfolio that think very carefully about what it will mean to have adversaries for their models, right?”
Guo: B2B software has not been a profitable industry in China
“B to B software has not actually been very profitable a industry in China so far.”
Guo: Sophisticated enterprises aren't looking to cut security budgets or headcount
“Especially for, you know, medium and large companies that are more sophisticated in security nobody's, like, necessarily trying to bring their overall security budget Down or cut headcount. They're just trying to figure out, like, am I doing things that are im…”
Guo: Greylock seeks investments in GDPR and CCPA compliance tools
“If somebody says they make it easier for Companies to deal with GDPR and CCPA and be compliant, understand what data they have. Like, I'm listening. I think we would actually welcome more of that in the portfolio.”
Guo: Tightening regulations will force companies to re-architect for data residency
“I feel pretty strongly like we're gonna move in one direction here, and companies are going to be asked to know much more about their data and have better controls and architect it differently, including in terms of you know, data residency, which people are j…”