Jul 6, 2026 · 55m · sourcery
Accel: The Quiet Firm Behind Facebook, Cursor, Nebius, Lovable, Vercel · Sourcery with Molly O'Shea
Arun Mathew · 14m spoken Matt Miller · 12m spoken Molly O'Shea · 11m spoken Miles Clements · 11m spoken
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 episode of the Sourcery podcast, host Molly O'Shea interviews Accel growth partners Arun Mathew, Miles Clements, and Matt Miller to discuss Accel's investment culture, their history from Facebook to modern AI leaders, and their deep theses on AI infrastructure, developer tooling, cybersecurity, and future public market capacity.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Molly holds 24.3% of the talking time here. How this is scored →
Molly as informed peer 4.1 Guest teaching 2.7 Guest disagreement 0.3 Molly pushing back 0.3
speaking balance: gold is Molly, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 42:42 Miles challenges token overconsumption narrative
Miles directly counters the conventional media narrative about token waste by citing Accel's proprietary developer survey where CFOs instructed teams to accelerate rather than restrict spend.
Hardest push from Molly ▶ 44:22 Molly tests the limits of token expenditure optimismMolly challenges the panel with a direct bullish/bearish test on half-billion-dollar monthly token spends and teases Matt's predictable optimism before probing for the eventual tipping point.
Biggest teaching moment ▶ 23:37 Matt explains vertical integration and inference economicsMatt provides a clear technical tutorial on why proprietary hardware and vertical integration are essential for next-generation hyperscalers handling non-linear agentic inference curves.
Molly holds their own ▶ 39:48 Molly connects tool integration risks to agent threat surfacesMolly synthesizes insights from her interview with Merge to demonstrate how unmonitored employee-driven agent connections create explosive new vulnerability surfaces for enterprise security.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Molly as informed peer | Guest teaching | Guest disagreement | Molly pushing back | Why |
|---|---|---|---|---|---|---|
| Accel's Growth Team Structure and Strategy | 4 | 2 | 0 | 0 | Molly opens with lighthearted banter about Accel's office aesthetic and references public market data from Coatue regarding Meta's historic IPO. The guests recount Accel's early growth fund days and the Series A/secondary Facebook investments. | |
| Generalist Mindset and Specialized Focus in AI Investing | 4 | 3 | 0 | 0 | Molly inquires how a generalist team develops high-conviction theses in rapidly moving AI categories. Miles and Arun explain how Accel balances deep thematic domain focus with cross-sector exploration. | |
| Accel's "Silent Partner" Ethos and Winning Competitive Deals | 4 | 1 | 1 | 0 | Molly characterizes Accel as a 'silent killer' compared to flashier VC competitors and asks how they win competitive deals. Miles gently reframes the label to 'silent partner' and shares how organic founder references won them the Cursor deal. | |
| Navigating Global AI Opportunities Across the Tech Stack | 3 | 2 | 0 | 0 | Molly asks how late-stage funding dynamics have shifted amid current market frenzy. Miles outlines Accel's full-stack, multi-stage, multi-geography approach managing a $7B AI portfolio. | |
| Massive Check Sizes and Private Market Late-Stage Returns | 5 | 3 | 0 | 0 | Molly points to recent massive quarterly capital concentrations into select AI labs. Arun and Matt explain why larger check sizes are economically rational and how late-stage top-quartile returns are converging with early-stage returns. | |
| The Expanding Trillion-Dollar Market and Office Interlude | 3 | 1 | 0 | 0 | Molly brings up conversations with Brian Singerman on parking capital late stage, followed by an office interlude joking about Loonin water and mid-roll sponsor spots. | |
| Nebius Deep Dive: Building the Next-Generation AI Hyperscaler | 4 | 5 | 0 | 0 | Molly asks for a breakdown of Nebius and why inference infrastructure is gaining strategic importance. Matt provides an in-depth explanation of Arkady Volozh's background, vertical integration, and inference demands. | |
| GTC Takeaways and the $150 Million Nebius PIPE Investment | 4 | 2 | 0 | 0 | Molly shares her on-the-ground observations of Nebius marketing at Nvidia GTC. Miles highlights Matt's contrarian $150M PIPE investment into Nebius that grew over 13x. | |
| Enterprise Agentic Adoption, Supabase Growth, and Vibe Coding | 4 | 4 | 0 | 0 | Arun details Supabase's rapid inbound growth driven by enterprise agentic adoption. Molly shares a colorful anecdote about her non-technical father using vibe coding to build a marine navigation app. | |
| The "Agentic Influence" Thesis and Interconnected Portfolio Dynamics | 5 | 3 | 0 | 0 | Molly questions whether AI agents organically recommend products or if there is paid placement, citing Coatue's growth charts. Arun and Miles introduce their 'agentic influence' thesis connecting Supabase, Linear, and Lovable. | |
| Investment Focus: Developer Tooling, Infra, and Ex-xAI Teams | 5 | 2 | 0 | 0 | Molly probes into Accel's category bets and investment in RadixArk. Matt and Arun discuss being deliberately slow on the application layer while aggressively targeting infrastructure bottlenecks. | |
| Expanding Threat Surfaces and Value Accrual in AI Security | 5 | 4 | 0 | 0 | Molly connects the rise of agent connections to insights from her Merge interview regarding vulnerability surfaces. Arun details Accel's Cyera investment journey and explains how AI widens security threat vectors. | |
| Token Maxing Debates and Developer Spending Habits | 5 | 4 | 2 | 3 | Molly presses the panel on token maxing and asks whether massive token bills will trigger budget limits. Miles counters using Accel developer survey data showing CFOs urging teams to increase spend, while Matt defends multi-hundred-million-dollar monthly spends. | |
| Sponsor Announcements: VCX by Fundrise, Public.com, and Deel | 4 | 3 | 1 | 1 | Following mid-roll ad reads, Molly asks how public markets will absorb impending mega-IPOs like SpaceX. The panel argues that index inclusion and retail access will smoothly absorb the capital. | |
| Real-World AI Impact, Medical Applications, and Accel's Next Generation | 3 | 2 | 0 | 0 | The episode concludes on future outlooks: Matt highlights real-world AI utility, Arun gives a case study of hospital ER triage AI, and Miles praises Accel's next generation of rising investors. |
Statements from this episode (31)
Miller: Accel relies on about six investors for late-stage capital deployment
“So like on the late stage in terms of investors, we've got about six investors that deploy most of that capital.”
Clements: OpenAI and Anthropic parallel Facebook's platform era
“The thing that still kind of inspires us, you know, through the Facebook parallel was, it was like the first platform company that we saw. So all the companies As associates that we were building relationships with, we're figuring out how they could like build…”
Mathew: Accel bought Facebook secondary shares around a $20B valuation
“We, so we initially invested in Facebook in 2005. We led the series A out of our early stage fund, and we created this growth vehicle in 2008. Miles and I joined in 2008 and 2009, and one of the first opportunities was to buy some Facebook secondary that came …”
Mathew: Accel modeled Facebook secondary investment at 5X upside
“And so I remember that partner meeting around the table saying, we actually think this could be a hundred billion dollar business or more that we could generate north of a five X. Lo and behold, that was way under shooting what the opportunity and the potentia…”
Clements: VCs must know who will win before the first meeting
“Like, you have to have an outside-in viewpoint on which company is going to win before you even get the first meeting, and then you have to really show up and add some value and you know, Convince this entrepreneur that there's a compelling reason to work with…”
Miller: Being the first institutional investor was Accel's religion historically
“Up until about five to seven years ago, it was almost a religion here to be the first institutional investor into a company. And as a result, you're almost organizationally introverted. You're spending your time going and finding something that someone hasn't …”
Miller: VCs making founders' funding rounds about themselves is 'cringeworthy'
“I mean, we always, I think we always kind of looked at the social media posts about a founder raising around and a VC making about themselves as like pretty cringeworth. Like, don't throw your shoulder out patting yourself on the back type of situation.”
Clements: Accel has assembled a $7 billion portfolio of AI investments
“So holistically, when we reflect on our work over the last couple of years, I think what we see is A seven billion dollar portfolio of really thoughtful, nuanced investments that we've selected very carefully.”
Mathew: Accel now regularly invests $500M to over $1B per company
“Our first growth fund was four hundred and eighty million when we first joined. It is not unusual for us to invest five hundred million dollars into a company, actually to invest over a billion dollars into a company.”
Mathew: AI mega-round funding is rational and AI potential is underestimated
“I actually think the market is totally rational on this. I think if anything, a lot of us have underestimated the potential of AI and the value creation that it offers.”
Miller: Top-quartile late-stage venture returns will soon match early-stage returns
“The returns in the private market at the late stage are actually going to start matching the returns at the early stage. If you look at top quartile performance of early stage funds versus late stage funds, I think they're going to be pretty close to each othe…”
Clements: The market will see $10 trillion companies in the next cycle
“I mean, if you think about it, 10 years ago, there were zero publicly traded companies worth a trillion dollars. Five years ago, there were five companies worth a trillion dollars. As of today, there's 14 with probably three or four private companies that are …”
Miller: Volozh correctly refused to sell his 30% ClickHouse stake to Accel
“And actually the early innings of that was me trying to convince Arkady to sell me his 30% stake of ClickHouse, which he rightfully refused to do.”
Miller predicts AI agents will exponentially increase compute demand for inference
“Agents start using inference in an exponential way. The demand curve goes way up for what we, what we're going to need from an inference perspective.”
Miller: Full-stack vertical integration creates an edge in delivering AI inference
“When you control capacity by owning the data centers and the GPUs, and you control the software, you're in an advantageous position to be able to meet the moment and deliver inference for a market that, you know, we fundamentally believe is going to grow expon…”
Clements: Accel's $150M PIPE investment in Nebius is up 13x in 16 months
“16 months ago, a hundred and fifty million dollar investment that I think is up 13 x today, and the world is now awakened to this company and how special it could potentially be.”
Mathew: Supabase is growing 350% at hundreds of millions in scale
“So we can take Supabase, for example, which is the backend database to a lot of agent workflows and agent applications. That's growing 350% at hundreds of million dollars of scale, and they have virtually no salespeople. It's all inbound.”
Mathew: Most agentic workflow usage is coming from inside enterprises
“Most of the usage that we're seeing is actually coming from inside the enterprise.”
Mathew: AI models are the new distribution channel, favoring best developer UX
“In this era it's AI optimization. And so these products are now the distribution mechanism for all the downstream products and services that you can use. And they tend to prioritize the tools that have the best developer and user experience.”
Mathew: 60% of Y Combinator companies now choose Supabase
“So, 60% of YC companies now choose Supabase.”
Mathew: Supabase grew from under 1M to 9M developers in over a year
“They just crossed nine million developers You know, last week when we first invested, it had under a million developers, and that was at the beginning of last year.”
Clements: Accel Doubled Down on Cursor as a Workflow Choke Point
“We were investors in cursor at the time, but we certainly saw a lot of, you know, cursor as a choke point for downstream workflow creation and downstream tool recommendation. That probably emboldened us to make our second investment in cursor.”
Miller: Accel moved slowly on AI apps due to foundation model risks
“We were admittedly slower than our peers on the application layer for AI because we were very worried about how fickle they were when the wakes of the models were getting wider and wider, and we spent a lot of time in infrastructure, and we're continuing to sp…”
Miller: Accel Is Investing in Ex-xAI Inference Startup Radix Arc
“We are investing in pure play inference software providers like Ying and the team at Radix Arc that just came out of XAI. They're the inference team there.”
Mathew: AI model releases create tailwinds for cybersecurity platforms like CrowdStrike
“When Mythos first came out, I think the stock market thought that Mythos was going to kill a bunch of our security companies. Just look at CrowdStrike's stock price over the last six weeks. It actually is such a tailwind behind a lot of these companies, especi…”
Mathew: Cyera is the highest-valued private security company
“And so, they're actually one of the, I think they're the highest valued private security company on the market.”
Clements: 7x more CFOs are telling developers to increase AI token spending
“We did this survey of developers, and we asked them how many of them had a CFO who was telling them to spend less on Consumption versus more, and we actually found that seven times more companies are being told to, like, let it rip and spin more.”
Miller: Monthly AI token spend will go well above $500M
“I'm very bullish that we're going to go well above that.”
Miller: Token infrastructure forecasts will consistently be revised upward
“I think it's going to be just like every underestimated infrastructure forecast. We're going to spend X and we always spend X plus some multiple of that every single quarter. Everyone revises up. I think you're going to see revisions up on tokens for sure.”
Mathew: Elon Musk is reserving 30% of the SpaceX IPO for retail
“And so Elon reserving 30% of the SpaceX IPO for retail investors, I think is a really great thing.”
Mathew: US Financial Markets Will Absorb Multi-Trillion-Dollar Tech IPOs
“And so I think this is just the power of the financial markets in the United States specifically, I think we can absorb it. I think we will. I think you'll see a number of other countries and investors from all over the world and mom and pop, you know, it's no…”