Feb 13, 2026 · 51m · neon-show
Is AI changing EVEN the Top Roles in startups?
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
Venture capitalist Umesh Padval joins Siddhartha Ahluwalia on The Neon Show to analyze the economic realities of generative AI, founder evaluation frameworks, SaaS consolidation, and the expanding US-India cross-border corridor. Drawing from four decades of Silicon Valley leadership, Padval offers actionable insights into sustainable unit economics, executive team scaling, and long-term enterprise value creation.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Siddhartha holds 10.4% of the talking time here. How this is scored →
Siddhartha as informed peer 3.6 Guest teaching 5.0 Guest disagreement 0.5 Siddhartha pushing back 0.9
speaking balance: gold is Siddhartha, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 5:00 Dismantling vibe-coding hype and lack of defensibility
Padval pushes against uncritical enthusiasm for coding assistant startups, warning that foundation model owners will integrate those capabilities while startups struggle with negative unit economics.
Hardest push from Siddhartha ▶ 23:00 Host challenges public market viability for mid-tier ARR companiesAhluwalia presses Padval on the structural squeeze faced by companies growing modestly at $200M ARR when public markets demand billion-dollar revenue profiles.
Biggest teaching moment ▶ 47:30 Explaining NVIDIA as a systems and software businessPadval reframes common market perceptions by detailing how NVIDIA's true moat lies in CUDA, developer software libraries, and data-center networking rather than just silicon.
Siddhartha holds their own ▶ 23:00 Host articulates banker incentives and market precedenceAhluwalia demonstrates market fluency by analyzing how mega-cap private rounds distort public liquidity and leave mid-stage enterprise software firms without banker sponsorship.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Siddhartha as informed peer | Guest teaching | Guest disagreement | Siddhartha pushing back | Why |
|---|---|---|---|---|---|---|
| Analyzing the AI Landscape, Infrastructure, and Startup Unit Economics | 4 | 6 | 2 | 2 | Ahluwalia prompts Padval on the fast-growing coding assistant landscape and unit economics. Padval breaks down the stark cost reality behind rapid ARR growth, pointing out that $500M in revenue requires massive LLM and compute outlays that crush margins. | |
| Investment Thesis Behind Cohere and Enterprise AI Differentiation | 3 | 5 | 0 | 0 | Padval articulates Thomvest's early investment thesis for Cohere, highlighting enterprise security, cloud-agnostic deployment, and geographic talent cost advantages in Toronto over Silicon Valley. | |
| Evaluating Founder Traits and Scaling Executive Teams for IPOs | 4 | 4 | 1 | 1 | Ahluwalia asks whether the balance between technical and commercial founders has shifted in AI. Padval explains that while deep technical ability is essential early on, scaling to IPO still requires seasoned executive hires across GTM and operations. | |
| Building Founder Relationships and Preempting Venture Capital Deals | 3 | 5 | 0 | 1 | Padval contrasts rapid seed deals done in three hours through existing founder trust with multi-month strategic relationship building needed to preempt competitive venture rounds. | |
| The Partnership Behind Jyoti Bansal and Harness's Growth | 3 | 4 | 0 | 0 | Padval details his longstanding relationship with serial entrepreneur Jyoti Bansal and the thesis of dominating continuous deployment before expanding across the entire DevOps stack. | |
| State of the IPO Market, M&A, and Private Capital | 6 | 5 | 1 | 4 | Ahluwalia pushes on the structural dilemma facing mid-tier growth companies ($200M-$300M ARR) when elite outliers stay private longer. Padval outlines the three realistic paths: strategic sales, defensive mergers for scale, or discounted IPOs. | |
| Salesloft-Clari Merger Dynamics and Public Profitability Benchmarks | 4 | 5 | 0 | 1 | Padval outlines the rationale behind the Salesloft-Clari merger, explaining how combining TAM and cutting redundant costs accelerates free cash flow breakeven required by public markets. | |
| Unlocking the Future of the US-India Cross-Border Corridor | 5 | 5 | 0 | 1 | Ahluwalia shares Neon's mission of expanding the US-India corridor relative to Israel. Padval draws on his board experience with Mellanox to argue that India's deep engineering bench can make the corridor 100x larger over the next decade. | |
| Separating Signal from Noise in Early-Stage AI Investing | 3 | 5 | 0 | 0 | Padval discusses vetting early-stage AI startups through open-source monetization dynamics, technical stack defensibility, and direct sanity checks with his network of enterprise CIOs. | |
| Umesh Padval's Journey From IIT Bombay to Thomvest Ventures | 2 | 5 | 0 | 0 | Padval recaps his four-decade path from IIT Bombay and Stanford to AMD, VLSI, public company CEO roles, and venture leadership across Bessemer and Thomvest. | |
| Behind the Scenes with Jensen Huang and NVIDIA's Dominance | 3 | 6 | 1 | 0 | Padval shares personal anecdotes about Jensen Huang's management practices, including having 60 direct reports without one-on-ones, and explains how building CUDA transformed NVIDIA from a chip maker into a complete systems platform. |
Statements from this episode (18)
Padval: AI cost per token dropped 95% in two years
“Luckily, over the last two years, the cost per token has come down 95%”
Padval: $500M ARR in AI coding requires billions in compute costs
“What I heard is the five hundred million of ARR needs about six hundred million of investment from the LLM company they have to pay, which also requires hosting costs from Amazon and the hyperscalers of seven hundred million, and that needs chips worth eight h…”
Padval: Cohere refused hyperscaler capital to stay cloud-neutral
“They didn't take investment from any of the hyperscalers. They had opportunities to do that, but they chose not to. So give the user and the enterprises a choice of any cloud. Rather than be tied to one cloud.”
Padval: Toronto offers lower costs and less churn for AI talent than Bay Area
“In the Bay Area, the cost of that talent, you know, is very expensive. Whereas this talent is in Toronto, which has a great University of Toronto where all the AI researchers coming. The cost in Toronto is low. The churn is less.”
Padval: Repeat founders make up 75% of his 40+ private investments
“I've invested over 40 private companies now, and I've seen the trend. The repeat founders is 75% of my investments, and the reason is they've gone through it once, and so they've learned from it and don't make the same mistake again.”
Padval funded StackGen in seven days based on a prior relationship
“And we made a decision in the seed round in literally three hours... We were convinced of that, that we were able to present the following Tuesday to our IC, and then make the decision, go through IC, and actually fund the company in seven days, because we wer…”
Padval: Preempted Relyance AI Series B after 9-month courtship
“Reliance AI, which I led around series B, along with Microsoft and Menlo Ventures and others... I had to actually court him for nine months, develop a relationship where he got confident with me, And then I was able to preempt the round because he got total co…”
Padval: Harness grew 10x in five years and 30-40% this year
“The company has grown 10 X last year in five years, and this year another 30, 40% growth, which is incredible.”
Padval: Over 500 software companies above $300M ARR are awaiting IPO
“I was at the conference last 10 days back, and The banking conference they said there were over 500 companies waiting to grow IPO just in the software space, which are over three hundred million in ARR.”
Padval predicts Databricks and Stripe will IPO in 9 to 12 months
“And so they will go public in the next nine to 12 months, I think.”
Padval: Unprofitable $200M ARR startups must sell, merge, or discount IPO
“There's a bunch of companies which are not cash flow positive, and growing slowly, as you said, and so if you're at 203 hundred million going slowly, and it's competitive, two things have to happen. They have to sell to the large company, Or they merge within …”
Padval: Clari grew ARR 20x after 2018 Series B investment
“Since I invested in Clary in 2018, they grew more in 2018. It was a series B investment at very early. They've grown since they are 20 X on ARR.”
Padval: Cross-border startups work best with founders in both the US and India
“One founder is in India, And one founder is in the U.S. That's the best combination set up from the get-go, even better than trying to set up something and coming here.”
Padval: US-India cross-border startup ecosystem could grow 100x in 15 years
“So I'm very optimistic That India is falling very fast and could be a hundred X of the companies over next 10 to 15 years just like Israel did.”
Padval: AI companies mirror open-source software through viral, easy adoption
“AI companies are starting to be like open source companies because it's viral with the easy to use product.”
Padval: Cyber defense is an asymmetric problem ensuring endless market demand
“Cybersecurity is an asymmetric problem. The hackers, all they need is one IP address to get into a enterprise and then spread from there and disrupt the company. Whereas the enterprises have to be protecting it a hundred percent and no one's going to protect i…”
Padval: NVIDIA CEO Jensen Huang manages 60 direct reports
“He has 60 reports. I don't know any CEO who can manage 60 reports, but I think Jensen can.”
Padval: Jensen Huang avoids one-on-ones to prevent workplace politics
“He says, one-on-one means politics, because I say something in a different context to one person, they go off and use that context to get things done. And sometimes reality of what our conversation was and what happens is different. So he says, when I get a re…”