EP.012: Hydra Host: "The GPU Whisperer"

Free Personalized AI Moderator Trial “If you want to generate alpha, you can’t be sharing all this information.” “That interviewer needs to think and feel and act like you do. It needs to have all the context that you do.” “We’re just making primary research faster, cheaper, and embedded. If I had to summarize the whole thing, that’s what Qualitate is doing.” Sagar Kadakia has spent nearly 20 years in primary research. He joined Tom DelVecchio at ETR in 2011, ran buy-side survey work for six years, and later built research and data products at 7Park Data and CB Insights. In 2023, as ChatGPT and voice capture crossed human parity, he founded Qualitate: an AI interviewer that runs 45–65 minute expert conversations autonomously and is grounded in what he describes as the fastest-growing AI-moderated interview library in the world. Crew Capital wrote the first $1M check when few others wanted to fund a company building its own expert panel. This week, Qualitate announced a $25M Series A led by Insight Partners, with net revenue retention above 300% and run-rate revenue up 10x in nine months. Sagar walks through Qualitate’s two products: a proactive library built using Qualitate’s own capital and questions, and a private channel-check product where clients bring their own questions, interviewer, library, and internal inputs — without that proprietary information being mixed back into the broader system. We also get into why the architecture matters: grounding instead of retraining for every new model, persistent memory so call five knows calls one through four, and transcripts structured so investors can query them for time series instead of simply searching. “You can search through transcript files, but you can’t actually query that data. On Qualitate, you can actually query transcripts.” “We don’t train or fine-tune anything. Because if you do that and the next version comes out, all of that work you just did goes to zero. We ground it.”

Carmen Li, CEO of Silicon Data and Compute Exchange, joins the show to explain how compute is becoming a brand-new asset class. A former energy options market maker at DRW who later ran global data integrations at Bloomberg, Carmen was pre-seeded by DRW founder Don Wilson to build the GPU industry benchmarks. Silicon Data is the referee — "the Bloomberg for compute" — while Compute Exchange is the liquidity layer where GPUs, tokens and servers change hands. We get into why token price indexes are falling, recent token & model consumption trends, why GPU residual values are above accounting depreciation schedules, and how neoclouds, their lenders, and compute users can hedge GPU exposure to reduce business volatility. • "The A100 has been around for six years at this point, and we still see three-year contracts signed on A100s." Residual values are still hovering above five-year straight-line depreciation. • “we'd definitely be looking at 30 trillion of [GPU] derivatives every year." Topics: Compute as an asset class, GPU indices, CME compute futures, GPU depreciation and useful life, neoclouds, token pricing, AI chip supply chain, 💡 This episode is presented by: Oxford Data Plan–The Home of Alternative Data Not Investment Advice.

My long time confidant, Jarrett Harris, Director of Research at IronAdvisor Insights joins the show to share his latest insights. He is quintessentially mid-western and was my go to source for learning about the on the ground pulse of the industrial economy when I was at Citadel and Millennium. As part of IronConnect, a heavy equipment industry service provider, Jarrett has unparalleled access to equipment dealers, rental companies, and industry executives.

Prediction markets are surging, but DraftKings is moving in the opposite direction. Is the issue market share? Rising customer acquisition costs? Or simply a reset from last year’s growth expectations? In this episode, Bernie McTernan, CFA, Managing Director and Senior Analyst at Needham & Company, breaks down the rapidly changing prediction-market landscape. We discuss Kalshi’s first-mover advantage, DraftKings’ strategy to replicate its regulated sportsbook playbook in currently unregulated states, Novig’s aggressive customer-acquisition push, and whether prediction markets meaningfully expand the long-term TAM for companies like DKNG. With negative estimate revisions weighing on the stock, elevated short interest, and DraftKings trading at roughly 11x forward EV/EBITDA, the debate becomes increasingly interesting: how much of the bad news is already priced in—and could today’s investments create a much larger earnings opportunity over time? “The stock's probably fifty percent lower than it was and the TAM's doubled.” “Prediction markets aren't just sports. Soon you'll be able to bet the whisper number.” “This is gonna be Robinhood on steroids.” “The numbers are going down, short interest is going up. Why do you think that the investments and lower revisions this year are actually gonna have a benefit next year?” Stocks: $DKNG, $FLUT Topics: DraftKings, Kalshi, Prediction Markets, Sports Betting, FanDuel, iCasino, Novig, Polymarket, Customer Acquisition Cost, Market Share, Regulation, EBITDA, Sportsbooks Not Investment Advice [00:00:00] Introduction to Bernie McTernan, Managing Director and Senior Analyst at Needham & Company. [00:01:40] Kalshi's rumored $40 billion private valuation. [00:02:45] What DraftKings' first NFL week of prediction-market share actually looked like. [00:04:30] The retention data that killed the original bear case on sportsbooks. [00:05:00] Mid-teens to 30–40% share — the playbook DraftKings wants to run again. [00:08:00] Why the prediction-market regulatory battle could end up at the Supreme Court. [00:11:40] The data showing almost no customer overlap between Kalshi and DraftKings. [00:18:00] Hedge fund hit rates and betting the number instead of the stock. [00:20:15] Fast futures, crypto, and the iCasino land grab. [00:21:50] The potential path from $1 billion to $5–6 billion of EBITDA. [00:27:45] Why states legalize gambling — budget deficits, not morality. [00:31:25] A trough multiple on estimates burdened by TAM expansion investments. [00:43:45] Novig's $50 million ad and the Caesars cautionary tale. 💡 This episode is presented by: Carbon Arc — The Infrastructure for the AI Economy. Free trial with code PTPM30: https://www.carbonarc.co/onboarding?flow=professional&billing=monthly&source=lenses Research Sponsors: Fiscal.AI — Delivering Modern Financial Data Infrastructure. 15% discount with code PITCH: https://fiscal.ai/?via=pitch Oxford Data Plan — The Home of Alternative Data: https://oxford-dp.com/demo AlphaSense — Decision Grade AI: https://www.alpha-sense.com/Pitch/ Pitch The PM Links: 📩 Subscribe to our Substack for early episode drops, research updates, episode feedback, and the Pitch The PM Job Board: https://pitchthepm.substack.com/ Doug Garber on LinkedIn: https://linkedin.com/in/doug-garber-42aa508 Bernie McTernan, CFA Links: LinkedIn: https://www.linkedin.com/in/bernie-mcternan-cfa-84045a23/ Needham & Company: https://www.linkedin.com/company/needham-&-company/

Chandler Bocklage spent nearly two decades working alongside Steve Cohen, managing a portfolio before leading Business Development at Point72. In this episode, we discuss what separates good investors from great ones, how Point72 develops portfolio managers, why intellectual curiosity matters more than almost anything else, and how AI is reshaping the future of fundamental investing. If you've ever wondered what it really takes to become a successful PM, this is one of the most candid conversations from inside the walls of legendary investment firm Point72. “Alpha decay has been compressing and compressing. There are more and more people chasing the same alphas.” “You have to have different duration, you have to have different catalyst paths, you have to have different thought processes because” “Honestly, I think that's what it really comes down to. The sacrifices…” "We don't think AI replaces analysts or PMs."