TMTB Podcast Intel is produced by Ethmos, research agents that listen to every episode in TMTB’s coverage universe and write the briefs below.
Follow OpenAI or APP on Ethmos to get a brief whenever their executives go on a podcast. TMTB readers get 14 days free and 25% off their first month.
Today’s Essential Episodes:
OpenAI CFO Sarah Friar on 70% Quarterly Growth, a $1.25 Trillion Valuation and Winning Over Data-Center Towns
Squawk on the Street (CNBC), live from Citi’s TMT conference in Pebble Beach, with Sarah Eisen, and OpenAI CFO Sarah FriarChamath Palihapitiya and David Sacks on the Data-Center Freeze, Hyperscalers as Kingmakers and SpaceX’s 90-Day Out With Anthropic
All-In Podcast, with Jason Calacanis, Chamath Palihapitiya, David Sacks and David FriedbergJosh Chandley, Matt and Vincent on AppLovin vs. Unity: Axon’s Lead, an AI Arms Race That Could Squeeze Margins and Silver Lake’s Unity Stake
Mobile Signal, with Josh Chandley, Matt and VincentBen Thompson and Andrew Sharp on America’s Vote Against Data Centers, Why Hyperscalers Win the Bottleneck and Google’s AI Mode
Sharp Tech, with Ben Thompson and Andrew Sharp
Produced by Ethmos: AI-assisted summaries of public podcast episodes. For information only. May miss context or contain errors. Verify before relying on them. Summaries are not substitutes for the episodes, which belong to their creators. Not investment advice, offer, or solicitation. Not affiliated with or endorsed by the shows, hosts, guests or companies discussed. Corrections and publisher opt-outs: support@ethmos.ai.
Listen · Follow OpenAI on Ethmos
THE SIGNAL
Ads launched early 2026 and reached a $1B run rate by August, under 200 days; Friar claims ChatGPT handles nearly 20% of Google’s query volume and calls ads a high-gross-margin tailwind for 2027.
Revenue is now split 50/50 consumer/enterprise after enterprise doubled QoQ in Q3; with total growth at 70%, enterprise likely drove most of the quarter’s growth (our read).
Friar says the $1.25T pre-money is almost 50% above the March round; she also floated consumer hardware as a possible next growth option next year.
OpenAI is moving into hyperscaler resale channels, with AWS launched in the summer and GCP announced last week; Friar says enterprise agents come first partly because limited compute is being steered to the highest-value uses.
01 Midterms and the data-center backlash
CNBC’s Sarah Eisen opens by noting that Democrats campaigning against data center construction flipped seats in Texas, Ohio, South Carolina and Georgia in the midterms, framing it as a voter rebuke of the AI buildout. Friar responds that demand for AI is real and that her job is managing through scenarios, pointing to the 70% QoQ growth as evidence usage is genuine rather than hype. She describes a community-engagement playbook — commitments not to raise electricity bills, grid investment with utilities, water use framed as roughly two Olympic pools per data center with most water serving on-site workers, and a 300-foot berm in Georgia to control noise — alongside local investments like funded preschool lunches (a reported 70% lift in kids eating meals) and a small-business AI summit. When Eisen presses that the policy rollout looks reactive given its timing right after the election, Friar insists the community approach predates the vote and reflects grassroots engagement rather than reacting to cable-news framing.
02 Growth, jobs, and enterprise traction
Friar cites OpenAI’s Q3 figures as the core evidence of demand: 70% sequential growth, enterprise revenue doubling, and 1.2 billion people weekly across ChatGPT, including 300 million for health/wellness and 200 million for personal finance. On jobs, she references that morning’s jobs report (productivity up, unemployment down, wages up) as consistent with OpenAI’s internal economic research, claiming roughly 40% of ChatGPT usage falls outside a user’s formal job description, and that 90% of tasks on her own finance team are now agentic. She cites Cisco deploying OpenAI technology to speed its roadmap by 20% and save a large number of coding hours monthly, plus JPMorgan Chase adopting the tools for cyber use cases.
03 Cyber, safety, and the Navier-Stokes model
Friar frames cybersecurity as a strategic enterprise priority, noting the newly launched GPT-6 Cyber and warning that Chinese models are becoming increasingly cyber-capable, which she says requires arming defenders with the best tools. As an example, she says OpenAI found and helped fix two major PHP vulnerabilities in August before public disclosure, protecting what she estimates as 70% of the internet. On safety more broadly, she says OpenAI has paced the frontier by holding back a model that solved the Navier-Stokes Millennium Math problem (relevant to things like artificial heart fluid dynamics) until internal and regulatory confidence in its safety is established — an assertion made without further detail on timeline or specific risks identified.
04 Products: agents, ads, and chips
Friar describes the proactive agentic assistant OpenAI launched at Dev Day: Uber started it with about 1,000 employees and is now rolling it out to 30,000. Asked how this compares to Meta’s Muse, she says OpenAI deliberately started agentic deployment in the enterprise because the tasks are harder, the security bar is higher, and compute constraints favor starting where value is maximized — later extending it to the 1.2 billion-person consumer base. She also discusses OpenAI’s first-party chip program, announced over the summer, as a way to serve customers at lower cost and a long-term competitive advantage tied to the AGI mission.
05 Fundraising, valuation, and the Anthropic comparison
Pressed on reports of a lower valuation and smaller raise than expected, Friar firmly denies this, reiterating the $1.25 trillion pre-money figure and calling it evidence of Q3 momentum carrying into 2027; Eisen notes this compares to an earlier reported $1.4 trillion target. Friar says investors were impressed by the 70% growth rate at OpenAI’s scale despite a rough October for the public AI trade and an ugly prior trading day. Asked whether Anthropic’s S-1 — which Eisen cites as showing 85% gross margins and 45% compute margins — creates pressure, Friar declines to share OpenAI’s own margins, instead stressing OpenAI’s broader platform (consumer plus enterprise, multiple frontier modalities) and new hyperscaler partnerships with AWS and Google Cloud. On IPO timing, she repeatedly frames fundraising as opportunistic, says a public offering, which she ties to spreading AI’s economic value broadly, is in OpenAI’s future while today is about managing the business, and says the midterm results have not changed OpenAI’s plans. Asked by Eisen about a reported sale of data-center loans that included one tied to Stargate’s New Mexico buildout, Friar does not directly address the report, instead asserting that capital remains available for the strongest, most exciting stories.
THE SIGNAL
Chamath argues that if constraints push spot GPU-hour and $/MW-month prices high enough, even long-term offtakes turn negative-EV, so 90-day outs get exercised and capacity goes to Google/Meta-type balance sheets rather than OpenAI-type renters.
Chamath says SpaceX likely negotiated its 90-day Anthropic out because it saw this coming, which flips the clause from a feared SpaceX risk into pricing optionality; repricing at spot, he says, could lift SpaceX’s revenue from Anthropic 2-3x.
Sacks says gigawatt sites cost about $50B and need certainty; with governors in the biggest data-center states leaning toward moratoria, he expects the 2028 pipeline to close or shrink sharply and calls Gavin Baker’s calm read too optimistic.
Chamath expects incoming governors to drop tax breaks and demand measurable local economic impact upfront, so data-center deal terms likely get costlier for hyperscalers investing tens of billions (our read).
01 The $4 trillion wipeout
Jason opens by noting NVIDIA, SpaceX, Tesla, AMD, Micron and related stocks suffered what he calls the worst single-day dollar loss in stock market history, part of a near-$4 trillion market cap shed across the week. He links this directly to uncertainty over whether data centers will proceed, says it casts a cloud over Anthropic’s IPO, and cites a Wall Street Journal report that a valuation haircut could now be baked in, putting Anthropic’s expected IPO valuation well below what Gavin Baker had projected on the show.
02 Texas as a bellwether
Sacks frames Democrat Gina Hinojosa’s defeat of longtime Texas governor Abbott as the night’s biggest shock, attributing it not to culture-war issues (which she reportedly declined to engage on in a New York Times interview) but to economic populism around data centers — the perception that they raise utility bills and consume water while well-connected companies profit. Chamath adds color from a viral video of SpaceX employees Riley and Katie presenting a six-gigawatt Cochise County, Arizona data-center plan (promising 40% lower electricity costs, no water use, 20% lower property taxes) to a hostile audience, concluding the backlash reflects real anger at seemingly gratuitous tax breaks for the most powerful companies rather than genuine engagement with the facts.