Jan 23, 2026 · 1h 35m · allin
Coinbase CEO's Top 3 Crypto Trends for 2026 + More from Davos!
Jason Calacanis · 34m spoken Andrew Feldman · 21m spoken Brian Armstrong · 19m spoken Jake Loosararian · 10m 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
Broadcasting live from the World Economic Forum in Davos, host Jason Calacanis interviews industry leaders Brian Armstrong (Coinbase), Andrew Feldman (Cerebras Systems), and Jake Loosararian (Gecko Robotics). Together, they explore emerging trends in crypto regulation, next-generation AI chip hardware, energy grid demands, physical-world robotics, and the macro shift toward economic growth.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 39.8% of the talking time here. How this is scored →
The hosts as informed peer 5.7 Guest teaching 5.1 Guest disagreement 1.6 The hosts pushing back 2.7
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:15:35 Andrew Feldman Rejects AI Layoffs Premise
Andrew Feldman forcefully rejects the host's framing that current tech layoffs are driven by AI, correcting the premise by explaining it is actually the delayed impact of SaaS tools reducing middle management scope.
Hardest push from the hosts ▶ 11:39 Host Probes Coinbase Exposure to Tether RiskJason pushes back against Brian Armstrong's cautious stance on Tether, questioning whether listing or facilitating Tether trading introduces unacceptable regulatory and collapse risks for Coinbase.
Biggest teaching moment ▶ 1:02:05 Andrew Feldman Teaches GPU Memory Architecture BottlenecksAndrew Feldman thoroughly educates the host on the three pillars of computer architecture, explaining how GPU memory limitations stall AI inference and why hyperscalers are spending billions on fast-memory architectures.
The host holds their own ▶ 19:23 Jason Demonstrates Deep Private Market ExpertiseJason draws on his extensive venture capital background to detail how private secondary market SPVs have devolved into predatory boiler rooms charging high upfront fees to unsophisticated investors.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Brian Armstrong on Global Banking Partnerships | 5 | 3 | 1 | 2 | Jason sets up the interview by drawing parallels between crypto regulation and earlier tech battles like Uber and Airbnb. Brian explains Coinbase's growing institutional partnerships with five of the top twenty global banks and credits the political shift toward clear rules. | |
| The GENIUS Act & Stablecoin Reserve Models | 6 | 4 | 1 | 2 | Brian explains the 100% short-term US Treasury reserve requirement for stablecoins under the GENIUS Act compared to fractional reserve banking. Jason highlights his firsthand experience during the Silicon Valley Bank run to illustrate liquidity risk. | |
| Coinbase Business Incentives & Banking Pushback | 6 | 4 | 2 | 5 | Jason presses Brian on whether incumbent banking trade groups are actively trying to retrade the GENIUS Act. Brian clarifies the legal distinction between yield and activity-based rewards while defending the law as a strict industry red line. | |
| Stablecoin Dynamics: USDC Compliance vs. Tether | 6 | 3 | 2 | 6 | Jason challenges Brian on the risks of listing Tether given its regulatory scrutiny and historical lack of full audits. Brian diplomatically notes Tether's utility in high-inflation emerging markets while emphasizing USDC's compliance focus. | |
| Platform Standards & Rating Crypto Assets | 5 | 5 | 2 | 3 | Jason asks if Coinbase should issue credit-rating style scores for listed crypto assets. Brian explains why Coinbase relies on minimum listing standards instead, citing past failures with user star ratings where traders simply voted their portfolio interests. | |
| Tokenizing Private Markets & SEC Modernization | 7 | 3 | 1 | 3 | Jason demonstrates deep venture knowledge when discussing private market tokenization, criticizing predatory SPV brokerages charging high loads without carry. Brian agrees that on-chain capital formation should be modernized through updated SEC accreditation rules. | |
| Coinbase Tokenized & Expanding Investor Access | 5 | 4 | 1 | 2 | Brian introduces Coinbase Tokenized and shares data showing four billion adults globally lack brokerage access. Jason contrasts wealth building through stock ownership against retail gambling on sports betting apps. | |
| California Policy Failures & Wealth Flight | 6 | 3 | 1 | 2 | Jason and Brian discuss California's fiscal crisis, wealth tax proposals, and executive departures. Jason notes how sky-high housing costs make recruiting out-of-state talent nearly impossible for employers. | |
| Brian Armstrong's Longevity Venture: New Limit | 5 | 4 | 1 | 1 | Brian outlines his longevity startup New Limit and its progress in epigenetic reprogramming. Jason cites current macroeconomic figures to argue that pro-growth private sector policies outperform government spending. | |
| AI Work Displacement & Internal Coinbase AI Agents | 6 | 5 | 1 | 3 | Brian describes internal Coinbase AI agents that reverse-prompt executives on schedule management and organizational strategy. Jason references self-driving taxi pushback in China and his hands-on experience using Claude for workplace automation. | |
| Introducing Andrew Feldman & Cerebras Wafer Scale Engine | 5 | 6 | 1 | 2 | Andrew Feldman displays the physical Cerebras Wafer Scale Engine and explains its architecture. He educates the host on chip size scaling, transistor density, and cost structures across cloud and on-premise systems. | |
| Early AI Compute Workloads Before LLMs | 6 | 6 | 1 | 2 | Andrew outlines early AI workloads in national labs and pharma before the rise of LLMs. Jason describes how AI deep research cut his podcast preparation workflows from forty-eight hours down to minutes. | |
| Cerebras-OpenAI Cloud Partnership & Speed as Moat | 6 | 6 | 1 | 3 | Andrew details Cerebras's 750-megawatt deal with OpenAI and explains why megawatt capacity has superseded square footage as the primary metric for data center scale. Jason references classic search research showing latency reductions drive user engagement. | |
| Water-Cooling Closed-Loops & Data Center PR Myths | 5 | 6 | 2 | 2 | Andrew corrects public misperceptions regarding data center water consumption, clarifying that modern AI hardware utilizes closed-loop chilling systems. Jason compares the public narrative to golf course water usage. | |
| Community Relations, Utility Guarantees, & Grid Resilience | 6 | 5 | 2 | 2 | Andrew critiques how tech hyperscalers mishandled rural community relations by causing local utility rate hikes. Jason suggests deploying residential battery storage to smooth grid demand and share benefits with host communities. | |
| US Power Grid Underinvestment vs. Nuclear (SMR) Future | 6 | 5 | 1 | 2 | Andrew points out four decades of US power grid underinvestment compared to China. Jason brings up French nuclear power subsidies for local residents as a model for incentivizing energy infrastructure. | |
| Evaluating Orbital Compute & Space-Based Data Centers | 5 | 7 | 2 | 3 | Andrew evaluates Elon Musk's proposal for space-based solar data centers. While validating solar collection benefits, he educates the host on space vacuum thermal dissipation and cross-satellite communication bottlenecks, putting the realistic timeline at eight to ten years. | |
| Enterprise AI Demand Curve & Inference Scaling | 6 | 7 | 2 | 3 | Andrew breaks down the fundamental balance of computation, memory storage, and transport bandwidth. He explains why GPU memory architecture limits fast inference and details the supply-chain bullwhip effect causing current memory shortages. | |
| Andrew Feldman on Chipmaking Talent and US-China AI Race | 5 | 6 | 2 | 2 | Andrew describes the geographic concentration of top chip design talent in Santa Clara while noting China's rapid progress in open-source AI models and grid buildouts. | |
| Evaluating US AI Policy, Global Allies, and Industrial Competition | 7 | 6 | 3 | 3 | Andrew praises policy shifts that allow chip exports to global allies like the UAE while offering a nuanced take on China restrictions. Jason highlights Huawei's historical expansion across global telecom infrastructure and Chinese automotive dominance in Latin America. | |
| Labor Market Impact, Middle Management, and Dilbert Tribute | 6 | 7 | 4 | 4 | Andrew explicitly rejects the host's premise that current tech layoffs are driven by AI, arguing instead that SaaS productivity tools flattened middle management information-relaying roles. Jason challenges him on upcoming job displacement dynamics. | |
| Interview with Jake Loosararian: Davos ROI Shift and Gecko Robotics | 5 | 5 | 1 | 2 | Jake Loosararian discusses the pivot at Davos away from virtue signaling toward concrete AI ROI. He points out that industrial asset owners face major data gaps before AI can deliver productivity gains. | |
| Gecko Robotics in Defense, Infrastructure, and Industrial Data | 6 | 6 | 2 | 3 | Jake explains Gecko's defense and energy work inspecting military ships and infrastructure welds. He dismisses consumer humanoid applications like folding laundry as low ROI compared to solving critical industrial bottlenecks. | |
| Physical World AI, Dangerous Labor, and Humanoid Robotics Outlook | 5 | 6 | 2 | 2 | Jake estimates a three-year timeline for physical task AI execution while noting the lack of internet training data for dangerous industrial environments. Jason concludes the interview program. |
Statements from this episode (69)
Armstrong: Five of top 20 global banks use Coinbase infrastructure
“Five of the top 20 global banks are now using Coinbase to build their crypto infrastructure into their products.”
Armstrong: Biden administration unlawfully tried to kill US crypto industry
“The Biden administration really tried to unlawfully kill this industry in America”
Armstrong: 52 million Americans have used cryptocurrency
“There's like, fifty-two million Americans who've used crypto now.”
Armstrong: China announced interest payments on central bank digital currency
“China just announced that they're going to pay interest on their central bank digital currency.”
Armstrong: Roughly 500 million people globally have used crypto
“Something like five hundred million people have used it globally.”
Armstrong: Bitcoin was the best-performing asset class of the last decade
“Bitcoin was the best performing asset class of the last decade.”
Armstrong: GENIUS Act requires US-regulated stablecoins to hold 100% Treasury reserves
“Under the Genius Act that got passed into law last year, US regulated stable coins have to have 100% of the assets stored in short term US treasuries.”
Coinbase passes ~100% of stablecoin yield economics to active rewards users
“So when customers do that, we actually pass along about a hundred percent of the economics to them for holding those stable coins with us.”
Armstrong: Undoing the GENIUS Act is a red line for Coinbase
“I think there's the bank trade groups, which I believe are trying to undo the genius act. Got it. It just got passed into law four months ago. And for us, that's a red line.”
Armstrong: USDC is the largest regulated stablecoin globally
“USDC is the largest regulated stablecoin, because it is compliant under Genius in the US, they're compliant under Mika in Europe, et cetera.”
Armstrong: Coinbase does not have an exclusivity deal with Circle
“We don't have like an exclusive with circle or anything like that. We actually list other stable coins on our platform.”
Armstrong: Tether is not compliant under the US GENIUS Act
“It's not currently compliant under the genius act in the U S and so it doesn't follow those same hundred percent reserves in short term U S treasuries is my understanding.”
Armstrong: Coinbase lists any legal crypto asset meeting minimum cybersecurity standards
“So what we try to do is have minimum listing standards. So if we believe that there's a cybersecurity risk to it the developer could, you know, rug everyone, or if it's certainly, it's illegal from a compliance point of view, you know, there's a few different …”
Armstrong: Coinbase attempted user crypto ratings, but abandoned them due to user bias
“We tried making user ratings at one point. It didn't go that well because people were basically voting with their, whatever they liked, you know.”
Armstrong: All assets are coming on-chain for trading
“So anyway, we're long, I think that, yeah, the biggest trends are all assets are coming on chain for trading. Prediction markets are growing like crazy, and stablecoin payments are growing like crazy. Those are the, probably the three biggest trends in crypto …”
Armstrong: B2B cross-border payments were crypto's top growth area last year
“So the biggest growth area over the last year has been B to B cross border payments and cross border.”
Armstrong: On-chain capital formation will be massive for private companies
“But anyway, I think on-chain capital formation is going to be massive for private companies.”
Armstrong: Companies will eventually go public entirely on-chain
“I think eventually, like, you'll actually just be able to go public totally on-chain too.”
Armstrong: Coinbase launched Coinbase Tokenized for funds and real estate
“Coinbase launched a product actually called Coinbase tokenized. So we're helping any fund or real estate project or anybody who wants to tokenize their products”
Armstrong: BlackRock and Apollo publicly stated plans to tokenize all products
“BlackRock, Apollo, like these, the top funds in the world are putting, they've come out publicly and said they want to tokenize every single one of their products.”
Armstrong: Four billion adults globally are unbrokered without investment access
“There's actually four billion adults who are unbrokered as well, which means they don't have any ability to invest in these high quality assets.”
Armstrong: Coinbase integrated an AI agent to teach users financial strategies
“We've integrated one into Coinbase app. It can kind of teach people about dollar cost averaging and, you know, tax loss harvesting.”
Armstrong estimates 20% of California billionaires left, creating $10B tax deficit
“My estimates would probably be 20% of the billionaires have already left in California, that it's already created a negative ten billion tax hole, even if, even with the amount that they hope to raise from the people who stay.”
Armstrong: California budget rose dramatically over 10 years while services worsened
“The history of the last 10 years in California is that the budget has gone up dramatically and the services have gotten worse.”
Armstrong: New Limit demonstrated cell reprogramming to restore youthful function
“Within the first two or three years, they were able to successfully demonstrate reprogramming of human cells to sort of function they had when they were younger.”
Armstrong: New Limit's first drug candidate likely enters clinical trials next year
“And the first drug candidate is probably going to go into clinical trials next year.”
Brian Armstrong: Poorest populations fare best in high economic freedom nations
“Even the poorest people in society are, they do the best in high economic freedom countries that are, anyway.”
Jason Calacanis: US GDP growth reached 5.6%, highest since COVID stimulus
“We have not seen this level of GDP since we pumped a bunch of money and printed a bunch of dollars during COVID. 5.6 GDP is pretty spectacular.”
Brian Armstrong: Economic growth comes from deregulation and energy, not government spending
“It seems like growth does not come from government spending, right? That's like the key that this is this Keynesian economic argument that I think is basically wrong. Like growth comes from having like deregulation, having low cost energy, allow the private ma…”
Armstrong: AI agents will use stablecoins and crypto wallets for payments
“AI agents, I think, are gonna use stablecoins and wallet, crypto wallets.”
Armstrong: AI job displacement will lead to net economic abundance
“I basically, I think if the AI plays out as we all think it will with robots and a lot of, there will be job displacement, but it means that we'll be in a world of more abundance and people are going to have jobs that there's like, they're streaming on video g…”
Armstrong: Coinbase deployed internal AI agents connected to all company data
“One of the big pushes we made in the last year was we got our own internal hosted AI model that was connected to all of our data sources, right? So it's like every Slack message, every Google doc, every Salesforce data confluence, you know, so now this is all …”
Feldman: Cerebras Wafer Scale Engine is 56x larger than Nvidia B200
“And so this is, say, 56 times larger than a B 200. And it's a four trillion transistor part.”
Feldman: Cerebras spent $500M to make its first chip
“Well, the first one cost us half a billion to make.”
Feldman: On-premise Cerebras systems cost $1M to $1.5M
“On premise, they're about a million, million and a half, depending on how you have it configured.”
Feldman: Sam Altman and Ilya Sutskever accurately predicted current AI in 2015
“We talked to them in 2015, and what they were saying is now. It was unbelievable how right they've been.”
Feldman: Nvidia's stock price remained flat between 2004 and 2010
“If you look at NVIDIA's stock price from 2004 to 2010, it was flat.”
Calacanis: AI deep research yields 5x human output in under 10 minutes
“And when I tell you they would ask for 48 hours to do a dossier, that was 20% of what this does in under 10 minutes.”
Calacanis: AI tools replaced human assistants for podcast guest research
“Now it's 16 hours to I don't need them. Literally don't need them to do this work.”
Feldman: Cognition uses Cerebras to power its AI coding engine
“So we have customers, ah, like Cognition, who use us to power their coding engine.”
Feldman: OpenAI is a Cerebras customer, but Anthropic is not
“And Claude is not a, Anthropic is not a customer, but we recently announced OpenAI.”
Feldman: Slow AI model latency causes immediate user churn
“While you're waiting for Claude, or you're waiting for ChatGPT, you get a coffee, or you poke around somewhere else, and you've lost the customer. That's the cost of being slow, is the customer has gone somewhere else.”
Feldman: Cerebras announced 750-megawatt deal with OpenAI
“What we announced was, 750 megawatts, is what was announced.”
Feldman: Power footprint is always the limiting constraint on data centers
“So the limiting constraint on data centers is always their power footprint. And right now for large deployments, the limiting constraint is how much power can be delivered.”
Feldman: Hydro is world's cheapest power source, followed by natural gas
“So the cheapest power in the world is, is Hydro. Without question. After that is natural gas.”
Feldman: Cerebras was among the first production AI systems using water cooling
“We were some of the first production AI systems to use water.”
Feldman: Funding local infrastructure would have prevented data center pushback
“If instead you'd have gone in and said look, great. We, we're going to be good citizens. We're going to be big taxpayers here. Let's build more schools. We can build a school for you. It's a rounding error in the cost of this facility. We're going to make a bu…”
Calacanis: Microsoft guaranteed data center power won't raise utility costs
“They basically said, just to catch the audience up, we guarantee you our usage of energy will not increase the cost of your utilities.”
Feldman: Small modular reactors probably won't power data centers for 3-4 years
“I think it's both the obviously the right thing to do and probably not the source of data centers for the next three or four years.”
Feldman: Nuclear power is vastly more efficient over 20-30 years than alternatives
“Nuclear has, ah, over 20 or 30 years is vastly more efficient than any other power we know how to create.”
Feldman: Space-based data centers are 8 to 10 years away
“I think it's in the eight to 10 year category, not in the three to five.”
Feldman: Very few enterprises have meaningfully adopted AI workflows yet
“I think we are still really early in the demand for AI compute, and I think if you think about, you know, what portion of enterprises have really adopted AI in a meaningful way that has changed their workflow, it's tiny.”
Feldman: GPUs face a major inference bottleneck due to slow memory speed
“The GPU, for example, has a lot of capacity of memory, but it's really slow. All right, and that's a huge bottleneck in inference. It's why they can't be fast.”
Feldman: Memory market will take 18 months to digest surge, keeping prices high
“And it will take us about 18 months to digest. The prices will stay high.”
Feldman: Santa Clara houses six of the world's top ten chip teams
“Within a few square miles of Santa Clara you had Intel, you had AMD, you had NVIDIA, you have our team, you have ARM, you have one of ARM's great teams. You have maybe six. You have six of the world's great 10 chip teams.”
Feldman: China has pushed ahead in open-source AI models
“I think on the other side, I think they have pushed ahead in the open model category. The open source model is an area where they've pushed ahead.”
Feldman: China leads US in expanding electrical grid capacity for AI
“I think because they're a top down economy, they were able to Make decisions like we're going to bring a huge amount of power onto our grid. We're going to modernize our grid. So they were able to bring on huge amounts of power, and that's something that we're…”
Feldman: Recursive AI systems compound early leads into winner-take-all outcomes
“Systems that are recursive, right, that build on themselves at rapid rates, have a huge winner-take-all feel, right, right? By getting ahead, you get further ahead, your iteration speed accelerates, And even small differences at the beginning are magnified ver…”
Feldman: Huawei clobbered the U.S. in developing world telecommunications
“They clobbered us in Africa. They clobbered us in the developing world. They absolutely ran the table.”
Feldman: Cerebras is navigating local regulations across 14 data center locations
“I think one of the things that kills a company like us is trying to grow extremely quickly. Is having to deal with different regulations in each of 14 localities we're trying to put data centers.”
Calacanis: Chinese government subsidizes BYD electric vehicles by 30-50%
“There's no way that the government isn't subsidizing those by 30, 40, 50%.”
Feldman: Middle management layoffs stem from SaaS efficiency, not AI
“What I think has happened is this is the delayed impact of good SaaS tools. That what's happened is your ability to extend your reach as a leader and as a manager to stay abreast of what's happening. Your scope is much, much bigger. And so the role of middle m…”
Loosararian: Data Gaps Limit AI ROI in Heavy Industry
“A lot of it comes down to this really interesting gap that exists in AI, which is like all the information and data set that you need to actually turn all of this into actual return on investment and productivity gains, especially for these, like, large infras…”
Gecko Robotics generates roughly 30% of its business from defense
“We do about 30% of our businesses defense.”
Gecko Robotics tech delivers 90% faster manufacturing per US Navy Admiral
“Admiral Houston has talked about 90% improvements to speed and manufacturing using the technology that Gecko builds.”
Loosararian: Folding laundry and washing dishes are low-ROI robot tasks
“Folding laundry and cleaning dishes is not a high ROI use case.”