The Future of Everything: What CEOs of Circle, CrowdStrike & More See Coming in 2026
Speakers
Jeremy Allaire
— CEO, Circle
George Kurtz
— Founder and CEO, CrowdStrike
Adam Goldstein
— Founder and CEO, Archer
Chase Lochmiller
— CEO, Crusoe
Jason Calacanis
— Angel Investor / Founder, LAUNCH
Summary
Jason Calacanis hosts four CEO interviews from Davos 2026. Circle's Jeremy Allaire explains why the GENIUS Act, network effects and falling interest rates favor USDC's growth; CrowdStrike's George Kurtz describes AI-accelerated attacks, autonomous malware and the emerging market for securing AI agents; Archer's Adam Goldstein lays out the 2026 eVTOL pilot-city timeline, the LA28 Olympics deal and a defense partnership with Anduril; Crusoe's Chase Lochmiller discusses accelerating AI compute demand, gas-turbine shortages, batteries, small modular reactors and how the buildout is financed.
- Allaire: USDC is the fastest-growing major stablecoin, second only to Tether, with record bank engagement and roughly $60T of non-interest-bearing money to grow into.
- Allaire: USDC circulation is inversely correlated with interest rates; rate cuts since late 2023 drove multi-hundred-percent supply growth.
- Allaire: the GENIUS Act invites more competition, but stablecoins are network businesses and Circle's integrations and multi-jurisdiction licenses are its moat.
- Kurtz: AI compresses attack timelines and enables prompt-only autonomous malware; CrowdStrike counters with models trained on 14 years of attack data.
- Kurtz: AI detection and response for enterprise agents (about 90 per employee) is a massive new TAM; geopolitical tension and federal vendor consolidation support demand.
- Goldstein: DOT will name five eVTOL pilot cities in Q1 2026 with flights starting in summer; Archer bought Hawthorne airport and holds LA28 Olympics air-taxi rights.
- Goldstein: Archer is building attritable autonomous military aircraft with Anduril at about 90% lower cost than crewed helicopters.
- Lochmiller: AI compute demand is accelerating; gas turbines are a major supply constraint; Crusoe is bullish on SMRs with an Oklo-partnered plant targeted for 2027.
Ideas
USDC adoption accelerating into $60T money pool
Circle's USDC has been the fastest-growing of the major stablecoins for two years straight, is the second largest overall and by far the largest regulated stablecoin network, with transaction volume growing even faster than supply. Adoption is broadening well beyond crypto: Circle is having more engagement with banks (global money-center and regional) than ever, including a globally systemically important bank moving its own money between branches in USDC because it is faster than correspondent banking; banks want it for payments, collateral posting and capital markets, Cash App, Coinbase, Revolut, Visa, Stripe and Shopify merchants have integrated it, and the heaviest users are the largest electronic-market firms plus remittance senders avoiding 12% Western Union fees. The addressable pool is roughly $120 trillion of legal electronic money, of which about $60 trillion is physical cash and non-interest-bearing demand deposits that stablecoins can grow into, and Circle is extending its platform with the largest euro stablecoin, a conditionally approved national trust bank charter and a technology stack for other issuers' national stablecoins. Risks he flags are the still-missing market-structure and tokenization legislation and a fragmenting geopolitical landscape for a US-headquartered global network.
AI-era threats need CrowdStrike's AI defense
AI is minting new adversaries and compressing attack timelines: a mid-level hacker can use LLMs to automate attacks with nation-state sophistication, and CrowdStrike is now seeing prompt-only autonomous malware that interacts with an LLM, produces a unique fingerprint on every machine and never phones home, so signature-based antivirus cannot catch it, while most attacks are identity-based (stolen credentials, session-token replay, help-desk social engineering, North Korean operatives hired as remote employees). Kurtz argues you need AI to counter AI: CrowdStrike has used machine learning since 2011 and now trains its own models on a 14-year data set of indicators of attack, detecting the behavior of a break-in regardless of tooling, and pairs that with identity protection and a newly acquired browser-security layer that runs beneath any browser, because the browser is now the front door for both work and adversaries. It runs small language models on its own data with guardrails around the frontier models it uses.
Geopolitics and federal consolidation lift CrowdStrike demand
Whenever geopolitical tensions rise (Greenland, China, Russia), security activity rises, which is good for CrowdStrike's business; nation-state actors are already prolific, with Russia patient and low-and-slow and China commercially motivated, handing stolen trade secrets to domestic companies and raising hardware supply-chain concerns. Separately, the Trump administration is taking a business-first approach to security: it wants to save money, consolidate vendors, buy platforms and get better outcomes instead of piecemeal purchasing across government, and it is actively engaging CrowdStrike, which favors large platform vendors over point solutions.
Archer 2026 pilot cities and LA28 catalysts
Certification is the hard part, but President Trump's executive order fast-tracked eVTOLs: DOT and the FAA will name five US pilot cities in Q1 2026 and Archer expects regular flights around those cities to begin in summer 2026, building public comfort ahead of certification; Goldstein expects mostly permissive red-state locations such as Texas and Florida plus his push for Huntington Beach and Los Angeles. Archer holds exclusive air-taxi rights for the LA28 Olympics, bought Hawthorne airport next to SpaceX for about $170 million as an LA hub with real-estate residual value, flies piloted test flights most days in California and the UAE (Mubadala and IHC are investors), plans New York operations from the three existing heliports, and is signing deals with Gulf, African and Asian governments at Davos. He says the platform will certify to an order of magnitude safer than helicopters thanks to 24 redundant motors and a 50-foot glide-capable wing, and argues the retail-investor base gives the stock liquidity that lets institutions participate, so as long as Archer keeps executing toward the Morgan Stanley-projected $9 trillion market it can keep stair-stepping its valuation and raising the capital this capital-intensive business needs (about $4 billion raised so far, more likely over time).
AI power buildout strains gas turbine makers
Gas turbines are a massive supply-chain constraint as the industry rushes to energize compute: Crusoe pioneered behind-the-meter, off-grid natural-gas-powered data centers (a 350 MW on-site gas plant fed by pipeline into 10 turbines plus a 1.2 GW substation in Abilene, and a gas-rich 10 GW campus planned in Wyoming with carbon capture and 45Q credits), many others have followed that pathway, and that has created supply-chain challenges at the major producers, namely GE Vernova, Caterpillar's Solar Turbines (a close Crusoe partner), Siemens and Mitsubishi. The shortage is severe enough that Crusoe placed a $1.2 billion first order with Boom Supersonic for jet-engine-derived turbines, and he calls power a great business that will make a lot of money as AI infrastructure scales.
Batteries and solar essential for AI datacenters
Batteries are becoming core AI data-center equipment: a giant GPU cluster runs as a single computer, alternating compute cycles with network synchronization, which causes massive power-draw fluctuations that utilities hate and that are terrible for turbines, so Crusoe deployed a one-hour medium-voltage battery energy storage system to normalize load, beyond the usual low-voltage UPS role. Separately, with JB Straubel's Redwood Materials it built a fully off-grid solar-plus-storage plant using second-life EV batteries that powers an AI data center 24 hours a day at a lower power price than Northern Virginia, showing that solar plus batteries can undercut grid power for AI loads; Abilene itself was chosen because curtailed Texas wind and solar made power prices frequently negative.
AI compute demand accelerating; infrastructure retains value
Demand for AI compute is still accelerating into 2026: none of the leading labs (OpenAI, Anthropic, Google) or application companies can get enough compute, data centers rather than GPUs alone are the scarce bottleneck, and Crusoe's inbound has gone from knocking on doors to having its door knocked down, with 8,000 workers on site in Abilene, about 3,000 in Armstrong County, a planned 10 GW campus in Wyoming and 45 GW in the pipeline, while rack power density climbs from 130 kW (Blackwell) to 250 kW (Vera Rubin), 600 kW (Rubin Ultra) and eventually 1 MW. His answer to the AI-bubble question is that the buildout is financed with project equity and debt (JPMorgan, Bank of America, Apollo, SMBC) against long-term offtakes from investment-grade counterparties such as Oracle's 15-year Abilene lease, collateralized by data centers with inherent value; even if OpenAI failed, the lab that beat it would jump to take over that capacity, so the infrastructure remains valuable to whoever wins.
SMR-powered AI factory with Oklo by 2027
Crusoe is very bullish on small modular reactors: it has signed four SMR contracts and aims to energize the first SMR-powered AI factory in 2027 at the Idaho National Laboratory, where experimental technology sits outside the NRC's regulatory domain; the partner announced on the show, transcribed as Alo Energy, is Oklo. He expects SMR and next-generation geothermal costs to follow the steep decline of the solar cost curve as the industry relentlessly pursues efficiency, and argues that because grid power and data-center capacity are saturated, tech companies must bring their own new generation, which is sized to peak load and therefore leaves surplus power that lowers costs for local ratepayers.
Up Next
This All-In Podcast video, published January 25, 2026, features Jeremy Allaire, George Kurtz, Adam Goldstein, Chase Lochmiller discussing CRCL, CRWD, CIBR, ACHR, GEV, CAT, Battery energy storage, SOLAR, AI Data Center Infrastructure, OKLO. 8 trade ideas extracted by AI with direction and confidence scoring.
Speakers: Jeremy Allaire, George Kurtz, Adam Goldstein, Chase Lochmiller · Tickers: CRCL, CRWD, CIBR, ACHR, GEV, CAT, Battery energy storage, SOLAR, AI Data Center Infrastructure, OKLO