Frazis Capital Podcast

00:00 Why AI stocks are selling off02:23 The warning from credit markets04:30 Why the hyperscalers are different05:43 Micron, SK Hynix and the memory sell-off07:50 AI adoption and real customer demand09:27 Hedge fund deleveraging and forced selling12:43 Data centre debt and financing risks15:30 Semiconductor valuations after the sell-off17:16 Anthropic, open source and AI profitability18:47 Politics and data centre regulation20:48 What a recovery could look like22:42 Opportunities in healthcare23:35 ASX tech and fund closures24:38 Lessons from the software rebound27:55 The liquidity crunch and equity issuance29:55 Interest rates and the path to recovery

Webinar update

September 1st: Frazis Capital Partners Market Update

Market Update from Frazis Capital Partners

00:00 Welcome back01:20 Building the portfolio gradually02:10 Semiconductor sell-off04:00 Why active management matters05:30 AI infrastructure07:30 Software weakness11:30 Risk management13:30 Growth investing misconceptions15:30 Current opportunities in healthcare

ASX: ROAR

ASX:ROAR#ROARetf  Lioncrest Partners Pty Ltd (ACN 677 520 473 | CAR 001315184), an authorised representative of Frazis Capital Management Pty Ltd (ABN 91 638 965 910 | AFSL 521445). The Lion Active ETF (ASX:ROAR | ARSN 685 354 518) is issued by K2 Asset Management Ltd (ABN 95 085 445 094 | AFSL 244 393). All investments carry risk. You may lose some or all of your investment. Past performance is not a reliable indicator of future performance. This is for general information purposes only and does not constitute an offer. Before investing, seek independent financial advice and read the PDS and TMD available at https://lioncrestpartners.com/.

00:00 Why this AI cycle feels fundamentally different.03:00 The shift from chatbots to fully autonomous AI agents.07:00 Productivity explodes as software and workflows are built in minutes.11:00 Security risks of AI with full computer access.15:00 AI swarms and large-scale agent coordination emerge.19:00 Hardware, compute, and memory demand return.23:00 Software development accelerates ~100x.27:00 Legacy SaaS and CRM models face disruption.31:00 Market implications and underestimated software risk.35:00 Healthcare, infrastructure, and AI beneficiaries.39:00 Final thoughts on why this moment matters.

00:00 Mike joins from Thailand and reflects on Southeast Asia’s scale, energy and contrasts.02:00 This AI cycle feels fundamentally different after years of incremental updates.04:00 Productivity leaps as AI moves from copy-paste to orchestrating code and workflows.06:30 Recent breakthroughs suggest AI may already be entering a self-improving phase.09:00 Feedback loops from current usage will materially strengthen next-generation models.11:00 FRA Capital is cancelling major SaaS subscriptions and rebuilding tools internally.13:30 Hands-on AI usage is directly influencing investment positioning and sector views.15:30 The shift reinforces semiconductors and compute while pressuring legacy software.17:30 The buy-versus-build decision is swinging sharply toward in-house AI systems.19:30 Wegovy’s PBS listing signals obesity being treated as a serious medical condition.21:00 AI’s role in healthcare diagnostics continues to expand rapidly.22:30 Wrap-up on why the AI capex and compute cycle likely extends further.

0:00 – Intro0:45 – The AI takeoff moment: why early 2026 feels different2:20 – From ChatGPT to agents: the real step-change in capability4:30 – Senior developers don’t write code anymore: time compression explained6:40 – Agent swarms: why one AI is no longer the point8:30 – What’s still broken and why the gaps are closing fast10:10 – Building a diary + memoir app in days, not months12:40 – AI-powered journaling, email prompts & memory extraction14:40 – Investment meetings solved: auto-generating the next 6 questions16:30 – Vibe coding in real time18:20 – Wealth Wise: automating databases people charge fortunes for20:10 – Replacing expert networks: GLG, Tegus & mass-market disruption22:30 – $100k software projects becoming weeks-long builds24:30 – Phase 1 vs Phase 2 AI: new winners, new losers26:40 – Compute demand explodes: why GPUs, power & energy still win28:40 – Software pressure: seats, subscriptions & enterprise risk30:40 – The terminal replaces the IDE: dev tools under threat32:40 – Marketing becomes the bottleneck: attention software33:20 – Closing thoughts: massive opportunity, few people paying attention