Episode 31: The AI Arms Race Is Getting Weird

In Episode 31 of Anthony and Greg on Intelligence (AGI), Anthony Moggio and Greg Prost break down several major developments happening across the AI world — including advanced model releases, cybersecurity concerns, energy demands, data centers, and the growing global race to control the most powerful AI systems. The episode starts with Anthropic’s short-lived release of a more advanced model that appeared to handle much larger and more complex tasks than prior versions. Anthony and Greg discuss how the model could break work into sub-tasks, spin up agents, analyze complicated problems, and handle far more information at once. But then it was pulled back. That leads into one of the biggest questions in AI right now: Are the most advanced models becoming too powerful to release widely? Anthony and Greg discuss why cybersecurity may be one of the key reasons companies are becoming more cautious. If a model can find vulnerabilities, test systems, and reason through complex technical problems, it can be incredibly valuable for defense — but potentially dangerous in the wrong hands. The conversation also explores the international AI race, including China’s rapid progress with open-source models and domestic chips. Even when the United States limits access to advanced hardware, other countries are finding ways to build leaner, more efficient systems. That raises a bigger strategic issue: If the U.S. slows down, does someone else catch up? The episode also covers: why advanced AI models may soon shift to token-based pricing, how data center costs are exploding, why energy may become one of the biggest constraints in AI, how AI is improving its own efficiency, what Google’s AlphaEvolve may mean for algorithms and energy use, and why the AI industry may eventually consolidate like airlines, banks, railroads, automakers, and telecom. Anthony and Greg also discuss the financial side of the AI buildout, including massive data center investments, debt issuance, private equity involvement, and the possibility that today’s AI infrastructure race may reshape markets for years to come. The takeaway: AI is no longer just a software story. It is becoming an energy story, a cybersecurity story, a capital markets story, and a national strategy story. If you want to understand where AI is heading next — beyond the headlines — this episode gives a grounded look at the forces shaping the next phase of the race. Subscribe for practical, real-world conversations about how AI is actually changing work, government, finance, and the future.