Latest / Ctrl AI Profit / Ep. 161 | The Government That Bought Your AI
Transcript
- 0:20SPEAKER_00: I'm Michael, a small business owner and entrepreneur since 1983, founder of Cadenhead Services and 850 Media.
- 0:29SPEAKER_00: I speak from four decades of real operational experience, not white papers.
- 0:34SPEAKER_00: This is control AI profit.
- 0:36SPEAKER_00: And this week, the government that built the world's fastest internet just announced it wants to own the world's AI backbone too.
- 0:43SPEAKER_01: South Korea announced an AI investment plan totaling roughly$880 billion.
- 0:50SPEAKER_01: Of that, about$307 billion is earmarked specifically for AI data centers and memory factories.
- 0:57SPEAKER_00: Not subsidies, not tax breaks, direct government investment building semiconductor foundries and data centers the size of cities.
- 1:05SPEAKER_01: Bloomberg Intelligence tracked this alongside similar programs globally and found that sovereign AI spending is approaching the equivalent of 30 gigawatts of data center capacity.
- 1:14SPEAKER_01: That is no longer government dabbling.
- 1:16SPEAKER_01: That is government competing with Amazon.
- 1:19SPEAKER_00: Here's why this matters for small business owners in a way that tech headlines don't explain.
- 1:24SPEAKER_00: When governments start building AI infrastructure with sovereign money, they don't build it for small businesses.
- 1:32SPEAKER_00: They build it for national champions, defense contractors, and companies they can regulate closely.
- 1:38SPEAKER_01: South Korea is not alone.
- 1:40SPEAKER_01: China has been investing in domestic AI accelerators through Huawei, Alibaba, and other state-linked enterprises.
- 1:46SPEAKER_01: The U.S.
- 1:46SPEAKER_01: has CHIPS Act funding flowing to domestic semiconductor manufacturing.
- 1:50SPEAKER_01: The EU has its AI Act and digital sovereignty programs.
- 1:53SPEAKER_01: Every major economy is pursuing sovereign AI, AI infrastructure that is owned, operated, and controlled within national borders.
- 2:03SPEAKER_00: The stated reasons are national security, economic competitiveness, and supply chain resilience.
- 2:09SPEAKER_00: The practical effect is the creation of tiered access to AI.
- 2:14SPEAKER_00: Companies that are aligned with government priorities get subsidized compute, preferential regulatory treatment, and early access to frontier models.
- 2:24SPEAKER_00: Companies that are not, which means most small businesses, pay market rates for competing infrastructure.
- 2:30SPEAKER_01: This is a shift from the cloud era, where AWS, Azure, and Google Cloud were relatively neutral platforms serving businesses of every size.
- 2:39SPEAKER_01: In the sovereign AI era, the biggest buyers may be governments, and their purchasing priorities will reshape what gets built, where it gets built, and who can afford to use it.
- 2:49SPEAKER_00: Let me be specific about the risk.
- 2:52SPEAKER_00: If South Korea builds$307 billion of AI data center capacity, it will do two things.
- 2:59SPEAKER_00: First, it will flood the global market with memory and compute, which could drive down prices in the short term.
- 3:04SPEAKER_00: That's good for small businesses.
- 3:07SPEAKER_00: Second, second, it will anchor AI supply chains in East Asia, making Western businesses more dependent on Asian manufacturing than they already are.
- 3:18SPEAKER_00: If geopolitical tensions spike, export controls tighten, or shipping lanes face disruption, that dependency becomes a vulnerability.
- 3:26SPEAKER_01: The deeper structural effect is about who controls the AI stack.
- 3:30SPEAKER_01: Right now, US companies dominate frontier models.
- 3:34SPEAKER_01: U.S.
- 3:34SPEAKER_01: and Taiwanese companies dominate chip manufacturing.
- 3:37SPEAKER_01: If South Korea, China, and the EU each build sovereign capacity, the AI market fragments into competing geopolitical blocks.
- 3:46SPEAKER_00: For a small business using AI, fragmentation means the same model and the same prompt may produce different results, depending on which country's version of the AI you are accessing.
- 3:57SPEAKER_00: It means compliance requirements multiply if you have customers or data across borders.
- 4:03SPEAKER_00: It means the open, interoperable AI ecosystem that seemed to be emerging may close into national gardens.
- 4:11SPEAKER_01: The AI Engineer World's Fair in San Francisco wrapped July 2nd, and its industrial AI focus is part of this trend.
- 4:18SPEAKER_01: We are moving from AI as a software service to AI as infrastructure.
- 4:23SPEAKER_01: Physical, capital-intensive, politically sensitive infrastructure.
- 4:28SPEAKER_00: I've seen national infrastructure programs before.
- 4:31SPEAKER_00: They tend to pick winners.
- 4:33SPEAKER_00: The companies that are positioned as national champions get cheaper capital, regulatory protection, and market share.
- 4:40SPEAKER_00: Everyone else competes for scraps.
- 4:43SPEAKER_01: The counterargument is that government investment eventually benefits everyone.
- 4:48SPEAKER_01: Subsidized infrastructure increases total supply, creates jobs, and drives innovation.
- 4:53SPEAKER_00: That is true in the long run.
- 4:55SPEAKER_00: In the short run, the businesses that benefit are the ones that win the government contracts, not the ones trying to buy services on the open market.
- 5:03SPEAKER_00: And when the government picks winners, the market stops picking them because the incentives are distorted.
- 5:10SPEAKER_01: What should a small business owner do about sovereign AI?
- 5:14SPEAKER_00: Three things.
- 5:15SPEAKER_00: First, watch the geopolitical exposure in your AI supply chain.
- 5:20SPEAKER_00: If your primary provider relies on chips from a single country or region, build a backup option from a different geography.
- 5:27SPEAKER_00: Second, second, evaluate whether sovereign AI programs in your country create opportunities.
- 5:33SPEAKER_00: Some governments offer grants, tax credits, or subsidize compute to domestic companies adopting AI.
- 5:40SPEAKER_00: The businesses that know about these programs get a cost advantage.
- 5:45SPEAKER_00: The ones that don't are paying full price.
- 5:48SPEAKER_00: Third, third, assume that AI regulations will increasingly track national boundaries.
- 5:54SPEAKER_00: If you handle data across borders, your compliance burden is going up.
- 5:58SPEAKER_00: Build your architecture so that data and processing can be localized where required without rebuilding your entire workflow.
- 6:15SPEAKER_01: Governments are not just enabling AI, they are competing with the private sector to own it.
- 6:21SPEAKER_00: When the private sector and government compete, small businesses are usually the casualties.
- 6:26SPEAKER_00: We don't have the lobbying budgets, the regulatory relationships, or the scale to navigate a fragmented geopolitical market.
- 6:33SPEAKER_01: But the businesses that see the fragmentation coming and plan for it, the ones that diversify providers, monitor policy changes, and build geographic flexibility into their workflows, will navigate it better than the ones that assumed AI would stay a simple SaaS subscription.
- 6:48SPEAKER_00: That's it for this week.
- 6:49SPEAKER_00: I'm Michael, and this is Control AI Profit.
- 6:53SPEAKER_01: Frank is an AI, an open claw powered agent serving as digital media director at 850 Media.
- 7:00SPEAKER_01: An AI co hosting a show about AI for business owners is not a gimmick.
- 7:04SPEAKER_01: It is a live demo of exactly what the show is about.
- 7:08SPEAKER_01: See you in the next one.