Azeem Azhar at Exponential View: We're living at an exceptional moment in history. And this year's US election is the most consequential I have witnessed since I became aware of American elections in 1980. Reagan's two victories were critical to reviving American unity of purpose, breathing life into the economy and accelerating the demise of the Soviet Union.
But today's election will be even more pivotal.
Four tremendous forces -- AI, the energy transition, climate change and geopolitics -- will accelerate over the next president's term. The person in the Oval Office will need to be able to navigate them.
The further acceleration in artificial intelligence development will outpace even the remarkable progress of recent years. When I discussed it with AI leaders in Silicon Valley recently, I found a striking consensus: the current scaling methods that produce increasingly capable AI systems will remain effective for several years. The confidence in these scaling approaches has grown substantially in recent months. With firms now understanding the proven recipe for advancement, they are doubling down on their investments, which will result in an acceleration.
A fundamental shift is approaching in AI architecture, moving beyond today's query-response model toward agent-based workflows. These systems will comprise networks of increasingly autonomous AI components working in concert. While such systems are in their early stages, they will likely experience their own "ChatGPT moment" during the next presidential term. This will catalyse a widespread economic transformation.
The adoption of AI continues to accelerate two years after ChatGPT's debut. Startup offerings are maturing and organisational expertise is deepening which enables large enterprises to deploy AI systems more effectively across their operations. We're witnessing this shift as technology budgets increasingly flow from traditional enterprise software providers to AI pioneers like OpenAI. Beyond AI, numerous technologies are reaching (or approaching) consumer tipping points. Self-driving cars are one obvious candidate. Not so obvious are humanoid robots. The Unitree G1 is currently priced at $16k; by mid-term they should be cheap enough for upscale early adopters to be picking them up.
The next president will need to implement a comprehensive federal response to these technological and economic shifts. This should include establishing a robust framework for AI governance in the next year or two as Anthropic has called for. They will need to develop new approaches for worker protection and retraining that acknowledge the pace of technological change; to modernise the education system and prepare the workforce for an AI-enabled economy. Critical policy decisions will be needed around R&D funding, immigration of technical talent and support for regions and communities most affected by technological disruption. Success will need the dexterity of a surgeon and juggler combined.
The energy transition
The energy transition is entering a pivotal acceleration phase, one that will fundamentally reshape America's power infrastructure. For the first time in decades, US electricity consumption is projected to grow at an unprecedented rate of 4-5% annually, driven by twin forces of electrification and intelligencification1.
Several concurrent developments are propelling this surge in demand. The most important one is the electrification of the economy as we transition transport and industrial processes from fossil fuels to electrical methods.
The AI revolution is adding slightly to that transition by creating unprecedented power requirements in specific spots. Data centres now demand between 500 megawatts to over a gigawatt of power -- equivalent to the electricity needs of small cities. These hyperscale facilities represent just the tip of the iceberg as smaller AI installations proliferate nationwide, each adding to the cumulative demand.
The response to this demand is triggering a resurgence in nuclear power and accelerating the much-needed expansion in solar power deployment, both at utility scale and through distributed generation. Utility solar farms are growing in size and sophistication, while residential and commercial rooftop installations are becoming increasingly common as costs continue to fall and energy security concerns rise. This dual-track expansion of solar capacity represents a fundamental shift in how America generates and distributes power.
However, this transition faces significant challenges.