In a newsletter edition themed around Halloween treats rather than tricks, Packy McCormick makes a startling claim: the most significant technological breakthroughs of the week aren't about consumer gadgets, but about fundamentally rewriting the physics of how we make chips, generate power, and defend against biological threats. This isn't just a roundup of good news; it is an argument that the bottleneck of the 21st century—energy and manufacturing constraints—is finally cracking under the weight of radical, capital-intensive innovation. For the busy professional trying to grasp where the next decade of value will be created, McCormick provides a rare map of the industrial renaissance currently underway.
The Physics of Manufacturing
McCormick opens with a deep dive into Substrate, a Bay Area startup that has emerged from stealth with a mission to dismantle the duopoly of ASML and TSMC. The author highlights the sheer audacity of the project, noting that ASML's current extreme ultraviolet lithography machines are engineering marvels where "its mirrors are so smooth that if they were blown up to the size of Mars, the biggest imperfection would be a credit card's height tall." Yet, Substrate proposes a different path entirely.
As McCormick writes, "Unlike ASML's extreme ultraviolet (EUV) systems, which rely on complex plasma-based light sources operating at 13.5 nm wavelengths and costing over $400 million per machine, Substrate's technology harnesses compact particle accelerators to generate X-ray beams with significantly shorter wavelengths that are billions of times brighter than the sun." This shift from light-based to particle-based lithography promises to slash the cost of leading-edge wafers from $100,000 to around $10,000 by the end of the decade. The argument here is not merely about cost reduction; it is about sovereignty. McCormick emphasizes that "Substrate could make America a vertically integrated chipmaking nation," reducing reliance on Dutch and Taiwanese partners for the most critical step in the supply chain.
This framing is compelling because it connects a niche semiconductor startup to the broader geopolitical imperative of supply chain security. However, critics might note that the history of lithography is littered with promising technologies that failed to scale due to materials science hurdles or yield issues. The path from a prototype to a reliable foundry is notoriously difficult, and the timeline for such a dramatic cost reduction is aggressive.
"God Bless America, and God Bless Vertical Integrators."
The Energy Equation
The commentary then pivots to the energy crisis that underpins the AI revolution. McCormick argues that without a massive increase in power generation, the digital future stalls. He highlights Mazama Energy's breakthrough in Enhanced Geothermal Systems (EGS), which recently achieved a bottomhole temperature of 629 °F at a pilot site in Oregon. This is a critical development because, as McCormick explains, "energy extraction potential from geothermal wells increases exponentially with temperature." By reaching these superhot temperatures, Mazama aims to extract ten times more power density per well, potentially driving costs down to under 5 cents per kilowatt-hour.
McCormick connects this to the broader energy landscape by referencing the limitations of current solutions. He notes that while solar and nuclear are scaling, "the challenge has been it's just not cheap enough." The article draws a parallel to historical energy transitions, much like the shift from radioisotope thermoelectric generators (RTGs) which have powered deep space missions for decades, to the need for terrestrial baseload power. McCormick writes, "Whether Extropic succeeds in making inference more energy efficient, we're going to need lots and lots of energy to power our electric future." This is a sobering reminder that efficiency gains in AI chips are meaningless without a corresponding explosion in clean power generation.
The piece also touches on Mersenne, a startup attempting to create an "ultra compact power box" that can generate energy for years without refueling. McCormick quotes co-founder Kevin Sekniqi: "Humans have never before built power that can last for years or decades and that can be used anywhere it is needed." This vision of portable, long-duration power challenges the current dichotomy between stationary grid power and short-duration batteries.
Defense and Domestication
The final major thrust of the piece addresses the dual-use nature of artificial intelligence. McCormick argues that the best defense against AI-driven bioweapons is AI-driven biodefense. He highlights Valthos, a new startup backed by major players like OpenAI and Founders Fund, which aims to "develop frontier AI systems that identify biological threats and design medical countermeasures in real time." The author's tone here is pragmatic: "The only thing that can stop a bad guy with an AI bioweapon is a good guy with an AI biodefense tech stack." This reframes the conversation from fear of AI to the necessity of AI as a shield.
On the consumer front, McCormick discusses the launch of the NEO Home Robot by 1X, priced at $20,000. While acknowledging that the robot is currently "slow and clunky" and largely teleoperated, he sees the launch as a strategic move to gather data. "The idea with any of these is to get as many of them out into the field, trying things and collecting data, as possible," McCormick writes. This suggests that the current limitations are a feature of the learning process, not a bug of the product.
"The best biodefense is a good biodefense."
Critics might argue that relying on AI to solve AI-generated threats is a dangerous feedback loop, where the speed of defense must perfectly match the speed of offense. Furthermore, the high cost of the NEO robot, while lower than a car, remains prohibitive for the average household, raising questions about the timeline for true mass adoption.
Bottom Line
McCormick's strongest argument is that the convergence of breakthroughs in lithography, geothermal energy, and AI safety represents a genuine inflection point in industrial capacity, moving beyond incremental improvements to structural change. The piece's biggest vulnerability is its optimism bias; it assumes that engineering breakthroughs will translate smoothly to commercial viability without the usual regulatory or scaling friction. The reader should watch closely to see if Substrate's X-ray lithography and Mazama's superhot geothermal can survive the transition from pilot to production, as these are the hinges upon which the next era of abundance will turn.