Cerebras Systems: The Hardware Revolutionaries
Data General’s team raced to build a new 32-bit computer from scratch. Cerebras Systems is their modern hardware heir, but instead of a minicomputer, they’re building AI accelerators the size of a dinner plate. Their Wafer-Scale Engine is a single, massive
chip designed to overcome the limitations of connecting thousands of smaller GPUs. It's a radical bet on a completely different hardware architecture to challenge Nvidia's dominance, particularly in the high-speed AI inference market. By creating a single piece of silicon that eliminates communication bottlenecks, Cerebras is making a high-stakes gamble that a fundamentally new design is what it takes to power the next wave of agentic AI. Like Data General's Eagle project, it's an all-or-nothing engineering feat.
Anduril Industries: The Determined Disruptor
Data General had one giant to slay: Digital Equipment Corporation (DEC). Anduril Industries has five: the established prime defense contractors like Lockheed Martin and Northrop Grumman. Founded by Palmer Luckey, Anduril applies a Silicon Valley-style rapid development model to the traditionally slow-moving defense industry. They proactively build hardware and software, like their AI-powered Lattice OS, anticipating military needs rather than waiting for government specs. This approach—building the product first using private capital and then selling it—is a direct challenge to the sector’s decades-old business model. It’s a classic “move fast and break things” strategy aimed at a deeply entrenched, bureaucratic giant.
Relativity Space: The Manufacturing Innovators
The story of Data General was about a team finding new ways to build something complex under immense pressure. Relativity Space embodies this spirit by trying to 3D print an entire orbital rocket. Their Terran R vehicle, with a first flight targeted for 2026, is a ground-up rethinking of aerospace manufacturing. Rather than just competing on launch prices, Relativity is betting its future on a fundamentally different, and hopefully faster and cheaper, way of building. This focus on reinventing the manufacturing process itself, using automation and novel techniques to challenge incumbents like SpaceX and ULA, is a direct parallel to the engineering creativity that defined Data General's skunkworks team.
Groq: The Specialized Speed Demons
Data General’s engineers had to make specific, high-stakes technical choices to beat the competition. Groq is doing the same in the AI chip space. While Nvidia’s GPUs are the versatile workhorses of AI, Groq has focused obsessively on one thing: speed in AI inference. They developed a new architecture, the Language Processing Unit (LPU), designed to run language models at speeds that GPUs can't match. Their approach is not to beat Nvidia at everything, but to be so dramatically better at one critical task—real-time AI conversations and interactions—that a new market emerges. This strategy of carving out a niche through specialized, brilliant engineering, rather than competing head-on across the board, is a classic innovator's play reminiscent of Data General's targeted assault on the 32-bit market.
Commonwealth Fusion Systems: The Moonshot Project
The ultimate high-stakes engineering project is one with world-changing implications. At Data General, it was the future of the company. At Commonwealth Fusion Systems (CFS), it’s the future of energy. A spin-off from MIT, CFS is racing to build a compact, commercially viable fusion power plant using high-temperature superconducting magnets. Their SPARC project represents a 'bet-the-planet' level of ambition, requiring a tight-knit team of scientists and engineers to solve one of the most complex technical challenges in history. Like the Eagle team, the CFS crew is driven by the profound mission of creating something that has never existed before—a machine that could fundamentally alter the world. The pressure is immense, and the payoff is almost unimaginable.
