So you want to revolutionize the energy system?

The occasion is that demand from data centers has met bottlenecks at every level of the power system, a state the authors had earlier called an electricity gauntlet and expect to last three to four decades. Against an asset base built over a century, that looks to them like an opening for generational companies. The difficulty is that the sector has resisted this before, and their explanation generalizes: the more an industry is constrained by the physical behavior of molecules and electrons, the harder it is to disrupt with bits. Internet-era productivity soared in the sectors software could conquer, media, finance and retail, and stagnated in those doing real physical work. The evidence for that contrast is a productivity chart this note cannot read. Five specific obstacles follow, and each one closes off a move that works elsewhere: first-of-a-kind projects cost more than venture investors are used to writing; research timelines are long and, worse, unpredictable; commodity output means there is rarely a premium to charge for a greener version; reliability requirements rule out moving fast and breaking things; and the incumbent fuels are cheap, dense and versatile, especially when nobody charges for the emissions.

Their framework is deliberately unoriginal and they say so. Team, total addressable market, technology and timing are what any early-stage investor looks at. The point is that the margin for error is thinner here, which shows up in where they set the thresholds: an addressable market in the billions at minimum, a technology at least twice as good as the best available and preferably five to ten times, and founders with experience, because they rate the chances of a student with a good idea navigating this maze as extremely low. Since no company scores well on all four, the useful question becomes which two a given company actually needs.

That is what the wave metaphor is for. Some revolutions are built by ordinary secular trends, demographics, consumer preference, steady improvement in component technologies, that happen to coalesce; the move is to ride them. Their case is the iPhone, propelled by a stack of waves Apple did not start, though they insist riders are not passive and that the best of them amplify what they catch into something much larger. Other companies begin in still water. Tesla is their case there: no precursor product had warmed up the market, the components were absurdly expensive, analysts doubted batteries could ever be cheap enough for a mass-market car, and there was no charging infrastructure and nobody else with an interest in building any. The payoff is the pairing. Riders live or die on timing and technology, because the window between too early and too late is narrow and a visible wave draws competitors, so a technology lead is both hard to hold and essential. Makers live or die on team and market size, because for years the founder’s actual job is persuading investors, staff, vendors and partners to bet on a paradigm shift, and only a big enough prize justifies that bet. They then qualify the whole thing: the two classes are a spectrum, most companies sit in the middle, and they offer the dichotomy as a tool for assessing opportunities rather than as a finding.