An EV battery does not die when it leaves the car. A pack that has faded to 75% or 80% of its original capacity is no longer good enough for a vehicle, where range is everything — but it may have years of useful life left for a job where weight does not matter and a fade is tolerable: stationary grid or commercial storage. The question is whether anyone can prove it.That proof problem is why the 2025 wave of AI health-estimation IP is a second-life story as much as a diagnostics one. Eatron's US12299549B1 claims an AI decision engine to extend battery lifespan, and Element Energy's US12228613B1 claims state-of-health prediction modeling. Reliable, cheap state-of-health measurement is the gatekeeper for the entire second-life market — without it, a used pack is an unknown risk nobody will pay for.Does it pencil? Second life only works if three numbers line up. First, the residual value: what is a 75%-health pack actually worth for storage? Second, the cost to assess, sort and repurpose it — testing thousands of heterogeneous used packs is expensive, and cheap, accurate health estimation is what drags that cost down. Third, the alternative: if new sodium-ion or LFP cells get cheap enough, a brand-new purpose-built pack may simply beat a refurbished one.The assessment cost is where the AI IP earns its place. The classic knock on second life is that every used pack is different — different age, chemistry, abuse history — so each must be individually characterized, and individual characterization by slow physical testing is uneconomic at scale. A model that can read a pack's health quickly and accurately from its data turns a bespoke, expensive process into something closer to a sortable commodity.There is a real competitive threat to the whole second-life thesis, and batteryfolio's job is to name it. If new cells — especially cheap sodium-ion or LFP designed for stationary storage — keep falling in price, the economic case for refurbishing aged, uncertain EV packs weakens. Second life competes not just with landfill but with cheap new cells, and that race is live.For readers, the diagnostic to apply is whether a second-life business has solved the assessment-cost problem credibly. The technology to repurpose packs has existed for years; what has been missing is cheap, trustworthy health measurement at scale. The 2025 AI health IP is the industry building exactly that — which is why these patents are an economics story, not just an engineering one.