It is March 2026, and the headlines are dominated by the escalating conflict in Iran. The reality on the ground, however, is far less futuristic than many tech optimists predicted.
Despite the massive valuations and the relentless hype surrounding defense tech startups, the glaring truth of this war is that companies like Anduril are largely missing from the main stage. The battlefield is not yet a polished software environment, and the highly publicized "AI kill chain" is still waiting in the wings.
The Silicon Valley Echo Chamber
Over the past few years, the narrative has been intoxicating for anyone in Silicon Valley who spends their days thinking about scaling complex backend infrastructure or optimizing large language models. The pitch was that software would eat the battlefield. We were told that platforms like Anduril's Lattice would fuse unstructured data at the tactical edge, filter out the noise, and orchestrate automated swarms to neutralize threats before human operators even broke a sweat.
It sounds incredible in a pitch deck. It looks flawless in a controlled desert testing facility. But as this current conflict is brutally demonstrating, pitch decks do not win firefights.
The Persistence of Steel and Powder
Look at the actual footage and reports coming out of the Middle East right now. The skies are not filled with autonomous interceptors gracefully countering Iranian ballistics. The fight relies heavily on Patriot missile batteries, traditional F-35 sorties, B1s, and JDAMs. It is the same loud, kinetic, and deeply human warfare we have seen for decades.
The traditional defense primes (the Lockheeds, the Raytheons, the General Dynamics of the world) are still the undisputed heavyweights. When the stakes are existential, the military is reverting to what is proven, battle-hardened, and manufactured at a massive industrial scale.
We are seeing a very similar story unfold in Ukraine. Despite early optimism about Western defense tech, Anduril has visibly lagged behind in Eastern Europe. Reports from the frontlines indicate that their drones have frequently crashed or been neutralized by basic Russian electronic warfare and GPS jamming. The reality is that the conflict is being sustained by cheap, locally manufactured FPV drones and traditional weapons, while expensive Silicon Valley hardware struggles to adapt to contested environments.
Why the Startups Aren't Ready for Prime Time
The gap between a successful pilot program and prime-time readiness is enormous, though the exact reasons for this delay remain speculative. It could be that deploying delicate sensor fusion engines into chaotic environments is just too difficult right now. Alternatively, it might simply be an issue of manufacturing scale and military procurement bureaucracy. We do not know for sure if the technology itself is fundamentally flawed or if the integration process is just taking longer than expected. Regardless, Anduril and other defense startups are currently too early in their maturity curve to be the backbone of a major conflict. They still need to prove themselves.
The Burden of Proof
This is not to say that defense tech startups will never change the face of war. The shift toward software-defined defense is inevitable over a long enough timeline. But right now, they are learning a humbling lesson.
Before you can replace the old guard, you have to prove your systems are as reliable as a mechanical rifle in the mud. Until these autonomous systems can guarantee absolute resilience in the face of chaotic, heavily contested edge cases, the military will continue to rely on traditional steel.