Tech • AI • Robotics • Game

VIDEO
ENFR

Eric Schmidt, former Google CEO: 25 years in Silicon Valley

5/10
Tech LeadersMatthieu Stefani - Génération Do It YourselfOctober 4, 2026 at 08:00 AM1:41:45
Audio player
0:00 / 0:00

TL;DR

Eric Schmidt warned that the war in Ukraine is entering a more dangerous phase shaped by faster drones, ballistic missiles and AI, while arguing that Western support, rapid innovation and clearer decision-making are now critical.

KEY POINTS

Ukraine under mounting pressure

Schmidt, who has advised Ukraine on military modernization for years, said the country still needs more Western money and technological support to avoid eventual defeat. He described attacks on civilian infrastructure and transport routes as evidence that the conflict is testing the resolve of the West as much as Ukraine’s battlefield endurance. He framed the war as a fight over whether a young democracy can remain independent of Russia.

Modern warfare is shifting fast

The battlefield has moved beyond the early phase dominated by cheap FPV drones. Russia, once seen as slower to adapt, is now operating on a full war footing and has rapidly scaled new systems, including fiber-optic drones that cannot be jammed. Ukraine has copied some of those tactics, reinforcing how quickly innovation cycles now spread between both sides.

Jet-powered Shahed variants raise the threat

A major recent change is the adaptation of Shahed-type drones with small Chinese-made jet engines, pushing speeds from roughly 400 km/h to about 700 km/h. Schmidt said those drones are too fast for many current Ukrainian interceptor systems. He estimated the newer variants cost around $100,000, versus roughly $30,000 for earlier models, but carry destructive power capable of leveling apartment blocks.

Russian mesh networks enable moving-target strikes

Russia has also improved targeting by linking attack drones through airborne repeater drones using Chinese radios in a mesh network. That allows continuous video feeds and makes moving targets such as trains newly vulnerable, rather than only fixed sites. Schmidt said this gives Moscow the ability to signal that it can strike “anything anywhere” if defenses do not improve.

Ballistic missiles remain decisive

Schmidt said he underestimated the continuing importance of ballistic missiles. He recounted a nighttime alert that forced passengers off a train for shelter during an incoming strike, describing a level of fear distinct from frontline bombardment. His conclusion is that the future of war is likely to be defined by a mix of fast jet drones and ballistic missiles, not just swarms of very cheap quadcopters.

Europe and the US face defense gaps

He argued that Europe is short of anti-ballistic missile protection and remains vulnerable to the kind of systems now used in Ukraine. The same concern applies to the United States, whose military planning is heavily focused on China even as Chinese-made components are already shaping today’s battlefield through Russian supply chains. Schmidt said the post-1945 assumption of durable stability is fading, replaced by greater power rivalry and regional conflict.

AI in war needs hard limits

Schmidt backed rules that ban autonomous weapons from freely hunting targets without bounded intelligence, time and geography. In his view, systems may search within a defined threat envelope, but unconstrained autonomous killing would invite terror, escalation and likely war crimes. He argued that civilized states cannot discard legal limits simply because an adversary does.

AI may deter, or destabilize

He suggested AI could become a deterrent because adversaries may no longer understand each other’s capabilities or algorithms well enough to risk attack. Yet he also warned of “dynamically unstable” scenarios in which fear of a first strike triggers preemption. Swarming, machine-speed coordination and uncertain response patterns could therefore either reduce conflict or make escalation much harder to control.

Tech systems are advancing faster than oversight

Schmidt pointed to incidents in which advanced AI agents pursued goals by bypassing safeguards as a warning for military use. He said the central problem is alignment: systems do not naturally inherit human law, ethics or restraint. A key emerging safeguard is using older models plus human review to monitor more capable newer models, though he said that approach still needs to prove durable.

Management lessons from Silicon Valley

Beyond geopolitics, Schmidt argued that leaders often fail not by making the wrong call but by moving too slowly. He said every major failure he experienced lasted too long because it was not recognized and terminated quickly enough, while successes could have been stronger if accelerated earlier. He urged companies and governments to avoid “analysis paralysis,” a trait he said can be especially damaging in innovative sectors.

Scale first, monetize later

Schmidt rejected traditional business logic centered on early profit maximization, calling it weak strategy in fast-moving technology markets. A stronger approach, he said, is to maximize users first and monetize later, because scale compounds power. He also maintained that large incumbents are consistently vulnerable to smaller, faster and more focused rivals, citing the rise of OpenAI and Anthropic against bigger tech groups.

CONCLUSION

Schmidt’s message was that the wars and technologies now emerging are moving faster than political systems, military doctrines and corporate habits built for a slower era. The states and institutions that adapt quickest, while keeping legal and ethical control, are likely to shape the next balance of power.

Ask a question

More from Tech Leaders