I recently hosted a conversation that brought together two powerful lenses on modern security: strategic military experience and cutting-edge technology. The discussion—presented by Google—featured General David H. Petraeus, former Commander of U.S. Forces in Iraq and Afghanistan and former Director of the CIA, and me, Kent Walker, President of Global Affairs at Google and Alphabet. Together we explored how technological advances—especially artificial intelligence—are reshaping geopolitics, changing the battlefield, and redefining digital defense. In this report I summarize the key takeaways, expand on the implications, and offer practical observations and recommendations for policymakers, technologists, and citizens alike.
📌 Executive summary
In this piece I lay out the central themes from my conversation with General Petraeus and provide context and analysis. Our discussion focused on four interconnected areas:
- The return of geopolitics and great-power competition, and how AI is an accelerant in that environment.
- What gives the United States its comparative advantage—and what it must invest in to keep pace.
- Real-world examples from Ukraine demonstrating rapid innovation, agile procurement, and the role of cyber and unmanned systems in modern conflict.
- Practical priorities for digital defense: cybersecurity, post-quantum cryptography, energy and infrastructure, procurement reform, and public-private collaboration.
Throughout the report I quote key lines from General Petraeus and share my own perspective on actionable next steps.
🌐 Technological advancements and geopolitics
One of the clearest themes General Petraeus emphasized to me is that geopolitics are very much back. The lull some described as a “holiday from history” has ended, and nation-states are increasingly erecting barriers that had been falling for decades. In this renewed era of great-power competition, technology acts as both an enabler and a force multiplier for states—amplifying power, sharpening vulnerabilities, and changing the dynamics of deterrence and conflict.
Why geopolitics matter again
During our conversation, General Petraeus framed today’s environment as one with “more plates in that tent” than at any point since the end of the Cold War, and possibly since World War II. He was pointing to the sheer number and complexity of challenges modern states face: conventional military competition, cyber operations, economic coercion and supply chain competition, influence operations, competition over critical technologies like AI and quantum computing, and risks in space and maritime domains.
That complexity means strategies that once worked in isolation are insufficient. National security now requires an integrated approach that includes defense, diplomacy, economic policy, and technology policy. AI and other digital technologies are woven through each of these domains.
AI as an accelerant
AI doesn’t create geopolitical competition; it accelerates it. It lowers costs, speeds decision cycles, and multiplies the impact of capabilities—from intelligence analysis to weapons systems, from cyber defense to disinformation campaigns. As General Petraeus put it, technology is a “critical enabler of the countries in this competition.”
Gen. David H. Petraeus: "Technology is a critical enabler of the countries in this competition."
That observation has two practical implications. First, if you control the best tools and can get them into the hands of the right people quickly, you gain asymmetric advantages. Second, control of technology is not solely a matter of devices or algorithms; it is about the ecosystem that produces and diffuses those tools: universities, venture capital, legal frameworks, talent pipelines, and infrastructure.
🇺🇸 The U.S. approach to keeping pace in technology
When I asked General Petraeus about the best way for the United States to keep up—especially in the context of massive state-level programs elsewhere, such as China’s ambitious AI-plus investments—his answer returned again and again to the U.S. system: its incentives, its universities, its legal and judicial institutions, and the ecosystem of capital and talent that make American technology hubs exceptional.
The unique strengths of the U.S. innovation ecosystem
I agree with General Petraeus: much of the American advantage is systemic. Our universities attract world-class researchers. Our venture capital ecosystem converts research into startups. Our regulatory environment—when thoughtfully calibrated—can enable product deployment and scale. And our legal infrastructure supports contracts, intellectual property protection, and entrepreneurship in ways that are difficult to replicate elsewhere.
Gen. David H. Petraeus: "There is something that's truly unique about our system, about our universities... these critical mass of intellectual talent, the ecosystems of venture capital... the judicial experience, the legal experience."
But a system alone is not sufficient. That system must be actively supported and reinforced. In our conversation I highlighted a set of pillars that I believe are crucial for the U.S. to maintain leadership in AI and digital defense.
Key pillars for maintaining technological leadership
- Public investment in research and development: Government-funded R&D remains essential for foundational breakthroughs. These investments seed long-term advances in AI, quantum, materials science, and energy.
- Attracting global talent: Immigration policies and visa systems must allow the most talented researchers and engineers to contribute to the U.S. ecosystem.
- Innovation-friendly regulation: Regulatory frameworks should protect citizens while enabling experimentation and safe deployment of new technologies.
- Public-private partnerships: Effective coordination between the government and industry accelerates diffusion of critical capabilities to defense and public institutions.
- Infrastructure investment: Energy generation, transmission, and digital infrastructure must scale to meet the demands of AI-driven systems.
General Petraeus and I both stressed that implementation—diffusion and deployment—is the crucial step. Coming up with ideas is only the first phase; the harder work is ensuring those ideas reach practical users in the field.
⚡ Energy, AI, and infrastructure: a hidden dependency
One point I made that often surprises audiences is that AI’s growth is deeply tied to energy. Large AI models and data centers consume significant electricity. Without a robust, modernized grid and new sources of generation and transmission, scaling safe and secure AI systems will face constraints.
Will we have enough electricity?
As I noted in the conversation, I’m involved with investment in energy generation, management, and transmission. It’s a real question whether the grid and generation capacity will be able to satisfy projected demand by the end of the decade if we assume AI, electrification of transport, and other demands continue to grow. The answer is: possibly, but only if we take smart steps now.
Those steps include expanding diverse energy sources—including renewables and smaller modular nuclear reactors—modernizing grid management through AI-driven systems, and investing in transmission infrastructure so that generation can be delivered where it’s needed.
Using AI to make the grid more efficient
AI itself is part of the solution. Smarter grid management—using algorithms to optimize supply, predict demand, and coordinate distributed energy resources—can unlock "found energy" by reducing waste and improving utilization. I pointed out our investments in projects such as Oak Ridge and other small nuclear efforts as concrete examples of diversifying generation sources.
Kent Walker: "We are very bullish on all sorts of new kinds of alternatives... small nuclear... we're working throughout actually Virginia, PGM and others to try and re-engineer the grid in some ways."
In short, technological leadership requires a whole-of-economy approach that includes energy. AI will not only demand power; it can also help us use power more intelligently.
⚔️ Technology's impact on the balance of power: lessons from Ukraine
One of the most illustrative case studies we discussed was Ukraine. From the onset of the Russian invasion, Ukraine’s rapid innovation demonstrates how technology, when applied creatively and swiftly, can offset material disadvantages and reshape the battlefield.
Innovation under fire
General Petraeus observed that Ukraine’s innovation and application of technology have been “breathtaking.” He pointed to several striking outcomes: the sinking of roughly one-third of the Russian Black Sea fleet despite Ukraine lacking what one would traditionally call a full-fledged navy, the aggressive deployment of unmanned systems, and the use of digital procurement and logistics solutions that function like an "Amazon" for weapons components.
Gen. David H. Petraeus: "How have they sunk one-third of the Russian Black Sea fleet without what we would regard as a navy? So how do you offset these enormous disadvantages? You innovate."
That innovation includes both hardware and software: loitering munitions, drones for ISR (intelligence, surveillance, reconnaissance), sophisticated targeting systems, and networked logistics enabling rapid replenishment of gear and components.
A civilian tech influx to the military problem set
Another notable point was the rapid mobilization of civilian technologists. As Petraeus described, many Ukrainian tech entrepreneurs and engineers returned home to support the fight. They worked in “civilian clothes” but built the tools that changed the contest—creative, faster-moving solutions that bypassed traditional procurement timelines.
This mobilization shows how national resilience depends on more than just military capacity; it depends on the ability of civil society and private sector innovators to convert commercial know-how into defense-relevant applications under pressure.
Procurement lessons: speed, modularity, and simplicity
Ukraine’s approach to sourcing optics, components, and systems—where units can order off a marketplace and receive items to the front in days—highlights a lesson that’s resonant for U.S. defense acquisition: speed and adaptability can be as decisive as scale.
Gen. David H. Petraeus: "Maybe there's a lesson there for American military procurement... it is really hard to transform at the pace at which we should be transforming."
U.S. procurement is constrained by many legitimate priorities: safety, interoperability, lifecycle support, and accountability. Yet the Ukrainian experience demonstrates that, where appropriate, a more modular, commercial-friendly acquisition approach could accelerate the delivery of critical capabilities to the field.
🛡️ Cybersecurity and the evolving digital battlefield
A key domain in today’s contests is cyber. Cyber operations now span the full spectrum—from intellectual property theft and espionage to election interference, criminal ransomware, and disruptive attacks on critical infrastructure. Both General Petraeus and I emphasized that cyber is no longer a narrow military problem; it’s a whole-of-society issue.
The different categories of cyber actors
During our talk, General Petraeus outlined how different actors use cyber for different objectives:
- Some states try to steal intellectual property to leapfrog development.
- Some actors vacuum up encrypted traffic in the hope that quantum breakthroughs will let them decrypt data in the future.
- Some seek to influence elections and democratic processes.
- Some are purely criminal—extorting money via ransomware or selling access to valuable systems.
Understanding those motives matters, because defending against them requires different strategies—ranging from deterrence and diplomatic engagement to technical defenses and law enforcement cooperation.
Project Shield and defending institutions
I shared how Google supported Ukraine early in the conflict with Project Shield, which helped defend websites and critical online infrastructure. The effort highlighted how private sector capabilities can provide immediate real-world protection for public institutions when nation-states are under cyber attack.
AI as a defensive tool: the "Deep Sleep" example
One example I highlighted of AI aiding digital defense was an algorithmic tool—an agentic AI called "Deep Sleep"—which we used to detect and neutralize a vulnerability that adversaries were preparing to exploit in the wild. That instance showcased AI’s potential to autonomously identify and mitigate emerging cyber threats at machine speed.
Kent Walker: "We had the first instance where we were able to detect and diffuse a vulnerability that was about to be used in the wild... With an algorithm... a tool called Deep Sleep... we're actually training the tools to be able to identify and deal with problems that are out there."
Agentic AI tools, when responsibly developed and tightly scoped, can provide early warning and automated mitigation—freeing human defenders to focus on strategy, attribution, and complex response. But they also introduce governance and safety questions: how much autonomy should we grant these systems, how do we ensure robust oversight, and how do we manage false positives or manipulation attempts?
Staying ahead: quantum and post-quantum cryptography
Another cyber front is the advent of quantum computing. Some nations are currently archiving encrypted communications in the hope that future quantum advances will let them decrypt sensitive data. To preempt that threat, I discussed parallel efforts: pioneering quantum research while simultaneously deploying post-quantum cryptography—new cryptographic primitives resistant to quantum attacks.
Kent Walker: "We're simultaneously pioneering work on quantum and work on post-quantum cryptography."
This dual-track approach—advancing quantum computing while securing our systems against its potentially disruptive effects—is essential for maintaining long-term data confidentiality and trust in digital systems.
🔍 The fight over information and influence
Digital technologies make information manipulation cheaper and more scalable. We discussed various forms of influence operations: state-sponsored disinformation, targeted micro-propaganda, and the use of stolen data to shape narratives. Such tactics are part of a broader strategy to erode trust in institutions, shape foreign policy debates, and tilt electoral outcomes.
Why resilience matters
Combatting influence operations is not solely a technical challenge; it is also societal. Robust institutions, media literacy, transparent platforms, and responsive governance all matter. While technical measures—detection algorithms, content provenance tools, and platform policies—play important roles, societal resilience reduces the potency of targeted misinformation.
I emphasized that defending digital spaces is a multidisciplinary mission requiring coordination across government agencies, civil society, platforms, and international partners.
🏛️ Public-private partnership: how government and industry can work together
Throughout the discussion, General Petraeus and I highlighted the critical role of public-private collaboration. Neither governments nor companies can secure digital ecosystems alone.
Areas for stronger collaboration
- Shared threat intelligence: Rapid, responsible sharing of cyber threat indicators between industry and government helps speed detection and remediation.
- Joint R&D: Collaborative research projects can push frontier capabilities in AI safety, post-quantum cryptography, and secure hardware.
- Procurement pilots: Government should explore faster, modular procurement mechanisms to access commercial innovation where appropriate.
- Workforce development: Training programs and pathways that move talent between sectors enhance preparedness across the entire ecosystem.
Coordination also extends to international norms. We need multilateral conversations about the responsible use of AI in military contexts, protecting critical infrastructure, and setting guardrails on cyber operations that threaten civilian populations.
🧭 Reconciling speed and safety in defense modernization
One underlying tension is speed versus safety. The Ukrainian example shows the value of fast, iterative innovation. But governments must also ensure reliability, interoperability, and accountability for systems used by soldiers and civilians.
Where flexibility can be safely introduced
I suggested a calibrated approach:
- Tiered acquisition pathways: Create distinct channels for acquiring low-risk commercial components quickly (e.g., optics, sensors, small drones), while preserving rigorous testing for high-risk systems (e.g., lethal weapons systems or nuclear command and control).
- Sandbox environments: Allow startups and defense labs to test novel technologies in controlled environments with fast feedback loops.
- Standards and modularity: Emphasize open standards and modular designs so components can be integrated and upgraded without lengthy recertification.
These measures can capture the speed of commercial innovation while preserving safety and operational effectiveness.
🔧 What should we be investing in now?
General Petraeus and I agreed that investments must be both strategic and broad. Below I lay out a prioritized list that blends principles from our conversation with practical suggestions for government and industry.
Top investment priorities
- Research and development (AI, quantum, materials): Public funding should target foundational research that industry may underinvest in due to long horizons.
- Energy and grid modernization: Invest in generation diversity, transmission, storage, and AI-driven grid management.
- Cybersecurity and post-quantum cryptography: Accelerate deployment of quantum-resistant encryption and strengthen cyber defenses for critical infrastructure.
- Workforce and talent pipelines: Expand STEM education, reskilling programs, and immigration policies that attract top-tier global talent.
- Acquisition reform and procurement pilots: Build faster lanes for commercial innovation that preserve accountability and national security standards.
- Resilience and redundancy in supply chains: Reduce single points of failure in critical component supply chains, especially semiconductors and rare earths.
- International alliances and norms: Invest in partnerships and diplomatic frameworks to coordinate on tech governance and deterrence strategies.
These investments are not mutually exclusive. They are complementary components of a robust national strategy to maintain a competitive edge and defend democratic norms.
🔒 Governance, ethics, and AI safety
No conversation about AI and defense is complete without addressing governance and ethics. General Petraeus and I touched on the importance of aligning capabilities with values—ensuring that technological superiority is paired with responsible stewardship.
Guiding principles for AI in national security
- Human oversight: Maintain clear lines of human responsibility, particularly for decisions that have lethal or significant societal impact.
- Transparency: Where possible, provide explainability for AI-assisted decisions and maintain audit trails for accountability.
- Robust testing: Demand rigorous testing in representative environments to reduce unexpected failures and adversarial vulnerabilities.
- International coordination: Work with allies to establish norms and red lines for use of AI in military and intelligence operations.
These principles must be operationalized through policy, procurement terms, workforce training, and interoperable standards across sectors and allies.
🔁 Diffusion: moving from ideas to impact
Both Petraeus and I stressed that generating ideas is easy compared to deploying them. Diffusion—ensuring technologies are adopted where they will matter—is operationally the hardest problem. I offered several tactics to accelerate effective diffusion.
How to accelerate diffusion
- Field-forward pilots: Launch real-world pilots with front-line units or public agencies that will directly use the technologies, gather feedback, and iterate quickly.
- Commercial integration: Facilitate commercial companies to package their products in defense-friendly formats (e.g., security hardening, warranties, supply assurances).
- Shared platforms and marketplaces: Build secure procurement marketplaces for vetted components, mirroring some of the fast-delivery concepts seen in Ukraine but with scale and accountability.
- Standards and certifications: Develop modular certification pathways to reduce requalification time when swapping or upgrading components.
Diffusion is about institutional agility: the ability to absorb, validate, and operationalize external innovation without being paralyzed by process or fear of change.
🧩 The human factor: talent, leadership, and culture
Technologies are ultimately tools that require skilled people and adaptive organizations to produce outcomes. During our conversation, General Petraeus highlighted how, in times of existential threat, talent returns and contributes in unique ways. That dynamic points to a broader truth: culture and leadership matter.
Attracting and retaining talent
To continue to lead, the United States must ensure pathways for talent to participate in both commercial and national security innovation. That includes:
- Streamlined visa and immigration policies for in-demand talent.
- Career pathways that allow people to move between industry, academia, and government without punitive consequences.
- Investment in education at all levels to build the domestic pipeline for engineers, data scientists, and security experts.
We must also cultivate cultures within public institutions that are receptive to experimentation, that reward prudent risk-taking, and that value collaboration with the private sector.
🧭 Policy recommendations: a checklist
To make the above priorities actionable, I offer a practical checklist for policymakers and public sector leaders. These recommendations synthesize our conversation with operational realities.
Immediate (0–2 years)
- Increase funding for dual-use R&D programs that benefit both civilian industries and defense applications.
- Deploy post-quantum cryptographic standards across federal systems and critical infrastructure partners on an accelerated timeline.
- Establish procurement pilot programs that allow vetted commercial components to be purchased rapidly for low-to-moderate risk uses.
- Create secure marketplaces for vetted defense-relevant components and services.
Medium-term (2–5 years)
- Modernize grid infrastructure investment plans, including incentives for energy storage and small modular nuclear deployment.
- Implement cross-sector threat intelligence sharing frameworks with privacy and civil liberties safeguards.
- Expand workforce development programs and simplify pathways for industry-government rotation programs.
Long-term (5–10 years)
- Invest in sustained public funding for frontier research (AI safety, quantum computing, advanced materials).
- Build interoperable international standards and treaties for the use of AI in military contexts.
- Strengthen supply chain resilience for semiconductors and critical components through diversified sourcing and domestic manufacturing incentives.
🔎 Addressing risks: what keeps me up at night
While I’m optimistic about what technology can do for defense and security, there are risks that require urgency and humility. Chief among them are:
- Escalation risks: Rapidly automated decision-making in kinetic contexts could make crises more volatile if not properly constrained.
- Proliferation: As dual-use technologies spread, more actors may have capabilities to disrupt infrastructure or wage asymmetric campaigns.
- Concentration of power: Control over critical datasets and AI training capabilities could create new vectors of strategic dependence.
- Long-term secrecy risks: Data being hoarded now that may be decrypted later by quantum advances.
Addressing these concerns requires governance—laws, norms, technical standards—and international coordination. It also requires transparency and clear accountability in how systems are designed and deployed.
📰 Final observations: a path forward
In our wide-ranging conversation, General Petraeus and I came back to a few recurring themes that I want to end with. First, the era of great-power competition makes technology central to national security. Second, the United States has unique systemic advantages—but they must be preserved, modernized, and amplified. Third, the hard part is execution: moving ideas from lab to field, scaling solutions, and building resilient institutions that can operate at commercial pace while preserving safety and accountability.
Ukraine’s experience provides an immediate, vivid case study of what innovation under pressure looks like: nimbleness, cross-sector collaboration, and relentless focus on practical utility. Project Shield and algorithmic defenders like Deep Sleep show how industry capabilities can immediately bolster national resilience. Energy and infrastructure investments remind us that new capabilities require new foundations—literally and figuratively.
My final, pragmatic takeaway is simple: do not treat AI or any single technology as a magic bullet. Treat them as part of an integrated national strategy that includes R&D, infrastructure, procurement reform, workforce development, and international partnerships. If we invest deliberately in these areas—and if we continue to design safety, ethics, and accountability into deployments—then we can maintain an advantage that supports both our security and our values.
📣 Closing and next steps
I remain optimistic but vigilant. Technology presents enormous opportunity to strengthen digital defense, to protect democratic institutions, and to make societies safer. But realizing those benefits requires purposeful policy, sustained investment, and close collaboration between government and industry.
If you’re a policymaker, prioritize durable R&D and procurement pilots that harness commercial innovation. If you’re in industry, strengthen public-private partnerships and share threat intelligence responsibly. If you’re a citizen, ask your leaders about investments in energy resilience, encryption preparedness, and skills development. Collective action now will shape whether the coming decade is defined by resilient democracies and secure infrastructures—or by mismanaged risks and fractured governance.
Gen. David H. Petraeus: "It's not just coming up with the big ideas, it's actually implementing it... you have to have the government investing in research and development... you need to have the government enabling you to continue to attract the best and brightest from around the world."
I appreciate the opportunity to reflect on these issues with General Petraeus and to share the insights we discussed. The conversation underscores a straightforward fact: the intersection of AI, energy, cyber, and geopolitics is where the future of national security will be decided. We must be deliberate, collaborative, and fast—balancing innovation with responsibility.
For readers who want to dive deeper, follow developments in post-quantum standards, support public-private cyber initiatives, and watch how procurement pilots evolve. The next several years will be pivotal, and I’m committed—personally and professionally—to working across sectors to ensure technology enhances security for all.
